Research ArticleGPS Accelerometer and Smartphone Fused Smart Sensor forSHM on Real-Scale Bridges

G Michel Guzman-Acevedo1 G Esteban Vazquez-Becerra1 Jesus R Millan-Almaraz2

Hector E Rodriguez-Lozoya3 Alfredo Reyes-Salazar 3 J Ramon Gaxiola-Camacho 3

and Carlos A Martinez-Felix1

1Department of Earth and Space Sciences Autonomous University of Sinaloa 80040 Culiacan Mexico2Department of Physics and Mathematics Autonomous University of Sinaloa 80040 Culiacan Mexico3Department of Civil Engineering Autonomous University of Sinaloa 80040 Culiacan Mexico

Correspondence should be addressed to J Ramon Gaxiola-Camacho jrgaxiolauasedumx

Received 7 March 2019 Revised 4 June 2019 Accepted 10 June 2019 Published 23 June 2019

Academic Editor Reza Akhavian

Copyright copy 2019 G Michel Guzman-Acevedo et al +is is an open access article distributed under the Creative CommonsAttribution License which permits unrestricted use distribution and reproduction in anymedium provided the original work isproperly cited

During the last two decades Global Positioning System (GPS) geodetic-grade receivers and accelerometers have been imple-mented in Structural Health Monitoring (SHM) Most recently the use of sensors integrated in smartphones has been evolvingAlthough some of their capabilities are validated for small and local structures there is a gap in knowledge about the use of sensorsembedded in smartphones and other electronic devices for SHM of complex structures as bridges To contribute in this area thispaper demonstrates the application of GPS receivers accelerometers and smartphones integrating a smart sensor for the SHM ofbridges In order to validate its capabilities the alternative smart sensor is used to study a particular bridge with vibrationproblems Semistatic and dynamic displacements are obtained by means of GPS measurements Accelerations in three directionsof the bridge are determined using the accelerometer and the smartphone Based on the results of the alternative smart sensorinappropriate structural behavior is detected in the vertical direction of the bridge In addition dynamic characteristics areextracted using the smart sensor applying the Fast Fourier Transformation (FFT) and periodogram to the structural responses Asa result it is verified the applicability of the fused smart sensor for SHM on real-scale bridges

1 Introduction

Since two decades ago both geodetic-grade GPS (GlobalPositioning System) receivers and commercial accelerom-eters have been used in several studies related to StructuralHealth Monitoring (SHM) [1ndash9] +e results of these in-vestigations demonstrated that they are efficient and accu-rate devices for SHM of roads dams buildings bridges etcHowever there are some deficiencies associated to them Forexample it is well known that GPS signal is generally affectedwhen it passes through the different layers of the atmosphere(ionosphere and troposphere) In addition some objectssurrounding the GPS antenna environment also producemultipath effects In consequence the GPS signal reflects tothe receiver ie the GPS antenna obtain both direct and

indirect signals [10] As an alternative it is documented inthe literature that SHM can be performed using commercialaccelerometers [11 12] Nevertheless they present reducedprecision in real time Furthermore static and semistaticdisplacements cannot be directly obtained [3] ie a doubleintegration must be implemented for calculating displace-ment from acceleration time series [11] producing differ-ences from decimeter to even meter level with respect to realvalues [13] Some researchers have proposed to integrateGPS devices and accelerometers in implementing SHM[2 6 14] but one of the biggest issues is that both in-struments are very expensive +is problem opens an op-portunity to implement different methodologies or even newinstruments to collect data One option may be to usesensors that are generally integrated in smartphones these

HindawiAdvances in Civil EngineeringVolume 2019 Article ID 6429430 15 pageshttpsdoiorg10115520196429430

devices have been studied and it is demonstrated in certainway that they are reliable in terms of accuracy and efficiencyfor relatively small and local structures [15 16] For exampleZhao et al [17] validated the implementation of smart-phones in extracting force measurement of cables in bridgesParticularly they proposed an easy-to-handle convenientlow cost time-saving software in SHM of bridges that can beoperated in smartphones [18] In addition Ding et al [19]documented a bridge inspection and management systemusing smartphones Such a system was demonstrated to bean opportunity for management and inspection of bridgeslocated in rural and mountainous areas where the access tothemmay be very difficult Furthermore very recently it wasreported in the literature the validation of smartphonesparticularly under seismic loading using shaking tables forthe excitation of small frames [20] However to increase thestate of the art in this area additional investigations must beperformed related to the use of smartphones in SHM ofcomplex structures under serviceability conditions in realtime Furthermore measurements of the structural re-sponses of such structures should be performed consideringthe real scale of them At present smartphones containelements for storing data RAM and advanced micropro-cessors and the recorded information can be transferredusing mobile Internet WiFi or Bluetooth connectivity Inaddition to integrated accelerometer smartphones possessother sensors such as gyroscope and magnetometer [15]Hence the main difficulty about using GPS receivers ac-celerometers and smartphones in SHM of complex struc-tures as bridges may be related to the next three questions(1) do GPS receivers accelerometers and smartphonesprovide appropriate accuracy in structural response mea-surements of real-scale bridges (2) can GPS receiversaccelerometers and smartphones be complemented to eachother and (3) can GPS receivers accelerometers andsmartphones be utilized to integrate an alternative smartsensor for the SHM of bridges To address the above threeissues the main objective and contribution of this paper tothe literature is that it will demonstrate the sensing capa-bilities of GPS receivers accelerometers and smartphones interms of an alternative fused smart sensor implementing itsapplication with the study of a bridge structure with seriousvibration problems

2 Fused Smart SensormdashIntegration of GPSReceivers an Accelerometer anda Smartphone

In this research the following instruments are used to pro-pose an alternative smart sensor for SHM of bridges (1)geodetic-grade GPS receivers (2) a commercial accelerom-eter and (3) a smartphone+e above instruments are used tocollect displacement and acceleration data and after properevaluation and signal processing they are used for in-terpretation of the structural behavior of a real-scale bridge

21 Geodetic-Grade GPS Receivers +e use of GPS receiversrelated to SHM has been implemented in several

investigations [1 3ndash9] resulting that GPS is an advancedtool for this kind of purposes +e main advantage of GPS isits capability of measuring directly displacements withoutthe requirement of double integration of accelerationsgenerally necessary in other devices GPS receivers candetermine dynamic static semistatic displacements andcharacteristic frequencies of structures [7] Furthermoremodern geodetic-grade GPS receivers allow data collectionwith a sampling rate of 50 to 100Hz [21] providing highprecision and automatic operation and they do not need aline of sight between receivers On the other hand GPS maypresent some disadvantages [2 6] that can be reduced byapplying different methodologies related to signal process-ing techniques as several filters that significantly improve theresults Such filters will be discussed later in this paper Forthe smart sensor presented in this research two similargeodetic-grade GPS receivers as illustrated in Figure 1 areused for the SHM on real-scale bridges

+e brand of the GPS receiver used is Septentrio modelAsteRx-U with PolaNt-x MF antenna +e second author ofthis paper contacted Septentrio Inc for the evaluation ofsuch an equipment Some of the main specifications of thesedevices are 544 channels for tracking satellite constellation ofGPS GLONASS GALILEO BEIDOU IRNSS and QZSSRTK horizontal and vertical accuracy of 06 cm+ 05 ppmand 1 cm+ 1 ppm respectively According to the manu-facturer it obtains data from L1 L2 and L5 carrier fre-quencies just considering GPS constellation In additiontheir maximum sampling frequency is 100Hz +ereforebased on the NyquistndashShannon theorem [22] the effectivebandwidth of this receiver ranges from 0 to 50Hz Eventhough the receiver employed can record data from differentsatellite constellations in this paper the authors onlyconsidered the GPS constellation

22 Commercial Accelerometer Same as GPS receiverscommercial accelerometers have been tested in SHM [2 6]For example it was documented by Celebi [11] that structuralresponse collected using accelerometer produced positiveresults when earthquake or wind loading excites structuresAcceleration data have been used for evaluation analysis oreven design finding a relationship between damage and re-action and developing better code provisions In addition incomparison with GPS receivers accelerometers provide highersampling frequency which is one of their main advantagesHence it can be obtained better information and more dataabout dynamic characteristics of the structure under con-sideration For example some commercial accelerometers canobtain dynamic information from structures with a naturalfrequency greater than 1000Hz [23] +e rate of samplingfrequency indicates the quality of the collected behavior Forhigh sampling rates the resolution of the measuredmovementimproves significantly Moreover according to Kaloop [22]the Nyquist theorem must be satisfied

fs ge 2f0 (1)

where fs is the sampling frequency and f0 is the funda-mental frequency of the structure being monitored

2 Advances in Civil Engineering

Another benefit of accelerometers with respect to GPSreceivers is that they are less affected by external conditionsie accelerometers are more independent In some cases theaccelerometer can improve the structural monitoring pro-cess because of the quantity of data acquired [6] For thesmart sensor presented in this paper the commercial ac-celerometer used is an ETNA high-dynamic range strongmotion accelerometer (see Figure 2) Some of the mainspecifications of this accelerometer are 18 bits of resolutionwith 108 dB dynamic range cost-effective solution that cansatisfy todayrsquos most demanding applications multitaskingoperating system that allows simultaneous data acquisitionand interrogation timing accuracy to 05milliseconds due tosynchronized sampling with optional GPS timing systemremote alerting capability for system event or auto-diagnostic failure interconnectivity with other Altus fam-ily recorders for common triggering and shared GPS (op-tional) common user interface file format and supporttools with other Altus family recorders

23 Brand-New Smartphones As previously mentionedmodern smartphones have different kind of sensors such asgyroscopes magnetometers and accelerometers Howeverbecause of the scope and objectives of this paper only theintegrated accelerometer is evaluated Several investigationshave examined such a sensor incorporated in smartphonesfor example its performance has been demonstrated withrespect to commercial accelerometers for small prototypes[15] +e results of both instruments seem to be very similarin time and frequency domain However there is a gap ofknowledge in the use of sensors embedded in smartphonesin the SHM of infrastructure as bridges considering theirreal scale As previously mentioned one of the main ob-jectives of this paper is to present an alternative to the aboveconcern integrating a smart sensor To implement smart-phones in SHM some issues must be addressed such as howto preprocess the information reported by accelerometersmart sensor how to fix the smart sensor to the structureand problems related to the batteryrsquos lack of energy [16] On

the other hand since sensors are perfectly embedded insmartphones they are portable friendly and easy to handleie anyone can use them without a background knowledgeabout their physics and they are less expensive than GPSreceivers and commercial accelerometers respectively +emodel of smartphone employed in this research to be in-tegrated in the smart sensor is Samsung Galaxy S8+ (seeFigure 3)+emain characteristics of this smartphone are sizeof 1595times 734times 81mm weight 0173 kg and QualcommSnapdragon 835 processor it is powered by the octa-coreExynos 8895 processor 235GHz (Quad Core) + 19GHz(Quad) 64 bit and technology of 10 nanometers 4GB ofRAM storage of 64GB and operative system Android 70(Nougat) and accelerometer LSM6DSL (resolution of00023956299msec2) To extract acceleration time series theembedded accelerometer from the S8+ smartphone wasutilized as well as an App named Accelerometer analyzer (SeeFigure 3(b)) +is free App was created by Mobile Tools [24]It will be demonstrated how this App is an excellent tool torecord acceleration data using smartphones Furthermore theApp plots the graphics in real time and stores the data in a textfile for postprocessing

24 Fused Smart Sensor +e above three devices are in-tegrated to form the alternative smart sensor as illustrated inFigure 4 +ey are mounted in a special mechanic steelstructure developed to hold the three of them (1) GPSreceiver (2) accelerometer and (3) smartphone +e sup-porting structure consists of 3 levels where each platesupports an instrument On the top is mounted the GPSantenna this plate allows GPS antenna to have a correctorientation to North Below in the middle level the com-mercial accelerometer is located At the lowest plate thesmartphone is properly fixed

3 Calculation of Displacements UsingGPS Receivers

As will be discussed later in this paper when bridges arebeing monitored two deformations need to be considered

(a) (b)

Figure 1 GPS sensors used (a) receiver unit (b) antenna

Advances in Civil Engineering 3

(1) short- and (2) long-term periods Hence it is crucial touse an instrument that can measure both movements at thesame time Short- and long-term deformations are producedby different loading conditions of the bridge but in terms ofrisk deformations related to short period seems to be morerelevant [25] Before calculating the semistatic and dynamicdisplacements the raw GPS time series (output data fromthe GPS software) formed by coordinates from the WorldGeodetic system 84 (WGS84) should be transformed toapparent displacement time series+e purpose of this step isto obtain the GPS data around its mean value as follows [21]

Ai Ri minus1n

1113944

n

i1Ri (2)

where Ai is the apparent displacement Ri is the raw GPSdata n is the number of sampling points and i 1 n

To study the structural behavior of bridges as accurate aspossible using GPS data several researchers recommendedto use the Moving Average and Chebyshev filters over theapparent displacement time series to calculate semistatic anddynamic displacements respectively [9 21 26] +e use ofsuch filters is justified next

(a) (b)

Figure 3 Samsung Galaxy S8+ (a) general view (b) App used for acceleration measurement

(a) (b)

Figure 2 ETNA accelerometer (a) outside view (b) interior view

4 Advances in Civil Engineering

31 Semistatic Displacements Using Moving Average FilterApparent displacement is formed by short- and long-periodcomponents +e long component has a slow movementlong wavelength and low frequency It is produced by someeffects such as foundation settlement deck creep thermalexpansioncontraction and loss of prestress [25] It is re-ported in the literature that long-period component ex-presses the semistatic displacement [27] which can be usedto evaluate the movement of the bridge [22] It can becalculated by applying the moving average filter with awindow of 41 samples [26] One problem that must be facedfor determining semistatic displacement from GPS data isthat long-period component presents colored noise (chaoticbehavior) produced by environmental bias (multipathionospheric and tropospheric delay) and instrumental noise(noise produced by each GPS receiver) [28] +erefore it isimportant to take care about how the filter is used +emoving average filter presents a good performance in timedomain Consequently it can be used for smoothing timeseries It takes a specific number of elements from the timeseries based on the windowrsquos size beginning in the firstelement and their average is calculated +en it takes thesame number of elements but starting in the second elementfor calculating their average and so on +e moving averagefilter is given as [29]

Y[i] 1

M1113944

Mminus1

j0x[i + j] (3)

where x is the input signal Y is the output signal and M isthe size of window

32 Dynamic Displacements Using Chebyshev Filter Tocalculate the short period component representing the dy-namic displacement of the structure under consideration ahigh-pass filter should be applied to the apparent dis-placement time series +e short period component is a fastmovement with high frequency and short wavelength +ismovement is generally produced by traffic people earth-quake activity and so on [25] To clarify this concept the

measured displacement of a bridge using GPS can beexpressed as [30 31]

Y(n) M(n) + D(n) + N(n) (4)

where Y(n) is the apparent displacement M(n) representsthe long period interferences (environment delays) D(n) isthe dynamic displacement of the bridge and N(n) is thenoise

It is documented in the literature that dynamic dis-placements can be used to understand the structural per-formance of bridges [32] +e filter used in this article toextract dynamic displacements is the Chebyshev type I andit has been tested by some scholars [9 21 26 31] Chebyshevtype I filter can be expressed as [30]

Gn(w) Hn(jw)1113868111386811138681113868

1113868111386811138681113868 1

1 + e2 + T2n ww0( 1113857

1113969(5)

where |Hn(jw)| is the transfer function with an absolutevalue equal to the gains as an angular function of frequencyw of order n e is the ripple factor T is the Chebyshevpolynomial order and w0 is the cut-off frequency +is filteris characterized by having the ripple in the passband

4 Calculation of Displacements UsingAcceleration Records

+ere are alternative methodologies in the literature forcalculating displacement using acceleration records [6 14]that is to obtain displacements from acceleration time seriesa double integration of them is required However theproblem is how to select the method that provides betterresults or the one losing less information ie the mostefficient and accurate Obtaining displacement from accel-eration records is a difficult process It introduces biascreating a difference from decimeter to even meter levelbetween direct and indirect results [13] +is process isdescribed in terms of four steps according to Feng et al [16]Firstly a band pass filter needs to be implemented to removenoise Secondly a double integration is applied as

s(t) s0 + v0xt + 1113946t

01113946

t

0a(t)dt1113888 1113889dt (6)

where s(t) is the displacement a is the acceleration s0 is theinitial position and v0 is the initial velocity

+irdly the initial values of position and velocity are setequal to zero Finally the linear trend must be removed [14]Hence the results of this process represent the dynamicdisplacement Unfortunately because the velocity and po-sition are supposed to be zero the semistatic and staticdisplacements are lost [6] In this process is essential to use acorrect band pass filter because noise is removed per it Forthis purpose there are several alternatives For example itcan be used the Discrete Fourier Transform (DFT) a FiniteImpulse Response (FIR) filter or an Empirical Model De-composition (EMD) filter +e first of them (DFT filter)provides a no complex structure and balance [14] On theother hand EMD filter generates a set of N functions from

Figure 4 Smart sensor

Advances in Civil Engineering 5

the input signal +e resulted set of functions from EMD isdefined as [14]

a(t) 1113944N

n1Cn(t) + rn(t) (7)

where a(t) is the input signal rn(t) is the residue and Cn(t)

is the nth function which is named intrinsic mode functions(IMF) +e EMD filter is expressed as [14]

ab(t) 1113944r

nq

Cn(t) (8)

where

1lt qlt rlt n (9)

Hence the above expressions can be used to calculatedynamic displacements from acceleration time historiesrecorded by smartphone and accelerometer respectively

5 Application of GPS ReceiversAccelerometers and Smartphones in SHM

As previously documented several are the advantages anddisadvantages of GPS receivers accelerometers and smart-phones in SHM Both GPS receivers and commercial accel-erometers have limitations that can be addressed using certainmethodologies For example integrating GPS and commer-cial accelerometer the combined result may recover lost dataeg missing measurements from GPS +is integration maysolve the lack of semistatic and static components +e re-dundancy of both systems (GPS and accelerometers) providesa good performance in SHM [6] In general it is reported inthe literature that high-frequency displacements may beobtained from accelerometers and low frequency ones fromGPS +e reason of considering low frequencies from GPSand high ones from accelerometers is that some scholars [14]demonstrated that GPS is a helpful tool for measuring staticand semistatic behaviors (low-frequency displacements) butit experiences problems related to collecting data of dynamicdisplacements from structures vibrating with high frequen-cies On the other hand accelerometer records dynamicresponses very accurately without low-frequency compo-nents +erefore it is reported in the literature that accel-erometers and GPS make a reasonable complement to eachother [2 6 14] Unfortunately GPS receivers and acceler-ometers are expensive instruments Commonly the accel-eration is recorded using commercial accelerometersHowever the improved quality of acceleration sensors in-tegrated in smartphones provides a feasible alternativeGenerally their level of accuracy may be lower than acommercial accelerometer but it is reported in the literaturethat their results are very similar between them for smallstructural systems [15 16] +e contribution of this article tothe literature is that accelerometer sensors of smartphones areused together with other devices to form a smart sensor to beused in the SHM of a reinforced concrete bridge in real time+en a parametric analysis is performed with respect to GPSreceivers and commercial accelerometers

6 Dynamic Characteristics of Structures

Before proceeding further with the discussion dynamiccharacteristic of structures must be introduced Structuresare designed to resist different types of loading conditions+ey must have the capacity to reliably transmit loadingsfrom the applied point to their foundations when this ishappening movements may be produced in structures as aresponse Movements are generated following several fre-quencies depending mainly on the degrees of freedom of thestructure one of the most important frequencies is thenatural or fundamental frequency [33] To analyze fre-quencies is necessary to change the analysis from time tofrequency domain generally using the Fast Fourier Trans-formation (FFT) In general sense FFT is a computationalcomplexity improvement of DFT that separates a complexsignal into a set of sinusoidal ones where each signal has acorresponded frequency and potential Similarly there areother techniques such as periodogram It is well-documented that both FFT and periodogram algorithmscan be utilized for analyzing movements of bridge structureson frequency domain [34] Because FFT is a well-knownalgorithm its mathematical formulation is not describedhere On the other hand periodogram equation can beobserved in Equation (10) where x(t) represents the timedomain input signal X(f) is its Fourier transform and Xp(f )is the output periodogram which is calculated as the Fouriertransform of the autocorrelation function from x(t) It alsocan be alternatively calculated as the multiplication betweenX(f ) and its conjugated counterpart as [34]

Xp(f) F x(t)lowastx(minust) X(f) middot Xlowast(f) (10)

7 StructuralDescriptionof theAnalyzedBridge

+e analyzed structure in this paper using the fused smartsensor is the Juarez Bridge (see Figure 5) Such a structure islocated in Culiacan Mexico It is a reinforced concretebridge with an approximate length of 200meters +is is animportant civil infrastructure of the city Some importantfindings about the structural performance of this bridge weredocumented earlier by Vazquez-Becerra et al [26] It wasreported that the bridge presented deformations in averageof about 8 cm in the vertical direction representing anundesirable structural behavior In the above researchmaximum displacements were recorded at the ZEN2 station(see Figure 5(b)) which was reported to be the most affectedlocation over the bridge mainly because several vehicles arestopped by a traffic light close to it Considering displace-ments with respect to the vertical direction the calculatedaverage of probability of failure based on comparingsemistatic displacement with AASHTO (American Associ-ation of State Highway and Transportation Officials) de-flection limits was about 40 In addition according toVazquez-Becerra et al [26] the Juarez Bridgersquos probability offailure and its maximum displacements were calculated bythe average displacements extracted via GPS receivers lo-cated at critical points of the bridge However they only usedGPS receivers to extract the performance of the structure

6 Advances in Civil Engineering

under normal loading conditions On the other hand for theinvestigation presented in this paper the GPS performancewas improved by adding acceleration data using commercialaccelerometer and smartphone respectively +e combi-nation between GPS and commercial accelerometers hasbeen developed and reported in the literature [2 14 23]However very few information is documented in the lit-erature about using sensors embedded in smartphones torecord acceleration of real-scale bridges Consequently ifresults obtained via smartphone for bridge structures areviable this will provide a great advancement in SHM ofbridges because price and time would be drastically reducedin extracting their structural performance in real time

8 Measurement Campaign Using FusedSmart Sensor

For this study the measurement campaign was developed atthe ZEN2 station (see Figure 5(b)) because it was previouslydocumented by Vazquez-Becerra et al [26] that this locationpresented the worst performance with respect to averagedisplacements and probability of failure Additionally thesemeasurements were developed on Monday since this daypresented the highest probability of inappropriate structuralbehavior +e structural assessment of the bridge was de-veloped simultaneously in terms of the smart sensor in-tegrating GPS receivers accelerometer and smartphone+e process is comprehensively described in this section

81 GPS Receivers Campaign +e GPS measurementmethodology employed was the kinematic relative posi-tioning which involves two stations measuring simulta-neously one inmovement (the one located on the bridge) andthe other one static (outside of the bridge as static reference)[10] +e stable-based reference station used is the 373Hstation this station is located near to the bridge It is part ofthe active national geodetic network from the Mexican Na-tional Institute of Statistics and Geography (INEGI inSpanish) Figure 6(a) shows the position of station 373H withrespect to the ZEN2 station +e length of the baseline(from ZEN2 to 373H) is approximately 120meters +e

displacement measurement lasted 500 seconds using asampling rate of 100Hz when the Juarez Bridge was subjectedto in-service traffic loading conditions Figures 6(b) and 6(c)illustrate the ZEN2 and 373H stations respectively+ey weremonitored during the GPS displacements campaign

Previously to postprocessing the collected data raw GPSdata was debugged by Teqc (translation editing and qualitycheck) [35] +e preprocessing consists in insulate GPS dataand perform a quality check+e latest step is very interestingconsidering that it provides an understanding of the condi-tion of the GPS receiverrsquos position in terms of Multipath(MP) visible satellites and signal to noise ratio (S) For thiscase the results demonstrate that the rover reports multipathvalues of MP1RMS 00478m MP2RMS 00119m 19 GPSsatellites with observations S1 701 and S2 449 for thereference station MP1RMS 00428m MP2RMS 00114S1 712 and S2 449 the values of both receivers are verysimilar due to the short distance between them On the otherhand because relative positioning with double differencing(DD) over a short baseline (less than 10 kilometers of length)was used some error sources were either reduced or droppedfor example clock bias from the satellite-receiver and at-mospheric effects (ionospheric and tropospheric bias) be-cause both receivers are generally affected by the sameconditions [36] Additionally this positioning techniquemakes possible to achieve a good performance using com-mercial software for processing raw GPS data which can beseen in most of investigations related to the use of GPS re-ceivers in SHM [7 21 28 31 32] However due to the ex-cessive amount of data the software considered to process allthe information is the GAMITGLOBK TRACK developed bythe MIT (Massachusetts Institute of Technology) +is soft-ware processes GPS data in kinematic mode +e mainspecifications of GPS data processing included 15deg cut-offangle precise final orbits disseminated by IGS (InternationalGNSS Service) mode air (used for high sample) NationalGeodetic survey antenna calibration parameters and others

82 Accelerometer Campaign As previously discussedduring the 500 seconds measurement session one com-mercial accelerometer was utilized +is device was

(a) (b)

Figure 5 Juarez Bridge (a) lateral view (b) ZEN2 GPS location

Advances in Civil Engineering 7

recording acceleration at a rate of 100Hz+e accelerometerwas perfectly fixed to present the same physical center as theother devices to record the behavior of ZEN2 stationFigure 6(b) illustrates the smart sensor that is holding thethree devices One more important fact in the accelerometercampaign is that information was recorded at the same timethan GPS and smartphone providing a correct synchroni-zation [37]

83 SmartphoneCampaign As illustrated earlier in Figure 4the smartphone was fixed at the lowest plate of the mechanicstructure for the smart sensor +e measured informationfrom the accelerometer embedded in the smartphone wascollected by an open access App [24] +is application canrecord acceleration and then send it to the user in the form oftext file+e smartphone collected data at the same time thanthe GPS receiver and commercial accelerometer during500 seconds with a sampling rate of 400Hz

84 Flow Chart Diagram of the Measurement Process Insummary the data acquisition process of the bridge usingthe smart sensor is illustrated in Figure 7 Firstly theaccelerometer smartphone and GPS antenna are perfectlyfixed to the location of ZEN2 station +e particularissues related to this location were previously discussedSecondly acceleration time histories are extracted fromaccelerometer and smartphone respectively and dis-placements from GPS receivers As documented earlier thedata acquisition process is performed simultaneously usingthe three instruments +irdly the apparent displacementin three directions (N-S E-W and vertical) is calculatedusing GPS +en semistatic and dynamic displacementsare extracted using low- and high-pass filters respectivelyFourthly in order to study the dynamic characteristics ofthe Juarez Bridge FFTand periodogram are plotted for thelongitudinal transversal and vertical directions using rawacceleration of smartphone and accelerometer re-spectively Finally to compare the extracting responsecapabilities of the three instruments dynamic displace-ments reported by every device are used for the calculationof periodograms +e above discussion is clarified in thenext section

9 Structural Response Collected by theSmart Sensor

+e structural performance of the Juarez Bridge is studiedunder normal service loading conditions in real time ie thestructure is evaluated during a loading condition that is ex-perienced daily by users In fact a nonpermissible vibration inthe vertical direction has been reported by users and it hasbeen documented in the literature by Vazquez-Becerra et al[26] To verify the undesirable behavior of this bridge in thisstudy the authors decided to implement other devices duringthe SHM Hence the collected structural response is obtainedsimultaneously using GPS accelerometer and smartphone inreal time as a fused smart sensor

91 Acceleration TimeHistory One measurement session of500 seconds was performed using the three instrumentsunder consideration In terms of collected accelerationFigure 8 illustrates the recorded response of the smartphoneand accelerometer respectively Figures 8(a)ndash8(c) illustratethe longitudinal transversal and vertical acceleration re-spectively Several observations can be made correspondingto Figure 8 It can be observed that acceleration in thevertical direction is greater than longitudinal and transversaldirections In terms of displacements the same observationwas reported by [26] In addition it can be noted inFigure 8(a) that in the longitudinal direction the acceler-ation reported by smartphone and accelerometer is pre-senting some discrepancies +is issue may be justifiedbecause of the precision of commercial accelerometers insuch a direction However this structural response does notrepresent a threat of safety because of the stiffness of thebridge in the longitudinal direction In addition it can beobserved in Figure 8(c) that the critical component is as-sociated to accelerations acting in the vertical directionConversely in terms of accuracy in Figures 8(b) and 8(c) isshown that acceleration for both smartphone and acceler-ometer is matching very closely In general terms Figure 8demonstrates that acceleration recorded by smartphone andaccelerometer can be positively comparable

92 Semistatic Displacements Using GPS Considering dis-placements obtained via GPS semistatic displacements are

(a) (b) (c)

Figure 6 GPS campaign (a) location of ZEN2 and 373H (b) station ZEN2 (c) station 373H

8 Advances in Civil Engineering

extracted by applying a low-pass 13lter Semistatic dis-placements corresponding to the bridge under considerationrepresent slowmovements of the structure they are reportedin the N-S E-W and vertical direction in Figures 9(a)ndash9(c)respectively

It can be observed in Figures 9(a) and 9(b) that smallerdisplacements are present in both N-S and E-W horizontaldirections than those in the vertical one is is illustrated inFigure 9(c) it is noted that considerable great displacementsare reported in the vertical direction Hence in terms ofsemistatic displacements it can be justi13ed that the verticalcomponent is presenting critical behavior ere exist casesof bridge studies with longer dimensions such as the SanFrancisco-Oakland bay bridge that presented small dis-placements [38] However the analyzed bridge presented inthis paper is a very particular case of study whichreports unusual behavior with respect to the vertical di-rection comparing semistatic displacement with AASHTO

deection limits atrsquos why the authors of this paper de-cided to use it for the evaluation of the sensing capabilities ofthe proposed smart sensor

93 Dynamic Displacements Using GPS Movements relatedto short period component are represented by dynamicdisplacements (see Figure 10)

Dynamic displacements are obtained applying a high-pass 13lter to GPS apparent displacements Figure 10 illus-trates dynamic displacements corresponding to the bridge Itcan be noted in Figures 10(a) and 10(b) that dynamicdisplacements are relatively small for both N-S and E-Whorizontal components On the contrary the vertical di-rection is presenting excessive displacements and this isillustrated in Figure 10(c) One more time critical dis-placements are appearing for the vertical component but inthis case with respect to dynamic displacements

Longitudinal direction

100 200 300 400 5000Time (sec)

Smartphone Accelerometer

ndash1

0

1

Acc

eler

atio

n (m

s2 )

(a)

Transversal directionSmartphone Accelerometer

ndash1

0

1A

ccel

erat

ion

(ms

2 )

100 200 300 400 5000

Time (sec)

(b)

Vertical directionSmartphone Accelerometer

100 200 300 400 5000

Time (sec)

ndash2

0

2

Acc

eler

atio

n (m

s2 )

(c)

Figure 8 Acceleration time history

Accelerometer

Acceleration

High-pass filterFFTPeriodogram

Detrend

t

0

Ri ndash (1n)sumni=1 Ri

t

0

Dynamicdisplacement

Semistaticdisplacement

Low-pass filterBand-Pass filter

Smartphone Displacementfrom GPS

Figure 7 Flow chart of measurement process

Advances in Civil Engineering 9

Apparent displacementSemistatic displacement

ndash001

0

001D

ispla

cem

ent (

m)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(a)

Apparent displacementSemistatic displacement

ndash005

0

005

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(b)

Apparent displacementSemistatic displacement

ndash001

0

001

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(c)

Figure 9 Semistatic displacements in three directions (a) N-S (b) E-W and (c) vertical

NndashS

ndash001

0

001

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(a)

EndashW

ndash005

0

005

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(b)

Figure 10 Continued

10 Advances in Civil Engineering

94 FFT and Periodogram Using GPS DynamicDisplacements To identify structural dynamic character-istics of the structure under consideration FFT and perio-dograms are calculated and plotted in frequency domainusing dynamic displacements of GPS

Figure 11 illustrates FFT and periodogram in the threeGPS directions N-S (Figures 11(a) and 11(b)) E-W(Figures 11(c) and 11(d)) and vertical (Figures 11(e) and11(f)) To clarify the results they are normalized with respectto their maximum amplitude value

95 Acceleration Time History Using Accelerometer andSmartphone Acceleration time histories corresponding tothe longitudinal transversal and vertical direction of thebridge were earlier presented in Figure 8 considering bothsmartphone and accelerometer In the longitudinal direction(see Figure 8(a)) it can be noted that there are some dif-ferences in the acceleration reported by smartphone andaccelerometer such discrepancies were justi13ed earlier inSection 91 In the transversal (see Figure 8(b)) and verticaldirections (see Figure 8(c)) it is observed that smartphone

Vertical

ndash001

0

001D

ispla

cem

ent (

m)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(c)

Figure 10 Dynamic displacements in three directions (a) N-S (b) E-W and (c) vertical

NndashSFFT

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(a)

NndashSPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(b)

EndashWFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(c)

EndashWPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

VerticalFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(e)

VerticalPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 11 Dynamic frequencies in terms of normalized FFT and periodogram

Advances in Civil Engineering 11

and accelerometer are recording very similar results In timedomain these results are very promising However tocomplement them results are changed to the frequencydomain in terms of FFT and periodogram is is discussedin the next section

96 FFT and Periodogram Using Accelerometer and Smart-phone Accelerations In this section results are studied infrequency domain ey are presented in terms of FFT andperiodogram (see Figure 12)

Figure 12 summarizes the results related to FFT andperiodogram corresponding to longitudinal transversal andvertical directions It can be observed in Figure 12 that thefrequency that contributes the most to the structural re-sponse may be located approximately at 9Hz In addition itis noted that FFT in the longitudinal and transversal di-rection (see Figures 12(a) and 12(c)) is presenting the 13rstfrequency of interest at about 35Hz Based on the resultsplotted in Figure 12 it can be stated that both smartphoneand accelerometer are providing very accurate results with

respect to each other Hence it may result practical to usesmartphones for extracting dynamic characteristics of real-scale bridges

97 Comparison of Periodogram of GPS Accelerometers andSmartphones Using Dynamic Displacements To comparedata acquisition capabilities of smartphones accelerometersand GPS the periodogram of them with respect to dynamicdisplacements in the vertical direction is illustrated inFigure 13 Only vertical direction is studied because it wasdemonstrated that this direction presented a critical be-havior It is noted as reported in Figure 12 that the fre-quency that contributes the most to the structural responseis located approximately at 937Hz which is related toprevious acceleration spectra is frequency is well detectedby the three instruments Periodogram reported in Figure 13is plotted considering dynamic displacements of smart-phone accelerometer and GPS Hence the accelerationrecorded by the smartphone and accelerometer respectivelyis converted to dynamic displacement by carrying out a

SmartphoneAccelerometer

FFTmdashlongitudinal direction

0

05

1A

mpl

itude

5 10 15 200Frequency (Hz)

(a)

SmartphoneAccelerometer

Periodogrammdashlongitudinal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(b)

SmartphoneAccelerometer

FFTmdashtransversal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(c)

SmartphoneAccelerometer

Periodogrammdashtransversal direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

SmartphoneAccelerometer

FFTmdashvertical direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(e)

SmartphoneAccelerometer

Periodogrammdashvertical direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 12 Dynamic frequencies in terms of FFT and periodogram

12 Advances in Civil Engineering

double integration which may involve some discrepancies inobtaining displacements In addition to numerical in-tegration some of the dierences observed in Figure 13 maybe associated with the applied 13ltering techniques to obtaindynamic displacements On the other hand GPS reportsdisplacements directly improving considerably data pro-cessing and interpretations However it is considered im-portant to analyze the three periodograms from Figure 13 tohave a better understanding about the coincidences of somefrequencies that appear in two or more periodograms and itmay be useful to have a better idea about de dynamic be-havior of a bridge

10 Conclusions

An alternative methodology to perform SHM on real-scalebridges is presented in this paper using GPS an acceler-ometer and a smartphone integrated in a fused smartsensor Such sensor is validated for the SHM of complexstructures as bridges under normal serviceability conditionsin real time A reinforced concrete bridge was studied byusing the proposed smart sensor because it presented ex-cessive vibration problems e accuracy of this smart sensorwas demonstrated comparing the gathered data from GPSaccelerometers and smartphones in both time and fre-quency domain Semistatic and dynamic displacements aswell as accelerations demonstrated that there are criticalperformance problems on the analyzed bridge particularlyin the vertical direction Since this critical behavior wasreported by other scholars that performed the SHM withdierent devices it can be stated that this smart sensor canbe implemented for SHM on bridges in real time In ad-dition it was demonstrated that the proposed smart sensor isadequate to perform SHM in real-scale bridges Conse-quently for the bridge under consideration it was calculated

that the fundamental frequency in the horizontal direction islocated at 9Hz Furthermore it was detected the 13rst fre-quency of interest around 35Hz In order to improvefrequency domain analysis the proposed smart sensormethodology performs a periodogram algorithm over thevertical direction data to obtain a more accurate SHM It wasobserved that there is a dierence between GPS displace-ment data and accelerometer-smartphone based data It canbe justi13ed due to the nature of each measurement meth-odology where GPS base data comes from atmosphere andaccelerometer-based data comes from a vibration sensorBecause of this the noise nature of the aforementionedmeasurement devices is not the same due to its dierentmeasurement principle On the other hand it was observed acoincidence of the three instruments approximately at937Hz for the vertical component us it can explain theusefulness of fusing GPS accelerometer and smartphoneinformation in order to discriminate undesired noise thatcan be generated by atmosphere disturbances of doubleintegration error from acceleration data when it appearsonly in one device data Finally it can be stated that thispaper contributed in understanding how GPS accelerom-eters and smartphones can be successfully fused in a smartsensor to perform SHM on real-scale bridges

Data Availability

e data used to support the 13ndings of this study areavailable from the corresponding author upon request

Disclosure

Any opinions 13ndings conclusions or recommendationsexpressed in this paper are those of the authors and do notnecessarily reect the views of the sponsors

Conflicts of Interest

e authors do not have any conicts of interest to declareabout this research

Acknowledgments

e authors would like to thank Septentrio Inc for pro-viding the AsteRx-U Receiver and the PolaNt-x MF antennaused during the GPS data collection is study is based onwork partly supported by the Consejo Nacional de Ciencia yTecnologıa (CONACYT-Mexico) as scholarships for the 13rstand seventh author respectively In addition the work isalso supported by the Universidad Autonoma de Sinaloa(UAS)

References

[1] V Ashkenazi and G W Roberts ldquoExperimental monitoringof the Humber bridge using GPSrdquo Proceedings of the In-stitution of Civil Engineers-Civil Engineering vol 120 no 4pp 177ndash182 1997

[2] W S Chan Y L Xu X L Ding andW J Dai ldquoAn integratedGPSndashaccelerometer data processing technique for structural

1

09

08

07

06

05

04

03

02

01

00 5 10

Frequency (Hz)15 20

SmartphoneAccelerometerGPS (vertical)

Figure 13 Dynamic frequencies in terms of periodogram (verticaldirection)

Advances in Civil Engineering 13

deformation monitoringrdquo Journal of Geodesy vol 80 no 12pp 705ndash719 2006

[3] S B Im S Hurlebaus and Y J Kang ldquoSummary review ofGPS technology for structural health monitoringrdquo Journal ofStructural Engineering vol 139 no 10 pp 1653ndash1664 2011

[4] M R Kaloop and H Li ldquoMonitoring of bridge deformationusing GPS techniquerdquo KSCE Journal of Civil Engineeringvol 13 no 6 pp 423ndash431 2009

[5] M Kaloop E Elbeltagi J Hu and A Elrefai ldquoRecent advancesof structures monitoring and evaluation using GPS-time seriesmonitoring systems a reviewrdquo ISPRS International Journal ofGeo-Information vol 6 no 12 p 382 2017

[6] X Li C Rizos L Ge Y Tamura and A Yoshida ldquo+ecomplementary characteristics of GPS and accelerometer inmonitoring structural deformationrdquo in Proceedings of the ION2005 National Technical Meeting Institute of Navigation(ION) San Diego CA USA January 2005

[7] P Psimoulis S Pytharouli D Karambalis and S StirosldquoPotential of Global Positioning System (GPS) to measurefrequencies of oscillations of engineering structuresrdquo Journalof Sound and Vibration vol 318 no 3 pp 606ndash623 2008

[8] C Watson T Watson and R Coleman ldquoStructural moni-toring of cable-stayed bridge analysis of GPS versus modeleddeflectionsrdquo Journal of Surveying Engineering vol 133 no 1pp 23ndash28 2007

[9] C Xiong and Y Niu ldquoInvestigation of the dynamic behavior ofa super high-rise structure using RTK-GNSS techniquerdquo KSCEJournal of Civil Engineering vol 23 no 2 pp 654ndash665 2018

[10] B Hofmann-Wellenhof H Lichtenegger and E WasleGNSSndashGlobal Navigation Satellite Systems GPS GLONASSGalileo and More Springer Science amp Business MediaVienna Austria 2007

[11] M Celebi ldquoGPS in dynamic monitoring of long-periodstructuresrdquo Soil Dynamics and Earthquake Engineeringvol 20 no 5ndash8 pp 477ndash483 2000

[12] D K Kim J I Kim and M Q Feng ldquoInstrumentation ofbridges for structural health monitoringrdquo KSCE Journal ofCivil Engineering vol 5 no 3 pp 231ndash242 2001

[13] G Q Wang D M Boore H Igel and X Y Zhou ldquoSomeobservations on colocated and closely spaced strong ground-motion records of the 1999 Chi-Chi Taiwan earthquakerdquoBulletin of the Seismological Society of America vol 93 no 2pp 674ndash693 2003

[14] J Hwang H Yun S K Park D Lee and S Hong ldquoOptimalmethods of RTK-GPSaccelerometer integration to monitorthe displacement of structuresrdquo Sensors vol 12 no 1pp 1014ndash1034 2012

[15] S Dashti J D Bray J Reilly S Glaser A Bayen and E MarildquoEvaluating the reliability of phones as seismic monitoringinstrumentsrdquo Earthquake Spectra vol 30 no 2 pp 721ndash7422014

[16] M Feng Y Fukuda M Mizuta and E Ozer ldquoCitizen sensorsfor SHM use of accelerometer data from smartphonesrdquoSensors vol 15 no 2 pp 2980ndash2998 2015

[17] X Zhao R Han Y Ding et al ldquoPortable and convenient cableforce measurement using smartphonerdquo Journal of CivilStructural Health Monitoring vol 5 no 4 pp 481ndash491 2015

[18] X Zhao K Ri R Han Y Yu M Li and J Ou ldquoExperimentalresearch on quick structural health monitoring technique forbridges using smartphonerdquoAdvances inMaterials Science andEngineering vol 2016 no 1 pp 1ndash14 2016

[19] Y Ding R Han H Liu S Li X Zhao and Y Yu ldquoBridgeinspection and management system based on smart phonerdquoin Proceedings of the ASME 2016 Conference on Smart

Materials Adaptive Structures and Intelligent SystemsAmerican Society of Mechanical Engineers Stowe VT USASeptember 2016

[20] X Zhao R Han K J Loh B Xie J Li and J Ou ldquoShakingtable tests for evaluating the damage features under earth-quake excitations using smartphonesrdquo in Proceedings of theHealth Monitoring of Structural and Biological Systems XIIInternational Society for Optics and Photonics Denver COUSA May 2018

[21] F Moschas and S Stiros ldquoMeasurement of the dynamicdisplacements and of the modal frequencies of a short-spanpedestrian bridge using GPS and an accelerometerrdquo Engi-neering Structures vol 33 no 1 pp 10ndash17 2011

[22] M R Kaloop ldquoBridge safety monitoring based-GPS tech-nique case study Zhujiang Huangpu bridgerdquo Smart Struc-tures and Systems vol 9 no 6 pp 473ndash487 2012

[23] G W Roberts X Meng and A H Dodson ldquoIntegrating aglobal positioning system and accelerometers to monitor thedeflection of bridgesrdquo Journal of Surveying Engineeringvol 130 no 2 pp 65ndash72 2004

[24] Accelerometer Analyzer App 2017 httpsplaygooglecomstoreappsdetailsidcomlulaccelerometeramphles

[25] X Meng ldquoReal-time deformationmonitoring of bridges usingGPSaccelerometersrdquo Doctoral dissertation University ofNottingham UK 2002

[26] G E Vazquez-Becerra J R Gaxiola-Camacho R BennettG M Guzman-Acevedo and I E Gaxiola-CamacholdquoStructural evaluation of dynamic and semi-static displace-ments of the Juarez bridge using GPS technologyrdquo Mea-surement vol 110 no 1 pp 146ndash153 2017

[27] P Psimoulis I Peppa L Bonenberg S Ince and X MengldquoCombination of GPS and RTS measurements for themonitoring of semi-static and dynamic motion of pedestrianbridgerdquo in Proceedings of the 3rd Joint International Sym-posium on Deformation Monitoring Vienna Austria April2016

[28] F Moschas and S Stiros ldquoNoise characteristics of high-frequencyshort-duration GPS records from analysis ofidentical collocated instrumentsrdquo Measurement vol 46no 4 pp 1488ndash1506 2013

[29] S W Smithgte Scientist amp Engineerrsquos Guide to Digital SignalProcessing California Technical Publishing San Diego CAUSA 1997

[30] J Wang X Meng C Qin and J Yi ldquoVibration frequenciesextraction of the forth road bridge using high sampling GPSdatardquo Shock and Vibration vol 2016 Article ID 980786118 pages 2016

[31] J Yu XMeng X Shao B Yan and L Yang ldquoIdentification ofdynamic displacements and modal frequencies of a medium-span suspension bridge using multimode GNSS processingrdquoEngineering Structures vol 81 no 1 pp 432ndash443 2014

[32] E Elbeltagi M Kaloop and M Elnabwy ldquoMonitoring andassessment of Mansoura railway steel bridge using RTK-GPStechniquerdquo in Proceedings of the 2nd Joint InternationalSymposium on Deformation Monitoring pp 9ndash11 Notting-ham England September 2013

[33] N Raziq ldquoGPS structural deformation monitoring the mid-height problemrdquo Doctoral dissertation University ofMelbourneAustralia 2008

[34] S S HaykinAdaptive Filtergteory Pearson Education India5th edition 2014

[35] L H Estey and C M Meertens ldquoTEQC the multi-purposetoolkit for GPSGLONASS datardquo GPS Solutions vol 3 no 1pp 42ndash49 1999

14 Advances in Civil Engineering

Another benefit of accelerometers with respect to GPSreceivers is that they are less affected by external conditionsie accelerometers are more independent In some cases theaccelerometer can improve the structural monitoring pro-cess because of the quantity of data acquired [6] For thesmart sensor presented in this paper the commercial ac-celerometer used is an ETNA high-dynamic range strongmotion accelerometer (see Figure 2) Some of the mainspecifications of this accelerometer are 18 bits of resolutionwith 108 dB dynamic range cost-effective solution that cansatisfy todayrsquos most demanding applications multitaskingoperating system that allows simultaneous data acquisitionand interrogation timing accuracy to 05milliseconds due tosynchronized sampling with optional GPS timing systemremote alerting capability for system event or auto-diagnostic failure interconnectivity with other Altus fam-ily recorders for common triggering and shared GPS (op-tional) common user interface file format and supporttools with other Altus family recorders

23 Brand-New Smartphones As previously mentionedmodern smartphones have different kind of sensors such asgyroscopes magnetometers and accelerometers Howeverbecause of the scope and objectives of this paper only theintegrated accelerometer is evaluated Several investigationshave examined such a sensor incorporated in smartphonesfor example its performance has been demonstrated withrespect to commercial accelerometers for small prototypes[15] +e results of both instruments seem to be very similarin time and frequency domain However there is a gap ofknowledge in the use of sensors embedded in smartphonesin the SHM of infrastructure as bridges considering theirreal scale As previously mentioned one of the main ob-jectives of this paper is to present an alternative to the aboveconcern integrating a smart sensor To implement smart-phones in SHM some issues must be addressed such as howto preprocess the information reported by accelerometersmart sensor how to fix the smart sensor to the structureand problems related to the batteryrsquos lack of energy [16] On

the other hand since sensors are perfectly embedded insmartphones they are portable friendly and easy to handleie anyone can use them without a background knowledgeabout their physics and they are less expensive than GPSreceivers and commercial accelerometers respectively +emodel of smartphone employed in this research to be in-tegrated in the smart sensor is Samsung Galaxy S8+ (seeFigure 3)+emain characteristics of this smartphone are sizeof 1595times 734times 81mm weight 0173 kg and QualcommSnapdragon 835 processor it is powered by the octa-coreExynos 8895 processor 235GHz (Quad Core) + 19GHz(Quad) 64 bit and technology of 10 nanometers 4GB ofRAM storage of 64GB and operative system Android 70(Nougat) and accelerometer LSM6DSL (resolution of00023956299msec2) To extract acceleration time series theembedded accelerometer from the S8+ smartphone wasutilized as well as an App named Accelerometer analyzer (SeeFigure 3(b)) +is free App was created by Mobile Tools [24]It will be demonstrated how this App is an excellent tool torecord acceleration data using smartphones Furthermore theApp plots the graphics in real time and stores the data in a textfile for postprocessing

24 Fused Smart Sensor +e above three devices are in-tegrated to form the alternative smart sensor as illustrated inFigure 4 +ey are mounted in a special mechanic steelstructure developed to hold the three of them (1) GPSreceiver (2) accelerometer and (3) smartphone +e sup-porting structure consists of 3 levels where each platesupports an instrument On the top is mounted the GPSantenna this plate allows GPS antenna to have a correctorientation to North Below in the middle level the com-mercial accelerometer is located At the lowest plate thesmartphone is properly fixed

3 Calculation of Displacements UsingGPS Receivers

As will be discussed later in this paper when bridges arebeing monitored two deformations need to be considered

(a) (b)

Figure 1 GPS sensors used (a) receiver unit (b) antenna

Advances in Civil Engineering 3

(1) short- and (2) long-term periods Hence it is crucial touse an instrument that can measure both movements at thesame time Short- and long-term deformations are producedby different loading conditions of the bridge but in terms ofrisk deformations related to short period seems to be morerelevant [25] Before calculating the semistatic and dynamicdisplacements the raw GPS time series (output data fromthe GPS software) formed by coordinates from the WorldGeodetic system 84 (WGS84) should be transformed toapparent displacement time series+e purpose of this step isto obtain the GPS data around its mean value as follows [21]

Ai Ri minus1n

1113944

n

i1Ri (2)

where Ai is the apparent displacement Ri is the raw GPSdata n is the number of sampling points and i 1 n

To study the structural behavior of bridges as accurate aspossible using GPS data several researchers recommendedto use the Moving Average and Chebyshev filters over theapparent displacement time series to calculate semistatic anddynamic displacements respectively [9 21 26] +e use ofsuch filters is justified next

(a) (b)

Figure 3 Samsung Galaxy S8+ (a) general view (b) App used for acceleration measurement

(a) (b)

Figure 2 ETNA accelerometer (a) outside view (b) interior view

4 Advances in Civil Engineering

31 Semistatic Displacements Using Moving Average FilterApparent displacement is formed by short- and long-periodcomponents +e long component has a slow movementlong wavelength and low frequency It is produced by someeffects such as foundation settlement deck creep thermalexpansioncontraction and loss of prestress [25] It is re-ported in the literature that long-period component ex-presses the semistatic displacement [27] which can be usedto evaluate the movement of the bridge [22] It can becalculated by applying the moving average filter with awindow of 41 samples [26] One problem that must be facedfor determining semistatic displacement from GPS data isthat long-period component presents colored noise (chaoticbehavior) produced by environmental bias (multipathionospheric and tropospheric delay) and instrumental noise(noise produced by each GPS receiver) [28] +erefore it isimportant to take care about how the filter is used +emoving average filter presents a good performance in timedomain Consequently it can be used for smoothing timeseries It takes a specific number of elements from the timeseries based on the windowrsquos size beginning in the firstelement and their average is calculated +en it takes thesame number of elements but starting in the second elementfor calculating their average and so on +e moving averagefilter is given as [29]

Y[i] 1

M1113944

Mminus1

j0x[i + j] (3)

where x is the input signal Y is the output signal and M isthe size of window

32 Dynamic Displacements Using Chebyshev Filter Tocalculate the short period component representing the dy-namic displacement of the structure under consideration ahigh-pass filter should be applied to the apparent dis-placement time series +e short period component is a fastmovement with high frequency and short wavelength +ismovement is generally produced by traffic people earth-quake activity and so on [25] To clarify this concept the

measured displacement of a bridge using GPS can beexpressed as [30 31]

Y(n) M(n) + D(n) + N(n) (4)

where Y(n) is the apparent displacement M(n) representsthe long period interferences (environment delays) D(n) isthe dynamic displacement of the bridge and N(n) is thenoise

It is documented in the literature that dynamic dis-placements can be used to understand the structural per-formance of bridges [32] +e filter used in this article toextract dynamic displacements is the Chebyshev type I andit has been tested by some scholars [9 21 26 31] Chebyshevtype I filter can be expressed as [30]

Gn(w) Hn(jw)1113868111386811138681113868

1113868111386811138681113868 1

1 + e2 + T2n ww0( 1113857

1113969(5)

where |Hn(jw)| is the transfer function with an absolutevalue equal to the gains as an angular function of frequencyw of order n e is the ripple factor T is the Chebyshevpolynomial order and w0 is the cut-off frequency +is filteris characterized by having the ripple in the passband

4 Calculation of Displacements UsingAcceleration Records

+ere are alternative methodologies in the literature forcalculating displacement using acceleration records [6 14]that is to obtain displacements from acceleration time seriesa double integration of them is required However theproblem is how to select the method that provides betterresults or the one losing less information ie the mostefficient and accurate Obtaining displacement from accel-eration records is a difficult process It introduces biascreating a difference from decimeter to even meter levelbetween direct and indirect results [13] +is process isdescribed in terms of four steps according to Feng et al [16]Firstly a band pass filter needs to be implemented to removenoise Secondly a double integration is applied as

s(t) s0 + v0xt + 1113946t

01113946

t

0a(t)dt1113888 1113889dt (6)

where s(t) is the displacement a is the acceleration s0 is theinitial position and v0 is the initial velocity

+irdly the initial values of position and velocity are setequal to zero Finally the linear trend must be removed [14]Hence the results of this process represent the dynamicdisplacement Unfortunately because the velocity and po-sition are supposed to be zero the semistatic and staticdisplacements are lost [6] In this process is essential to use acorrect band pass filter because noise is removed per it Forthis purpose there are several alternatives For example itcan be used the Discrete Fourier Transform (DFT) a FiniteImpulse Response (FIR) filter or an Empirical Model De-composition (EMD) filter +e first of them (DFT filter)provides a no complex structure and balance [14] On theother hand EMD filter generates a set of N functions from

Figure 4 Smart sensor

Advances in Civil Engineering 5

the input signal +e resulted set of functions from EMD isdefined as [14]

a(t) 1113944N

n1Cn(t) + rn(t) (7)

where a(t) is the input signal rn(t) is the residue and Cn(t)

is the nth function which is named intrinsic mode functions(IMF) +e EMD filter is expressed as [14]

ab(t) 1113944r

nq

Cn(t) (8)

where

1lt qlt rlt n (9)

Hence the above expressions can be used to calculatedynamic displacements from acceleration time historiesrecorded by smartphone and accelerometer respectively

5 Application of GPS ReceiversAccelerometers and Smartphones in SHM

As previously documented several are the advantages anddisadvantages of GPS receivers accelerometers and smart-phones in SHM Both GPS receivers and commercial accel-erometers have limitations that can be addressed using certainmethodologies For example integrating GPS and commer-cial accelerometer the combined result may recover lost dataeg missing measurements from GPS +is integration maysolve the lack of semistatic and static components +e re-dundancy of both systems (GPS and accelerometers) providesa good performance in SHM [6] In general it is reported inthe literature that high-frequency displacements may beobtained from accelerometers and low frequency ones fromGPS +e reason of considering low frequencies from GPSand high ones from accelerometers is that some scholars [14]demonstrated that GPS is a helpful tool for measuring staticand semistatic behaviors (low-frequency displacements) butit experiences problems related to collecting data of dynamicdisplacements from structures vibrating with high frequen-cies On the other hand accelerometer records dynamicresponses very accurately without low-frequency compo-nents +erefore it is reported in the literature that accel-erometers and GPS make a reasonable complement to eachother [2 6 14] Unfortunately GPS receivers and acceler-ometers are expensive instruments Commonly the accel-eration is recorded using commercial accelerometersHowever the improved quality of acceleration sensors in-tegrated in smartphones provides a feasible alternativeGenerally their level of accuracy may be lower than acommercial accelerometer but it is reported in the literaturethat their results are very similar between them for smallstructural systems [15 16] +e contribution of this article tothe literature is that accelerometer sensors of smartphones areused together with other devices to form a smart sensor to beused in the SHM of a reinforced concrete bridge in real time+en a parametric analysis is performed with respect to GPSreceivers and commercial accelerometers

6 Dynamic Characteristics of Structures

Before proceeding further with the discussion dynamiccharacteristic of structures must be introduced Structuresare designed to resist different types of loading conditions+ey must have the capacity to reliably transmit loadingsfrom the applied point to their foundations when this ishappening movements may be produced in structures as aresponse Movements are generated following several fre-quencies depending mainly on the degrees of freedom of thestructure one of the most important frequencies is thenatural or fundamental frequency [33] To analyze fre-quencies is necessary to change the analysis from time tofrequency domain generally using the Fast Fourier Trans-formation (FFT) In general sense FFT is a computationalcomplexity improvement of DFT that separates a complexsignal into a set of sinusoidal ones where each signal has acorresponded frequency and potential Similarly there areother techniques such as periodogram It is well-documented that both FFT and periodogram algorithmscan be utilized for analyzing movements of bridge structureson frequency domain [34] Because FFT is a well-knownalgorithm its mathematical formulation is not describedhere On the other hand periodogram equation can beobserved in Equation (10) where x(t) represents the timedomain input signal X(f) is its Fourier transform and Xp(f )is the output periodogram which is calculated as the Fouriertransform of the autocorrelation function from x(t) It alsocan be alternatively calculated as the multiplication betweenX(f ) and its conjugated counterpart as [34]

Xp(f) F x(t)lowastx(minust) X(f) middot Xlowast(f) (10)

7 StructuralDescriptionof theAnalyzedBridge

+e analyzed structure in this paper using the fused smartsensor is the Juarez Bridge (see Figure 5) Such a structure islocated in Culiacan Mexico It is a reinforced concretebridge with an approximate length of 200meters +is is animportant civil infrastructure of the city Some importantfindings about the structural performance of this bridge weredocumented earlier by Vazquez-Becerra et al [26] It wasreported that the bridge presented deformations in averageof about 8 cm in the vertical direction representing anundesirable structural behavior In the above researchmaximum displacements were recorded at the ZEN2 station(see Figure 5(b)) which was reported to be the most affectedlocation over the bridge mainly because several vehicles arestopped by a traffic light close to it Considering displace-ments with respect to the vertical direction the calculatedaverage of probability of failure based on comparingsemistatic displacement with AASHTO (American Associ-ation of State Highway and Transportation Officials) de-flection limits was about 40 In addition according toVazquez-Becerra et al [26] the Juarez Bridgersquos probability offailure and its maximum displacements were calculated bythe average displacements extracted via GPS receivers lo-cated at critical points of the bridge However they only usedGPS receivers to extract the performance of the structure

6 Advances in Civil Engineering

under normal loading conditions On the other hand for theinvestigation presented in this paper the GPS performancewas improved by adding acceleration data using commercialaccelerometer and smartphone respectively +e combi-nation between GPS and commercial accelerometers hasbeen developed and reported in the literature [2 14 23]However very few information is documented in the lit-erature about using sensors embedded in smartphones torecord acceleration of real-scale bridges Consequently ifresults obtained via smartphone for bridge structures areviable this will provide a great advancement in SHM ofbridges because price and time would be drastically reducedin extracting their structural performance in real time

8 Measurement Campaign Using FusedSmart Sensor

For this study the measurement campaign was developed atthe ZEN2 station (see Figure 5(b)) because it was previouslydocumented by Vazquez-Becerra et al [26] that this locationpresented the worst performance with respect to averagedisplacements and probability of failure Additionally thesemeasurements were developed on Monday since this daypresented the highest probability of inappropriate structuralbehavior +e structural assessment of the bridge was de-veloped simultaneously in terms of the smart sensor in-tegrating GPS receivers accelerometer and smartphone+e process is comprehensively described in this section

81 GPS Receivers Campaign +e GPS measurementmethodology employed was the kinematic relative posi-tioning which involves two stations measuring simulta-neously one inmovement (the one located on the bridge) andthe other one static (outside of the bridge as static reference)[10] +e stable-based reference station used is the 373Hstation this station is located near to the bridge It is part ofthe active national geodetic network from the Mexican Na-tional Institute of Statistics and Geography (INEGI inSpanish) Figure 6(a) shows the position of station 373H withrespect to the ZEN2 station +e length of the baseline(from ZEN2 to 373H) is approximately 120meters +e

displacement measurement lasted 500 seconds using asampling rate of 100Hz when the Juarez Bridge was subjectedto in-service traffic loading conditions Figures 6(b) and 6(c)illustrate the ZEN2 and 373H stations respectively+ey weremonitored during the GPS displacements campaign

Previously to postprocessing the collected data raw GPSdata was debugged by Teqc (translation editing and qualitycheck) [35] +e preprocessing consists in insulate GPS dataand perform a quality check+e latest step is very interestingconsidering that it provides an understanding of the condi-tion of the GPS receiverrsquos position in terms of Multipath(MP) visible satellites and signal to noise ratio (S) For thiscase the results demonstrate that the rover reports multipathvalues of MP1RMS 00478m MP2RMS 00119m 19 GPSsatellites with observations S1 701 and S2 449 for thereference station MP1RMS 00428m MP2RMS 00114S1 712 and S2 449 the values of both receivers are verysimilar due to the short distance between them On the otherhand because relative positioning with double differencing(DD) over a short baseline (less than 10 kilometers of length)was used some error sources were either reduced or droppedfor example clock bias from the satellite-receiver and at-mospheric effects (ionospheric and tropospheric bias) be-cause both receivers are generally affected by the sameconditions [36] Additionally this positioning techniquemakes possible to achieve a good performance using com-mercial software for processing raw GPS data which can beseen in most of investigations related to the use of GPS re-ceivers in SHM [7 21 28 31 32] However due to the ex-cessive amount of data the software considered to process allthe information is the GAMITGLOBK TRACK developed bythe MIT (Massachusetts Institute of Technology) +is soft-ware processes GPS data in kinematic mode +e mainspecifications of GPS data processing included 15deg cut-offangle precise final orbits disseminated by IGS (InternationalGNSS Service) mode air (used for high sample) NationalGeodetic survey antenna calibration parameters and others

82 Accelerometer Campaign As previously discussedduring the 500 seconds measurement session one com-mercial accelerometer was utilized +is device was

(a) (b)

Figure 5 Juarez Bridge (a) lateral view (b) ZEN2 GPS location

Advances in Civil Engineering 7

recording acceleration at a rate of 100Hz+e accelerometerwas perfectly fixed to present the same physical center as theother devices to record the behavior of ZEN2 stationFigure 6(b) illustrates the smart sensor that is holding thethree devices One more important fact in the accelerometercampaign is that information was recorded at the same timethan GPS and smartphone providing a correct synchroni-zation [37]

83 SmartphoneCampaign As illustrated earlier in Figure 4the smartphone was fixed at the lowest plate of the mechanicstructure for the smart sensor +e measured informationfrom the accelerometer embedded in the smartphone wascollected by an open access App [24] +is application canrecord acceleration and then send it to the user in the form oftext file+e smartphone collected data at the same time thanthe GPS receiver and commercial accelerometer during500 seconds with a sampling rate of 400Hz

84 Flow Chart Diagram of the Measurement Process Insummary the data acquisition process of the bridge usingthe smart sensor is illustrated in Figure 7 Firstly theaccelerometer smartphone and GPS antenna are perfectlyfixed to the location of ZEN2 station +e particularissues related to this location were previously discussedSecondly acceleration time histories are extracted fromaccelerometer and smartphone respectively and dis-placements from GPS receivers As documented earlier thedata acquisition process is performed simultaneously usingthe three instruments +irdly the apparent displacementin three directions (N-S E-W and vertical) is calculatedusing GPS +en semistatic and dynamic displacementsare extracted using low- and high-pass filters respectivelyFourthly in order to study the dynamic characteristics ofthe Juarez Bridge FFTand periodogram are plotted for thelongitudinal transversal and vertical directions using rawacceleration of smartphone and accelerometer re-spectively Finally to compare the extracting responsecapabilities of the three instruments dynamic displace-ments reported by every device are used for the calculationof periodograms +e above discussion is clarified in thenext section

9 Structural Response Collected by theSmart Sensor

+e structural performance of the Juarez Bridge is studiedunder normal service loading conditions in real time ie thestructure is evaluated during a loading condition that is ex-perienced daily by users In fact a nonpermissible vibration inthe vertical direction has been reported by users and it hasbeen documented in the literature by Vazquez-Becerra et al[26] To verify the undesirable behavior of this bridge in thisstudy the authors decided to implement other devices duringthe SHM Hence the collected structural response is obtainedsimultaneously using GPS accelerometer and smartphone inreal time as a fused smart sensor

91 Acceleration TimeHistory One measurement session of500 seconds was performed using the three instrumentsunder consideration In terms of collected accelerationFigure 8 illustrates the recorded response of the smartphoneand accelerometer respectively Figures 8(a)ndash8(c) illustratethe longitudinal transversal and vertical acceleration re-spectively Several observations can be made correspondingto Figure 8 It can be observed that acceleration in thevertical direction is greater than longitudinal and transversaldirections In terms of displacements the same observationwas reported by [26] In addition it can be noted inFigure 8(a) that in the longitudinal direction the acceler-ation reported by smartphone and accelerometer is pre-senting some discrepancies +is issue may be justifiedbecause of the precision of commercial accelerometers insuch a direction However this structural response does notrepresent a threat of safety because of the stiffness of thebridge in the longitudinal direction In addition it can beobserved in Figure 8(c) that the critical component is as-sociated to accelerations acting in the vertical directionConversely in terms of accuracy in Figures 8(b) and 8(c) isshown that acceleration for both smartphone and acceler-ometer is matching very closely In general terms Figure 8demonstrates that acceleration recorded by smartphone andaccelerometer can be positively comparable

92 Semistatic Displacements Using GPS Considering dis-placements obtained via GPS semistatic displacements are

(a) (b) (c)

Figure 6 GPS campaign (a) location of ZEN2 and 373H (b) station ZEN2 (c) station 373H

8 Advances in Civil Engineering

extracted by applying a low-pass 13lter Semistatic dis-placements corresponding to the bridge under considerationrepresent slowmovements of the structure they are reportedin the N-S E-W and vertical direction in Figures 9(a)ndash9(c)respectively

It can be observed in Figures 9(a) and 9(b) that smallerdisplacements are present in both N-S and E-W horizontaldirections than those in the vertical one is is illustrated inFigure 9(c) it is noted that considerable great displacementsare reported in the vertical direction Hence in terms ofsemistatic displacements it can be justi13ed that the verticalcomponent is presenting critical behavior ere exist casesof bridge studies with longer dimensions such as the SanFrancisco-Oakland bay bridge that presented small dis-placements [38] However the analyzed bridge presented inthis paper is a very particular case of study whichreports unusual behavior with respect to the vertical di-rection comparing semistatic displacement with AASHTO

deection limits atrsquos why the authors of this paper de-cided to use it for the evaluation of the sensing capabilities ofthe proposed smart sensor

93 Dynamic Displacements Using GPS Movements relatedto short period component are represented by dynamicdisplacements (see Figure 10)

Dynamic displacements are obtained applying a high-pass 13lter to GPS apparent displacements Figure 10 illus-trates dynamic displacements corresponding to the bridge Itcan be noted in Figures 10(a) and 10(b) that dynamicdisplacements are relatively small for both N-S and E-Whorizontal components On the contrary the vertical di-rection is presenting excessive displacements and this isillustrated in Figure 10(c) One more time critical dis-placements are appearing for the vertical component but inthis case with respect to dynamic displacements

Longitudinal direction

100 200 300 400 5000Time (sec)

Smartphone Accelerometer

ndash1

0

1

Acc

eler

atio

n (m

s2 )

(a)

Transversal directionSmartphone Accelerometer

ndash1

0

1A

ccel

erat

ion

(ms

2 )

100 200 300 400 5000

Time (sec)

(b)

Vertical directionSmartphone Accelerometer

100 200 300 400 5000

Time (sec)

ndash2

0

2

Acc

eler

atio

n (m

s2 )

(c)

Figure 8 Acceleration time history

Accelerometer

Acceleration

High-pass filterFFTPeriodogram

Detrend

t

0

Ri ndash (1n)sumni=1 Ri

t

0

Dynamicdisplacement

Semistaticdisplacement

Low-pass filterBand-Pass filter

Smartphone Displacementfrom GPS

Figure 7 Flow chart of measurement process

Advances in Civil Engineering 9

Apparent displacementSemistatic displacement

ndash001

0

001D

ispla

cem

ent (

m)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(a)

Apparent displacementSemistatic displacement

ndash005

0

005

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(b)

Apparent displacementSemistatic displacement

ndash001

0

001

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(c)

Figure 9 Semistatic displacements in three directions (a) N-S (b) E-W and (c) vertical

NndashS

ndash001

0

001

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(a)

EndashW

ndash005

0

005

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(b)

Figure 10 Continued

10 Advances in Civil Engineering

94 FFT and Periodogram Using GPS DynamicDisplacements To identify structural dynamic character-istics of the structure under consideration FFT and perio-dograms are calculated and plotted in frequency domainusing dynamic displacements of GPS

Figure 11 illustrates FFT and periodogram in the threeGPS directions N-S (Figures 11(a) and 11(b)) E-W(Figures 11(c) and 11(d)) and vertical (Figures 11(e) and11(f)) To clarify the results they are normalized with respectto their maximum amplitude value

95 Acceleration Time History Using Accelerometer andSmartphone Acceleration time histories corresponding tothe longitudinal transversal and vertical direction of thebridge were earlier presented in Figure 8 considering bothsmartphone and accelerometer In the longitudinal direction(see Figure 8(a)) it can be noted that there are some dif-ferences in the acceleration reported by smartphone andaccelerometer such discrepancies were justi13ed earlier inSection 91 In the transversal (see Figure 8(b)) and verticaldirections (see Figure 8(c)) it is observed that smartphone

Vertical

ndash001

0

001D

ispla

cem

ent (

m)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(c)

Figure 10 Dynamic displacements in three directions (a) N-S (b) E-W and (c) vertical

NndashSFFT

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(a)

NndashSPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(b)

EndashWFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(c)

EndashWPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

VerticalFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(e)

VerticalPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 11 Dynamic frequencies in terms of normalized FFT and periodogram

Advances in Civil Engineering 11

and accelerometer are recording very similar results In timedomain these results are very promising However tocomplement them results are changed to the frequencydomain in terms of FFT and periodogram is is discussedin the next section

96 FFT and Periodogram Using Accelerometer and Smart-phone Accelerations In this section results are studied infrequency domain ey are presented in terms of FFT andperiodogram (see Figure 12)

Figure 12 summarizes the results related to FFT andperiodogram corresponding to longitudinal transversal andvertical directions It can be observed in Figure 12 that thefrequency that contributes the most to the structural re-sponse may be located approximately at 9Hz In addition itis noted that FFT in the longitudinal and transversal di-rection (see Figures 12(a) and 12(c)) is presenting the 13rstfrequency of interest at about 35Hz Based on the resultsplotted in Figure 12 it can be stated that both smartphoneand accelerometer are providing very accurate results with

respect to each other Hence it may result practical to usesmartphones for extracting dynamic characteristics of real-scale bridges

97 Comparison of Periodogram of GPS Accelerometers andSmartphones Using Dynamic Displacements To comparedata acquisition capabilities of smartphones accelerometersand GPS the periodogram of them with respect to dynamicdisplacements in the vertical direction is illustrated inFigure 13 Only vertical direction is studied because it wasdemonstrated that this direction presented a critical be-havior It is noted as reported in Figure 12 that the fre-quency that contributes the most to the structural responseis located approximately at 937Hz which is related toprevious acceleration spectra is frequency is well detectedby the three instruments Periodogram reported in Figure 13is plotted considering dynamic displacements of smart-phone accelerometer and GPS Hence the accelerationrecorded by the smartphone and accelerometer respectivelyis converted to dynamic displacement by carrying out a

SmartphoneAccelerometer

FFTmdashlongitudinal direction

0

05

1A

mpl

itude

5 10 15 200Frequency (Hz)

(a)

SmartphoneAccelerometer

Periodogrammdashlongitudinal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(b)

SmartphoneAccelerometer

FFTmdashtransversal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(c)

SmartphoneAccelerometer

Periodogrammdashtransversal direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

SmartphoneAccelerometer

FFTmdashvertical direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(e)

SmartphoneAccelerometer

Periodogrammdashvertical direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 12 Dynamic frequencies in terms of FFT and periodogram

12 Advances in Civil Engineering

double integration which may involve some discrepancies inobtaining displacements In addition to numerical in-tegration some of the dierences observed in Figure 13 maybe associated with the applied 13ltering techniques to obtaindynamic displacements On the other hand GPS reportsdisplacements directly improving considerably data pro-cessing and interpretations However it is considered im-portant to analyze the three periodograms from Figure 13 tohave a better understanding about the coincidences of somefrequencies that appear in two or more periodograms and itmay be useful to have a better idea about de dynamic be-havior of a bridge

10 Conclusions

An alternative methodology to perform SHM on real-scalebridges is presented in this paper using GPS an acceler-ometer and a smartphone integrated in a fused smartsensor Such sensor is validated for the SHM of complexstructures as bridges under normal serviceability conditionsin real time A reinforced concrete bridge was studied byusing the proposed smart sensor because it presented ex-cessive vibration problems e accuracy of this smart sensorwas demonstrated comparing the gathered data from GPSaccelerometers and smartphones in both time and fre-quency domain Semistatic and dynamic displacements aswell as accelerations demonstrated that there are criticalperformance problems on the analyzed bridge particularlyin the vertical direction Since this critical behavior wasreported by other scholars that performed the SHM withdierent devices it can be stated that this smart sensor canbe implemented for SHM on bridges in real time In ad-dition it was demonstrated that the proposed smart sensor isadequate to perform SHM in real-scale bridges Conse-quently for the bridge under consideration it was calculated

that the fundamental frequency in the horizontal direction islocated at 9Hz Furthermore it was detected the 13rst fre-quency of interest around 35Hz In order to improvefrequency domain analysis the proposed smart sensormethodology performs a periodogram algorithm over thevertical direction data to obtain a more accurate SHM It wasobserved that there is a dierence between GPS displace-ment data and accelerometer-smartphone based data It canbe justi13ed due to the nature of each measurement meth-odology where GPS base data comes from atmosphere andaccelerometer-based data comes from a vibration sensorBecause of this the noise nature of the aforementionedmeasurement devices is not the same due to its dierentmeasurement principle On the other hand it was observed acoincidence of the three instruments approximately at937Hz for the vertical component us it can explain theusefulness of fusing GPS accelerometer and smartphoneinformation in order to discriminate undesired noise thatcan be generated by atmosphere disturbances of doubleintegration error from acceleration data when it appearsonly in one device data Finally it can be stated that thispaper contributed in understanding how GPS accelerom-eters and smartphones can be successfully fused in a smartsensor to perform SHM on real-scale bridges

Data Availability

e data used to support the 13ndings of this study areavailable from the corresponding author upon request

Disclosure

Any opinions 13ndings conclusions or recommendationsexpressed in this paper are those of the authors and do notnecessarily reect the views of the sponsors

Conflicts of Interest

e authors do not have any conicts of interest to declareabout this research

Acknowledgments

e authors would like to thank Septentrio Inc for pro-viding the AsteRx-U Receiver and the PolaNt-x MF antennaused during the GPS data collection is study is based onwork partly supported by the Consejo Nacional de Ciencia yTecnologıa (CONACYT-Mexico) as scholarships for the 13rstand seventh author respectively In addition the work isalso supported by the Universidad Autonoma de Sinaloa(UAS)

References

[1] V Ashkenazi and G W Roberts ldquoExperimental monitoringof the Humber bridge using GPSrdquo Proceedings of the In-stitution of Civil Engineers-Civil Engineering vol 120 no 4pp 177ndash182 1997

[2] W S Chan Y L Xu X L Ding andW J Dai ldquoAn integratedGPSndashaccelerometer data processing technique for structural

1

09

08

07

06

05

04

03

02

01

00 5 10

Frequency (Hz)15 20

SmartphoneAccelerometerGPS (vertical)

Figure 13 Dynamic frequencies in terms of periodogram (verticaldirection)

Advances in Civil Engineering 13

deformation monitoringrdquo Journal of Geodesy vol 80 no 12pp 705ndash719 2006

[3] S B Im S Hurlebaus and Y J Kang ldquoSummary review ofGPS technology for structural health monitoringrdquo Journal ofStructural Engineering vol 139 no 10 pp 1653ndash1664 2011

[4] M R Kaloop and H Li ldquoMonitoring of bridge deformationusing GPS techniquerdquo KSCE Journal of Civil Engineeringvol 13 no 6 pp 423ndash431 2009

[5] M Kaloop E Elbeltagi J Hu and A Elrefai ldquoRecent advancesof structures monitoring and evaluation using GPS-time seriesmonitoring systems a reviewrdquo ISPRS International Journal ofGeo-Information vol 6 no 12 p 382 2017

[6] X Li C Rizos L Ge Y Tamura and A Yoshida ldquo+ecomplementary characteristics of GPS and accelerometer inmonitoring structural deformationrdquo in Proceedings of the ION2005 National Technical Meeting Institute of Navigation(ION) San Diego CA USA January 2005

[7] P Psimoulis S Pytharouli D Karambalis and S StirosldquoPotential of Global Positioning System (GPS) to measurefrequencies of oscillations of engineering structuresrdquo Journalof Sound and Vibration vol 318 no 3 pp 606ndash623 2008

[8] C Watson T Watson and R Coleman ldquoStructural moni-toring of cable-stayed bridge analysis of GPS versus modeleddeflectionsrdquo Journal of Surveying Engineering vol 133 no 1pp 23ndash28 2007

[9] C Xiong and Y Niu ldquoInvestigation of the dynamic behavior ofa super high-rise structure using RTK-GNSS techniquerdquo KSCEJournal of Civil Engineering vol 23 no 2 pp 654ndash665 2018

[10] B Hofmann-Wellenhof H Lichtenegger and E WasleGNSSndashGlobal Navigation Satellite Systems GPS GLONASSGalileo and More Springer Science amp Business MediaVienna Austria 2007

[11] M Celebi ldquoGPS in dynamic monitoring of long-periodstructuresrdquo Soil Dynamics and Earthquake Engineeringvol 20 no 5ndash8 pp 477ndash483 2000

[12] D K Kim J I Kim and M Q Feng ldquoInstrumentation ofbridges for structural health monitoringrdquo KSCE Journal ofCivil Engineering vol 5 no 3 pp 231ndash242 2001

[13] G Q Wang D M Boore H Igel and X Y Zhou ldquoSomeobservations on colocated and closely spaced strong ground-motion records of the 1999 Chi-Chi Taiwan earthquakerdquoBulletin of the Seismological Society of America vol 93 no 2pp 674ndash693 2003

[14] J Hwang H Yun S K Park D Lee and S Hong ldquoOptimalmethods of RTK-GPSaccelerometer integration to monitorthe displacement of structuresrdquo Sensors vol 12 no 1pp 1014ndash1034 2012

[15] S Dashti J D Bray J Reilly S Glaser A Bayen and E MarildquoEvaluating the reliability of phones as seismic monitoringinstrumentsrdquo Earthquake Spectra vol 30 no 2 pp 721ndash7422014

[16] M Feng Y Fukuda M Mizuta and E Ozer ldquoCitizen sensorsfor SHM use of accelerometer data from smartphonesrdquoSensors vol 15 no 2 pp 2980ndash2998 2015

[17] X Zhao R Han Y Ding et al ldquoPortable and convenient cableforce measurement using smartphonerdquo Journal of CivilStructural Health Monitoring vol 5 no 4 pp 481ndash491 2015

[18] X Zhao K Ri R Han Y Yu M Li and J Ou ldquoExperimentalresearch on quick structural health monitoring technique forbridges using smartphonerdquoAdvances inMaterials Science andEngineering vol 2016 no 1 pp 1ndash14 2016

[19] Y Ding R Han H Liu S Li X Zhao and Y Yu ldquoBridgeinspection and management system based on smart phonerdquoin Proceedings of the ASME 2016 Conference on Smart

Materials Adaptive Structures and Intelligent SystemsAmerican Society of Mechanical Engineers Stowe VT USASeptember 2016

[20] X Zhao R Han K J Loh B Xie J Li and J Ou ldquoShakingtable tests for evaluating the damage features under earth-quake excitations using smartphonesrdquo in Proceedings of theHealth Monitoring of Structural and Biological Systems XIIInternational Society for Optics and Photonics Denver COUSA May 2018

[21] F Moschas and S Stiros ldquoMeasurement of the dynamicdisplacements and of the modal frequencies of a short-spanpedestrian bridge using GPS and an accelerometerrdquo Engi-neering Structures vol 33 no 1 pp 10ndash17 2011

[22] M R Kaloop ldquoBridge safety monitoring based-GPS tech-nique case study Zhujiang Huangpu bridgerdquo Smart Struc-tures and Systems vol 9 no 6 pp 473ndash487 2012

[23] G W Roberts X Meng and A H Dodson ldquoIntegrating aglobal positioning system and accelerometers to monitor thedeflection of bridgesrdquo Journal of Surveying Engineeringvol 130 no 2 pp 65ndash72 2004

[24] Accelerometer Analyzer App 2017 httpsplaygooglecomstoreappsdetailsidcomlulaccelerometeramphles

[25] X Meng ldquoReal-time deformationmonitoring of bridges usingGPSaccelerometersrdquo Doctoral dissertation University ofNottingham UK 2002

[26] G E Vazquez-Becerra J R Gaxiola-Camacho R BennettG M Guzman-Acevedo and I E Gaxiola-CamacholdquoStructural evaluation of dynamic and semi-static displace-ments of the Juarez bridge using GPS technologyrdquo Mea-surement vol 110 no 1 pp 146ndash153 2017

[27] P Psimoulis I Peppa L Bonenberg S Ince and X MengldquoCombination of GPS and RTS measurements for themonitoring of semi-static and dynamic motion of pedestrianbridgerdquo in Proceedings of the 3rd Joint International Sym-posium on Deformation Monitoring Vienna Austria April2016

[28] F Moschas and S Stiros ldquoNoise characteristics of high-frequencyshort-duration GPS records from analysis ofidentical collocated instrumentsrdquo Measurement vol 46no 4 pp 1488ndash1506 2013

[29] S W Smithgte Scientist amp Engineerrsquos Guide to Digital SignalProcessing California Technical Publishing San Diego CAUSA 1997

[30] J Wang X Meng C Qin and J Yi ldquoVibration frequenciesextraction of the forth road bridge using high sampling GPSdatardquo Shock and Vibration vol 2016 Article ID 980786118 pages 2016

[31] J Yu XMeng X Shao B Yan and L Yang ldquoIdentification ofdynamic displacements and modal frequencies of a medium-span suspension bridge using multimode GNSS processingrdquoEngineering Structures vol 81 no 1 pp 432ndash443 2014

[32] E Elbeltagi M Kaloop and M Elnabwy ldquoMonitoring andassessment of Mansoura railway steel bridge using RTK-GPStechniquerdquo in Proceedings of the 2nd Joint InternationalSymposium on Deformation Monitoring pp 9ndash11 Notting-ham England September 2013

[33] N Raziq ldquoGPS structural deformation monitoring the mid-height problemrdquo Doctoral dissertation University ofMelbourneAustralia 2008

[34] S S HaykinAdaptive Filtergteory Pearson Education India5th edition 2014

[35] L H Estey and C M Meertens ldquoTEQC the multi-purposetoolkit for GPSGLONASS datardquo GPS Solutions vol 3 no 1pp 42ndash49 1999

14 Advances in Civil Engineering

(1) short- and (2) long-term periods Hence it is crucial touse an instrument that can measure both movements at thesame time Short- and long-term deformations are producedby different loading conditions of the bridge but in terms ofrisk deformations related to short period seems to be morerelevant [25] Before calculating the semistatic and dynamicdisplacements the raw GPS time series (output data fromthe GPS software) formed by coordinates from the WorldGeodetic system 84 (WGS84) should be transformed toapparent displacement time series+e purpose of this step isto obtain the GPS data around its mean value as follows [21]

Ai Ri minus1n

1113944

n

i1Ri (2)

where Ai is the apparent displacement Ri is the raw GPSdata n is the number of sampling points and i 1 n

To study the structural behavior of bridges as accurate aspossible using GPS data several researchers recommendedto use the Moving Average and Chebyshev filters over theapparent displacement time series to calculate semistatic anddynamic displacements respectively [9 21 26] +e use ofsuch filters is justified next

(a) (b)

Figure 3 Samsung Galaxy S8+ (a) general view (b) App used for acceleration measurement

(a) (b)

Figure 2 ETNA accelerometer (a) outside view (b) interior view

4 Advances in Civil Engineering

31 Semistatic Displacements Using Moving Average FilterApparent displacement is formed by short- and long-periodcomponents +e long component has a slow movementlong wavelength and low frequency It is produced by someeffects such as foundation settlement deck creep thermalexpansioncontraction and loss of prestress [25] It is re-ported in the literature that long-period component ex-presses the semistatic displacement [27] which can be usedto evaluate the movement of the bridge [22] It can becalculated by applying the moving average filter with awindow of 41 samples [26] One problem that must be facedfor determining semistatic displacement from GPS data isthat long-period component presents colored noise (chaoticbehavior) produced by environmental bias (multipathionospheric and tropospheric delay) and instrumental noise(noise produced by each GPS receiver) [28] +erefore it isimportant to take care about how the filter is used +emoving average filter presents a good performance in timedomain Consequently it can be used for smoothing timeseries It takes a specific number of elements from the timeseries based on the windowrsquos size beginning in the firstelement and their average is calculated +en it takes thesame number of elements but starting in the second elementfor calculating their average and so on +e moving averagefilter is given as [29]

Y[i] 1

M1113944

Mminus1

j0x[i + j] (3)

where x is the input signal Y is the output signal and M isthe size of window

32 Dynamic Displacements Using Chebyshev Filter Tocalculate the short period component representing the dy-namic displacement of the structure under consideration ahigh-pass filter should be applied to the apparent dis-placement time series +e short period component is a fastmovement with high frequency and short wavelength +ismovement is generally produced by traffic people earth-quake activity and so on [25] To clarify this concept the

measured displacement of a bridge using GPS can beexpressed as [30 31]

Y(n) M(n) + D(n) + N(n) (4)

where Y(n) is the apparent displacement M(n) representsthe long period interferences (environment delays) D(n) isthe dynamic displacement of the bridge and N(n) is thenoise

It is documented in the literature that dynamic dis-placements can be used to understand the structural per-formance of bridges [32] +e filter used in this article toextract dynamic displacements is the Chebyshev type I andit has been tested by some scholars [9 21 26 31] Chebyshevtype I filter can be expressed as [30]

Gn(w) Hn(jw)1113868111386811138681113868

1113868111386811138681113868 1

1 + e2 + T2n ww0( 1113857

1113969(5)

where |Hn(jw)| is the transfer function with an absolutevalue equal to the gains as an angular function of frequencyw of order n e is the ripple factor T is the Chebyshevpolynomial order and w0 is the cut-off frequency +is filteris characterized by having the ripple in the passband

4 Calculation of Displacements UsingAcceleration Records

+ere are alternative methodologies in the literature forcalculating displacement using acceleration records [6 14]that is to obtain displacements from acceleration time seriesa double integration of them is required However theproblem is how to select the method that provides betterresults or the one losing less information ie the mostefficient and accurate Obtaining displacement from accel-eration records is a difficult process It introduces biascreating a difference from decimeter to even meter levelbetween direct and indirect results [13] +is process isdescribed in terms of four steps according to Feng et al [16]Firstly a band pass filter needs to be implemented to removenoise Secondly a double integration is applied as

s(t) s0 + v0xt + 1113946t

01113946

t

0a(t)dt1113888 1113889dt (6)

where s(t) is the displacement a is the acceleration s0 is theinitial position and v0 is the initial velocity

+irdly the initial values of position and velocity are setequal to zero Finally the linear trend must be removed [14]Hence the results of this process represent the dynamicdisplacement Unfortunately because the velocity and po-sition are supposed to be zero the semistatic and staticdisplacements are lost [6] In this process is essential to use acorrect band pass filter because noise is removed per it Forthis purpose there are several alternatives For example itcan be used the Discrete Fourier Transform (DFT) a FiniteImpulse Response (FIR) filter or an Empirical Model De-composition (EMD) filter +e first of them (DFT filter)provides a no complex structure and balance [14] On theother hand EMD filter generates a set of N functions from

Figure 4 Smart sensor

Advances in Civil Engineering 5

the input signal +e resulted set of functions from EMD isdefined as [14]

a(t) 1113944N

n1Cn(t) + rn(t) (7)

where a(t) is the input signal rn(t) is the residue and Cn(t)

is the nth function which is named intrinsic mode functions(IMF) +e EMD filter is expressed as [14]

ab(t) 1113944r

nq

Cn(t) (8)

where

1lt qlt rlt n (9)

Hence the above expressions can be used to calculatedynamic displacements from acceleration time historiesrecorded by smartphone and accelerometer respectively

5 Application of GPS ReceiversAccelerometers and Smartphones in SHM

As previously documented several are the advantages anddisadvantages of GPS receivers accelerometers and smart-phones in SHM Both GPS receivers and commercial accel-erometers have limitations that can be addressed using certainmethodologies For example integrating GPS and commer-cial accelerometer the combined result may recover lost dataeg missing measurements from GPS +is integration maysolve the lack of semistatic and static components +e re-dundancy of both systems (GPS and accelerometers) providesa good performance in SHM [6] In general it is reported inthe literature that high-frequency displacements may beobtained from accelerometers and low frequency ones fromGPS +e reason of considering low frequencies from GPSand high ones from accelerometers is that some scholars [14]demonstrated that GPS is a helpful tool for measuring staticand semistatic behaviors (low-frequency displacements) butit experiences problems related to collecting data of dynamicdisplacements from structures vibrating with high frequen-cies On the other hand accelerometer records dynamicresponses very accurately without low-frequency compo-nents +erefore it is reported in the literature that accel-erometers and GPS make a reasonable complement to eachother [2 6 14] Unfortunately GPS receivers and acceler-ometers are expensive instruments Commonly the accel-eration is recorded using commercial accelerometersHowever the improved quality of acceleration sensors in-tegrated in smartphones provides a feasible alternativeGenerally their level of accuracy may be lower than acommercial accelerometer but it is reported in the literaturethat their results are very similar between them for smallstructural systems [15 16] +e contribution of this article tothe literature is that accelerometer sensors of smartphones areused together with other devices to form a smart sensor to beused in the SHM of a reinforced concrete bridge in real time+en a parametric analysis is performed with respect to GPSreceivers and commercial accelerometers

6 Dynamic Characteristics of Structures

Before proceeding further with the discussion dynamiccharacteristic of structures must be introduced Structuresare designed to resist different types of loading conditions+ey must have the capacity to reliably transmit loadingsfrom the applied point to their foundations when this ishappening movements may be produced in structures as aresponse Movements are generated following several fre-quencies depending mainly on the degrees of freedom of thestructure one of the most important frequencies is thenatural or fundamental frequency [33] To analyze fre-quencies is necessary to change the analysis from time tofrequency domain generally using the Fast Fourier Trans-formation (FFT) In general sense FFT is a computationalcomplexity improvement of DFT that separates a complexsignal into a set of sinusoidal ones where each signal has acorresponded frequency and potential Similarly there areother techniques such as periodogram It is well-documented that both FFT and periodogram algorithmscan be utilized for analyzing movements of bridge structureson frequency domain [34] Because FFT is a well-knownalgorithm its mathematical formulation is not describedhere On the other hand periodogram equation can beobserved in Equation (10) where x(t) represents the timedomain input signal X(f) is its Fourier transform and Xp(f )is the output periodogram which is calculated as the Fouriertransform of the autocorrelation function from x(t) It alsocan be alternatively calculated as the multiplication betweenX(f ) and its conjugated counterpart as [34]

Xp(f) F x(t)lowastx(minust) X(f) middot Xlowast(f) (10)

7 StructuralDescriptionof theAnalyzedBridge

+e analyzed structure in this paper using the fused smartsensor is the Juarez Bridge (see Figure 5) Such a structure islocated in Culiacan Mexico It is a reinforced concretebridge with an approximate length of 200meters +is is animportant civil infrastructure of the city Some importantfindings about the structural performance of this bridge weredocumented earlier by Vazquez-Becerra et al [26] It wasreported that the bridge presented deformations in averageof about 8 cm in the vertical direction representing anundesirable structural behavior In the above researchmaximum displacements were recorded at the ZEN2 station(see Figure 5(b)) which was reported to be the most affectedlocation over the bridge mainly because several vehicles arestopped by a traffic light close to it Considering displace-ments with respect to the vertical direction the calculatedaverage of probability of failure based on comparingsemistatic displacement with AASHTO (American Associ-ation of State Highway and Transportation Officials) de-flection limits was about 40 In addition according toVazquez-Becerra et al [26] the Juarez Bridgersquos probability offailure and its maximum displacements were calculated bythe average displacements extracted via GPS receivers lo-cated at critical points of the bridge However they only usedGPS receivers to extract the performance of the structure

6 Advances in Civil Engineering

under normal loading conditions On the other hand for theinvestigation presented in this paper the GPS performancewas improved by adding acceleration data using commercialaccelerometer and smartphone respectively +e combi-nation between GPS and commercial accelerometers hasbeen developed and reported in the literature [2 14 23]However very few information is documented in the lit-erature about using sensors embedded in smartphones torecord acceleration of real-scale bridges Consequently ifresults obtained via smartphone for bridge structures areviable this will provide a great advancement in SHM ofbridges because price and time would be drastically reducedin extracting their structural performance in real time

8 Measurement Campaign Using FusedSmart Sensor

For this study the measurement campaign was developed atthe ZEN2 station (see Figure 5(b)) because it was previouslydocumented by Vazquez-Becerra et al [26] that this locationpresented the worst performance with respect to averagedisplacements and probability of failure Additionally thesemeasurements were developed on Monday since this daypresented the highest probability of inappropriate structuralbehavior +e structural assessment of the bridge was de-veloped simultaneously in terms of the smart sensor in-tegrating GPS receivers accelerometer and smartphone+e process is comprehensively described in this section

81 GPS Receivers Campaign +e GPS measurementmethodology employed was the kinematic relative posi-tioning which involves two stations measuring simulta-neously one inmovement (the one located on the bridge) andthe other one static (outside of the bridge as static reference)[10] +e stable-based reference station used is the 373Hstation this station is located near to the bridge It is part ofthe active national geodetic network from the Mexican Na-tional Institute of Statistics and Geography (INEGI inSpanish) Figure 6(a) shows the position of station 373H withrespect to the ZEN2 station +e length of the baseline(from ZEN2 to 373H) is approximately 120meters +e

displacement measurement lasted 500 seconds using asampling rate of 100Hz when the Juarez Bridge was subjectedto in-service traffic loading conditions Figures 6(b) and 6(c)illustrate the ZEN2 and 373H stations respectively+ey weremonitored during the GPS displacements campaign

Previously to postprocessing the collected data raw GPSdata was debugged by Teqc (translation editing and qualitycheck) [35] +e preprocessing consists in insulate GPS dataand perform a quality check+e latest step is very interestingconsidering that it provides an understanding of the condi-tion of the GPS receiverrsquos position in terms of Multipath(MP) visible satellites and signal to noise ratio (S) For thiscase the results demonstrate that the rover reports multipathvalues of MP1RMS 00478m MP2RMS 00119m 19 GPSsatellites with observations S1 701 and S2 449 for thereference station MP1RMS 00428m MP2RMS 00114S1 712 and S2 449 the values of both receivers are verysimilar due to the short distance between them On the otherhand because relative positioning with double differencing(DD) over a short baseline (less than 10 kilometers of length)was used some error sources were either reduced or droppedfor example clock bias from the satellite-receiver and at-mospheric effects (ionospheric and tropospheric bias) be-cause both receivers are generally affected by the sameconditions [36] Additionally this positioning techniquemakes possible to achieve a good performance using com-mercial software for processing raw GPS data which can beseen in most of investigations related to the use of GPS re-ceivers in SHM [7 21 28 31 32] However due to the ex-cessive amount of data the software considered to process allthe information is the GAMITGLOBK TRACK developed bythe MIT (Massachusetts Institute of Technology) +is soft-ware processes GPS data in kinematic mode +e mainspecifications of GPS data processing included 15deg cut-offangle precise final orbits disseminated by IGS (InternationalGNSS Service) mode air (used for high sample) NationalGeodetic survey antenna calibration parameters and others

82 Accelerometer Campaign As previously discussedduring the 500 seconds measurement session one com-mercial accelerometer was utilized +is device was

(a) (b)

Figure 5 Juarez Bridge (a) lateral view (b) ZEN2 GPS location

Advances in Civil Engineering 7

recording acceleration at a rate of 100Hz+e accelerometerwas perfectly fixed to present the same physical center as theother devices to record the behavior of ZEN2 stationFigure 6(b) illustrates the smart sensor that is holding thethree devices One more important fact in the accelerometercampaign is that information was recorded at the same timethan GPS and smartphone providing a correct synchroni-zation [37]

83 SmartphoneCampaign As illustrated earlier in Figure 4the smartphone was fixed at the lowest plate of the mechanicstructure for the smart sensor +e measured informationfrom the accelerometer embedded in the smartphone wascollected by an open access App [24] +is application canrecord acceleration and then send it to the user in the form oftext file+e smartphone collected data at the same time thanthe GPS receiver and commercial accelerometer during500 seconds with a sampling rate of 400Hz

84 Flow Chart Diagram of the Measurement Process Insummary the data acquisition process of the bridge usingthe smart sensor is illustrated in Figure 7 Firstly theaccelerometer smartphone and GPS antenna are perfectlyfixed to the location of ZEN2 station +e particularissues related to this location were previously discussedSecondly acceleration time histories are extracted fromaccelerometer and smartphone respectively and dis-placements from GPS receivers As documented earlier thedata acquisition process is performed simultaneously usingthe three instruments +irdly the apparent displacementin three directions (N-S E-W and vertical) is calculatedusing GPS +en semistatic and dynamic displacementsare extracted using low- and high-pass filters respectivelyFourthly in order to study the dynamic characteristics ofthe Juarez Bridge FFTand periodogram are plotted for thelongitudinal transversal and vertical directions using rawacceleration of smartphone and accelerometer re-spectively Finally to compare the extracting responsecapabilities of the three instruments dynamic displace-ments reported by every device are used for the calculationof periodograms +e above discussion is clarified in thenext section

9 Structural Response Collected by theSmart Sensor

+e structural performance of the Juarez Bridge is studiedunder normal service loading conditions in real time ie thestructure is evaluated during a loading condition that is ex-perienced daily by users In fact a nonpermissible vibration inthe vertical direction has been reported by users and it hasbeen documented in the literature by Vazquez-Becerra et al[26] To verify the undesirable behavior of this bridge in thisstudy the authors decided to implement other devices duringthe SHM Hence the collected structural response is obtainedsimultaneously using GPS accelerometer and smartphone inreal time as a fused smart sensor

91 Acceleration TimeHistory One measurement session of500 seconds was performed using the three instrumentsunder consideration In terms of collected accelerationFigure 8 illustrates the recorded response of the smartphoneand accelerometer respectively Figures 8(a)ndash8(c) illustratethe longitudinal transversal and vertical acceleration re-spectively Several observations can be made correspondingto Figure 8 It can be observed that acceleration in thevertical direction is greater than longitudinal and transversaldirections In terms of displacements the same observationwas reported by [26] In addition it can be noted inFigure 8(a) that in the longitudinal direction the acceler-ation reported by smartphone and accelerometer is pre-senting some discrepancies +is issue may be justifiedbecause of the precision of commercial accelerometers insuch a direction However this structural response does notrepresent a threat of safety because of the stiffness of thebridge in the longitudinal direction In addition it can beobserved in Figure 8(c) that the critical component is as-sociated to accelerations acting in the vertical directionConversely in terms of accuracy in Figures 8(b) and 8(c) isshown that acceleration for both smartphone and acceler-ometer is matching very closely In general terms Figure 8demonstrates that acceleration recorded by smartphone andaccelerometer can be positively comparable

92 Semistatic Displacements Using GPS Considering dis-placements obtained via GPS semistatic displacements are

(a) (b) (c)

Figure 6 GPS campaign (a) location of ZEN2 and 373H (b) station ZEN2 (c) station 373H

8 Advances in Civil Engineering

extracted by applying a low-pass 13lter Semistatic dis-placements corresponding to the bridge under considerationrepresent slowmovements of the structure they are reportedin the N-S E-W and vertical direction in Figures 9(a)ndash9(c)respectively

It can be observed in Figures 9(a) and 9(b) that smallerdisplacements are present in both N-S and E-W horizontaldirections than those in the vertical one is is illustrated inFigure 9(c) it is noted that considerable great displacementsare reported in the vertical direction Hence in terms ofsemistatic displacements it can be justi13ed that the verticalcomponent is presenting critical behavior ere exist casesof bridge studies with longer dimensions such as the SanFrancisco-Oakland bay bridge that presented small dis-placements [38] However the analyzed bridge presented inthis paper is a very particular case of study whichreports unusual behavior with respect to the vertical di-rection comparing semistatic displacement with AASHTO

deection limits atrsquos why the authors of this paper de-cided to use it for the evaluation of the sensing capabilities ofthe proposed smart sensor

93 Dynamic Displacements Using GPS Movements relatedto short period component are represented by dynamicdisplacements (see Figure 10)

Dynamic displacements are obtained applying a high-pass 13lter to GPS apparent displacements Figure 10 illus-trates dynamic displacements corresponding to the bridge Itcan be noted in Figures 10(a) and 10(b) that dynamicdisplacements are relatively small for both N-S and E-Whorizontal components On the contrary the vertical di-rection is presenting excessive displacements and this isillustrated in Figure 10(c) One more time critical dis-placements are appearing for the vertical component but inthis case with respect to dynamic displacements

Longitudinal direction

100 200 300 400 5000Time (sec)

Smartphone Accelerometer

ndash1

0

1

Acc

eler

atio

n (m

s2 )

(a)

Transversal directionSmartphone Accelerometer

ndash1

0

1A

ccel

erat

ion

(ms

2 )

100 200 300 400 5000

Time (sec)

(b)

Vertical directionSmartphone Accelerometer

100 200 300 400 5000

Time (sec)

ndash2

0

2

Acc

eler

atio

n (m

s2 )

(c)

Figure 8 Acceleration time history

Accelerometer

Acceleration

High-pass filterFFTPeriodogram

Detrend

t

0

Ri ndash (1n)sumni=1 Ri

t

0

Dynamicdisplacement

Semistaticdisplacement

Low-pass filterBand-Pass filter

Smartphone Displacementfrom GPS

Figure 7 Flow chart of measurement process

Advances in Civil Engineering 9

Apparent displacementSemistatic displacement

ndash001

0

001D

ispla

cem

ent (

m)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(a)

Apparent displacementSemistatic displacement

ndash005

0

005

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(b)

Apparent displacementSemistatic displacement

ndash001

0

001

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(c)

Figure 9 Semistatic displacements in three directions (a) N-S (b) E-W and (c) vertical

NndashS

ndash001

0

001

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(a)

EndashW

ndash005

0

005

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(b)

Figure 10 Continued

10 Advances in Civil Engineering

94 FFT and Periodogram Using GPS DynamicDisplacements To identify structural dynamic character-istics of the structure under consideration FFT and perio-dograms are calculated and plotted in frequency domainusing dynamic displacements of GPS

Figure 11 illustrates FFT and periodogram in the threeGPS directions N-S (Figures 11(a) and 11(b)) E-W(Figures 11(c) and 11(d)) and vertical (Figures 11(e) and11(f)) To clarify the results they are normalized with respectto their maximum amplitude value

95 Acceleration Time History Using Accelerometer andSmartphone Acceleration time histories corresponding tothe longitudinal transversal and vertical direction of thebridge were earlier presented in Figure 8 considering bothsmartphone and accelerometer In the longitudinal direction(see Figure 8(a)) it can be noted that there are some dif-ferences in the acceleration reported by smartphone andaccelerometer such discrepancies were justi13ed earlier inSection 91 In the transversal (see Figure 8(b)) and verticaldirections (see Figure 8(c)) it is observed that smartphone

Vertical

ndash001

0

001D

ispla

cem

ent (

m)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(c)

Figure 10 Dynamic displacements in three directions (a) N-S (b) E-W and (c) vertical

NndashSFFT

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(a)

NndashSPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(b)

EndashWFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(c)

EndashWPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

VerticalFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(e)

VerticalPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 11 Dynamic frequencies in terms of normalized FFT and periodogram

Advances in Civil Engineering 11

and accelerometer are recording very similar results In timedomain these results are very promising However tocomplement them results are changed to the frequencydomain in terms of FFT and periodogram is is discussedin the next section

96 FFT and Periodogram Using Accelerometer and Smart-phone Accelerations In this section results are studied infrequency domain ey are presented in terms of FFT andperiodogram (see Figure 12)

Figure 12 summarizes the results related to FFT andperiodogram corresponding to longitudinal transversal andvertical directions It can be observed in Figure 12 that thefrequency that contributes the most to the structural re-sponse may be located approximately at 9Hz In addition itis noted that FFT in the longitudinal and transversal di-rection (see Figures 12(a) and 12(c)) is presenting the 13rstfrequency of interest at about 35Hz Based on the resultsplotted in Figure 12 it can be stated that both smartphoneand accelerometer are providing very accurate results with

respect to each other Hence it may result practical to usesmartphones for extracting dynamic characteristics of real-scale bridges

97 Comparison of Periodogram of GPS Accelerometers andSmartphones Using Dynamic Displacements To comparedata acquisition capabilities of smartphones accelerometersand GPS the periodogram of them with respect to dynamicdisplacements in the vertical direction is illustrated inFigure 13 Only vertical direction is studied because it wasdemonstrated that this direction presented a critical be-havior It is noted as reported in Figure 12 that the fre-quency that contributes the most to the structural responseis located approximately at 937Hz which is related toprevious acceleration spectra is frequency is well detectedby the three instruments Periodogram reported in Figure 13is plotted considering dynamic displacements of smart-phone accelerometer and GPS Hence the accelerationrecorded by the smartphone and accelerometer respectivelyis converted to dynamic displacement by carrying out a

SmartphoneAccelerometer

FFTmdashlongitudinal direction

0

05

1A

mpl

itude

5 10 15 200Frequency (Hz)

(a)

SmartphoneAccelerometer

Periodogrammdashlongitudinal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(b)

SmartphoneAccelerometer

FFTmdashtransversal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(c)

SmartphoneAccelerometer

Periodogrammdashtransversal direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

SmartphoneAccelerometer

FFTmdashvertical direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(e)

SmartphoneAccelerometer

Periodogrammdashvertical direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 12 Dynamic frequencies in terms of FFT and periodogram

12 Advances in Civil Engineering

double integration which may involve some discrepancies inobtaining displacements In addition to numerical in-tegration some of the dierences observed in Figure 13 maybe associated with the applied 13ltering techniques to obtaindynamic displacements On the other hand GPS reportsdisplacements directly improving considerably data pro-cessing and interpretations However it is considered im-portant to analyze the three periodograms from Figure 13 tohave a better understanding about the coincidences of somefrequencies that appear in two or more periodograms and itmay be useful to have a better idea about de dynamic be-havior of a bridge

10 Conclusions

An alternative methodology to perform SHM on real-scalebridges is presented in this paper using GPS an acceler-ometer and a smartphone integrated in a fused smartsensor Such sensor is validated for the SHM of complexstructures as bridges under normal serviceability conditionsin real time A reinforced concrete bridge was studied byusing the proposed smart sensor because it presented ex-cessive vibration problems e accuracy of this smart sensorwas demonstrated comparing the gathered data from GPSaccelerometers and smartphones in both time and fre-quency domain Semistatic and dynamic displacements aswell as accelerations demonstrated that there are criticalperformance problems on the analyzed bridge particularlyin the vertical direction Since this critical behavior wasreported by other scholars that performed the SHM withdierent devices it can be stated that this smart sensor canbe implemented for SHM on bridges in real time In ad-dition it was demonstrated that the proposed smart sensor isadequate to perform SHM in real-scale bridges Conse-quently for the bridge under consideration it was calculated

that the fundamental frequency in the horizontal direction islocated at 9Hz Furthermore it was detected the 13rst fre-quency of interest around 35Hz In order to improvefrequency domain analysis the proposed smart sensormethodology performs a periodogram algorithm over thevertical direction data to obtain a more accurate SHM It wasobserved that there is a dierence between GPS displace-ment data and accelerometer-smartphone based data It canbe justi13ed due to the nature of each measurement meth-odology where GPS base data comes from atmosphere andaccelerometer-based data comes from a vibration sensorBecause of this the noise nature of the aforementionedmeasurement devices is not the same due to its dierentmeasurement principle On the other hand it was observed acoincidence of the three instruments approximately at937Hz for the vertical component us it can explain theusefulness of fusing GPS accelerometer and smartphoneinformation in order to discriminate undesired noise thatcan be generated by atmosphere disturbances of doubleintegration error from acceleration data when it appearsonly in one device data Finally it can be stated that thispaper contributed in understanding how GPS accelerom-eters and smartphones can be successfully fused in a smartsensor to perform SHM on real-scale bridges

Data Availability

e data used to support the 13ndings of this study areavailable from the corresponding author upon request

Disclosure

Any opinions 13ndings conclusions or recommendationsexpressed in this paper are those of the authors and do notnecessarily reect the views of the sponsors

Conflicts of Interest

e authors do not have any conicts of interest to declareabout this research

Acknowledgments

e authors would like to thank Septentrio Inc for pro-viding the AsteRx-U Receiver and the PolaNt-x MF antennaused during the GPS data collection is study is based onwork partly supported by the Consejo Nacional de Ciencia yTecnologıa (CONACYT-Mexico) as scholarships for the 13rstand seventh author respectively In addition the work isalso supported by the Universidad Autonoma de Sinaloa(UAS)

References

[1] V Ashkenazi and G W Roberts ldquoExperimental monitoringof the Humber bridge using GPSrdquo Proceedings of the In-stitution of Civil Engineers-Civil Engineering vol 120 no 4pp 177ndash182 1997

[2] W S Chan Y L Xu X L Ding andW J Dai ldquoAn integratedGPSndashaccelerometer data processing technique for structural

1

09

08

07

06

05

04

03

02

01

00 5 10

Frequency (Hz)15 20

SmartphoneAccelerometerGPS (vertical)

Figure 13 Dynamic frequencies in terms of periodogram (verticaldirection)

Advances in Civil Engineering 13

deformation monitoringrdquo Journal of Geodesy vol 80 no 12pp 705ndash719 2006

[3] S B Im S Hurlebaus and Y J Kang ldquoSummary review ofGPS technology for structural health monitoringrdquo Journal ofStructural Engineering vol 139 no 10 pp 1653ndash1664 2011

[4] M R Kaloop and H Li ldquoMonitoring of bridge deformationusing GPS techniquerdquo KSCE Journal of Civil Engineeringvol 13 no 6 pp 423ndash431 2009

[5] M Kaloop E Elbeltagi J Hu and A Elrefai ldquoRecent advancesof structures monitoring and evaluation using GPS-time seriesmonitoring systems a reviewrdquo ISPRS International Journal ofGeo-Information vol 6 no 12 p 382 2017

[6] X Li C Rizos L Ge Y Tamura and A Yoshida ldquo+ecomplementary characteristics of GPS and accelerometer inmonitoring structural deformationrdquo in Proceedings of the ION2005 National Technical Meeting Institute of Navigation(ION) San Diego CA USA January 2005

[7] P Psimoulis S Pytharouli D Karambalis and S StirosldquoPotential of Global Positioning System (GPS) to measurefrequencies of oscillations of engineering structuresrdquo Journalof Sound and Vibration vol 318 no 3 pp 606ndash623 2008

[8] C Watson T Watson and R Coleman ldquoStructural moni-toring of cable-stayed bridge analysis of GPS versus modeleddeflectionsrdquo Journal of Surveying Engineering vol 133 no 1pp 23ndash28 2007

[9] C Xiong and Y Niu ldquoInvestigation of the dynamic behavior ofa super high-rise structure using RTK-GNSS techniquerdquo KSCEJournal of Civil Engineering vol 23 no 2 pp 654ndash665 2018

[10] B Hofmann-Wellenhof H Lichtenegger and E WasleGNSSndashGlobal Navigation Satellite Systems GPS GLONASSGalileo and More Springer Science amp Business MediaVienna Austria 2007

[11] M Celebi ldquoGPS in dynamic monitoring of long-periodstructuresrdquo Soil Dynamics and Earthquake Engineeringvol 20 no 5ndash8 pp 477ndash483 2000

[12] D K Kim J I Kim and M Q Feng ldquoInstrumentation ofbridges for structural health monitoringrdquo KSCE Journal ofCivil Engineering vol 5 no 3 pp 231ndash242 2001

[13] G Q Wang D M Boore H Igel and X Y Zhou ldquoSomeobservations on colocated and closely spaced strong ground-motion records of the 1999 Chi-Chi Taiwan earthquakerdquoBulletin of the Seismological Society of America vol 93 no 2pp 674ndash693 2003

[14] J Hwang H Yun S K Park D Lee and S Hong ldquoOptimalmethods of RTK-GPSaccelerometer integration to monitorthe displacement of structuresrdquo Sensors vol 12 no 1pp 1014ndash1034 2012

[15] S Dashti J D Bray J Reilly S Glaser A Bayen and E MarildquoEvaluating the reliability of phones as seismic monitoringinstrumentsrdquo Earthquake Spectra vol 30 no 2 pp 721ndash7422014

[16] M Feng Y Fukuda M Mizuta and E Ozer ldquoCitizen sensorsfor SHM use of accelerometer data from smartphonesrdquoSensors vol 15 no 2 pp 2980ndash2998 2015

[17] X Zhao R Han Y Ding et al ldquoPortable and convenient cableforce measurement using smartphonerdquo Journal of CivilStructural Health Monitoring vol 5 no 4 pp 481ndash491 2015

[18] X Zhao K Ri R Han Y Yu M Li and J Ou ldquoExperimentalresearch on quick structural health monitoring technique forbridges using smartphonerdquoAdvances inMaterials Science andEngineering vol 2016 no 1 pp 1ndash14 2016

[19] Y Ding R Han H Liu S Li X Zhao and Y Yu ldquoBridgeinspection and management system based on smart phonerdquoin Proceedings of the ASME 2016 Conference on Smart

Materials Adaptive Structures and Intelligent SystemsAmerican Society of Mechanical Engineers Stowe VT USASeptember 2016

[20] X Zhao R Han K J Loh B Xie J Li and J Ou ldquoShakingtable tests for evaluating the damage features under earth-quake excitations using smartphonesrdquo in Proceedings of theHealth Monitoring of Structural and Biological Systems XIIInternational Society for Optics and Photonics Denver COUSA May 2018

[21] F Moschas and S Stiros ldquoMeasurement of the dynamicdisplacements and of the modal frequencies of a short-spanpedestrian bridge using GPS and an accelerometerrdquo Engi-neering Structures vol 33 no 1 pp 10ndash17 2011

[22] M R Kaloop ldquoBridge safety monitoring based-GPS tech-nique case study Zhujiang Huangpu bridgerdquo Smart Struc-tures and Systems vol 9 no 6 pp 473ndash487 2012

[23] G W Roberts X Meng and A H Dodson ldquoIntegrating aglobal positioning system and accelerometers to monitor thedeflection of bridgesrdquo Journal of Surveying Engineeringvol 130 no 2 pp 65ndash72 2004

[24] Accelerometer Analyzer App 2017 httpsplaygooglecomstoreappsdetailsidcomlulaccelerometeramphles

[25] X Meng ldquoReal-time deformationmonitoring of bridges usingGPSaccelerometersrdquo Doctoral dissertation University ofNottingham UK 2002

[26] G E Vazquez-Becerra J R Gaxiola-Camacho R BennettG M Guzman-Acevedo and I E Gaxiola-CamacholdquoStructural evaluation of dynamic and semi-static displace-ments of the Juarez bridge using GPS technologyrdquo Mea-surement vol 110 no 1 pp 146ndash153 2017

[27] P Psimoulis I Peppa L Bonenberg S Ince and X MengldquoCombination of GPS and RTS measurements for themonitoring of semi-static and dynamic motion of pedestrianbridgerdquo in Proceedings of the 3rd Joint International Sym-posium on Deformation Monitoring Vienna Austria April2016

[28] F Moschas and S Stiros ldquoNoise characteristics of high-frequencyshort-duration GPS records from analysis ofidentical collocated instrumentsrdquo Measurement vol 46no 4 pp 1488ndash1506 2013

[29] S W Smithgte Scientist amp Engineerrsquos Guide to Digital SignalProcessing California Technical Publishing San Diego CAUSA 1997

[30] J Wang X Meng C Qin and J Yi ldquoVibration frequenciesextraction of the forth road bridge using high sampling GPSdatardquo Shock and Vibration vol 2016 Article ID 980786118 pages 2016

[31] J Yu XMeng X Shao B Yan and L Yang ldquoIdentification ofdynamic displacements and modal frequencies of a medium-span suspension bridge using multimode GNSS processingrdquoEngineering Structures vol 81 no 1 pp 432ndash443 2014

[32] E Elbeltagi M Kaloop and M Elnabwy ldquoMonitoring andassessment of Mansoura railway steel bridge using RTK-GPStechniquerdquo in Proceedings of the 2nd Joint InternationalSymposium on Deformation Monitoring pp 9ndash11 Notting-ham England September 2013

[33] N Raziq ldquoGPS structural deformation monitoring the mid-height problemrdquo Doctoral dissertation University ofMelbourneAustralia 2008

[34] S S HaykinAdaptive Filtergteory Pearson Education India5th edition 2014

[35] L H Estey and C M Meertens ldquoTEQC the multi-purposetoolkit for GPSGLONASS datardquo GPS Solutions vol 3 no 1pp 42ndash49 1999

14 Advances in Civil Engineering

31 Semistatic Displacements Using Moving Average FilterApparent displacement is formed by short- and long-periodcomponents +e long component has a slow movementlong wavelength and low frequency It is produced by someeffects such as foundation settlement deck creep thermalexpansioncontraction and loss of prestress [25] It is re-ported in the literature that long-period component ex-presses the semistatic displacement [27] which can be usedto evaluate the movement of the bridge [22] It can becalculated by applying the moving average filter with awindow of 41 samples [26] One problem that must be facedfor determining semistatic displacement from GPS data isthat long-period component presents colored noise (chaoticbehavior) produced by environmental bias (multipathionospheric and tropospheric delay) and instrumental noise(noise produced by each GPS receiver) [28] +erefore it isimportant to take care about how the filter is used +emoving average filter presents a good performance in timedomain Consequently it can be used for smoothing timeseries It takes a specific number of elements from the timeseries based on the windowrsquos size beginning in the firstelement and their average is calculated +en it takes thesame number of elements but starting in the second elementfor calculating their average and so on +e moving averagefilter is given as [29]

Y[i] 1

M1113944

Mminus1

j0x[i + j] (3)

where x is the input signal Y is the output signal and M isthe size of window

32 Dynamic Displacements Using Chebyshev Filter Tocalculate the short period component representing the dy-namic displacement of the structure under consideration ahigh-pass filter should be applied to the apparent dis-placement time series +e short period component is a fastmovement with high frequency and short wavelength +ismovement is generally produced by traffic people earth-quake activity and so on [25] To clarify this concept the

measured displacement of a bridge using GPS can beexpressed as [30 31]

Y(n) M(n) + D(n) + N(n) (4)

where Y(n) is the apparent displacement M(n) representsthe long period interferences (environment delays) D(n) isthe dynamic displacement of the bridge and N(n) is thenoise

It is documented in the literature that dynamic dis-placements can be used to understand the structural per-formance of bridges [32] +e filter used in this article toextract dynamic displacements is the Chebyshev type I andit has been tested by some scholars [9 21 26 31] Chebyshevtype I filter can be expressed as [30]

Gn(w) Hn(jw)1113868111386811138681113868

1113868111386811138681113868 1

1 + e2 + T2n ww0( 1113857

1113969(5)

where |Hn(jw)| is the transfer function with an absolutevalue equal to the gains as an angular function of frequencyw of order n e is the ripple factor T is the Chebyshevpolynomial order and w0 is the cut-off frequency +is filteris characterized by having the ripple in the passband

4 Calculation of Displacements UsingAcceleration Records

+ere are alternative methodologies in the literature forcalculating displacement using acceleration records [6 14]that is to obtain displacements from acceleration time seriesa double integration of them is required However theproblem is how to select the method that provides betterresults or the one losing less information ie the mostefficient and accurate Obtaining displacement from accel-eration records is a difficult process It introduces biascreating a difference from decimeter to even meter levelbetween direct and indirect results [13] +is process isdescribed in terms of four steps according to Feng et al [16]Firstly a band pass filter needs to be implemented to removenoise Secondly a double integration is applied as

s(t) s0 + v0xt + 1113946t

01113946

t

0a(t)dt1113888 1113889dt (6)

where s(t) is the displacement a is the acceleration s0 is theinitial position and v0 is the initial velocity

+irdly the initial values of position and velocity are setequal to zero Finally the linear trend must be removed [14]Hence the results of this process represent the dynamicdisplacement Unfortunately because the velocity and po-sition are supposed to be zero the semistatic and staticdisplacements are lost [6] In this process is essential to use acorrect band pass filter because noise is removed per it Forthis purpose there are several alternatives For example itcan be used the Discrete Fourier Transform (DFT) a FiniteImpulse Response (FIR) filter or an Empirical Model De-composition (EMD) filter +e first of them (DFT filter)provides a no complex structure and balance [14] On theother hand EMD filter generates a set of N functions from

Figure 4 Smart sensor

Advances in Civil Engineering 5

the input signal +e resulted set of functions from EMD isdefined as [14]

a(t) 1113944N

n1Cn(t) + rn(t) (7)

where a(t) is the input signal rn(t) is the residue and Cn(t)

is the nth function which is named intrinsic mode functions(IMF) +e EMD filter is expressed as [14]

ab(t) 1113944r

nq

Cn(t) (8)

where

1lt qlt rlt n (9)

Hence the above expressions can be used to calculatedynamic displacements from acceleration time historiesrecorded by smartphone and accelerometer respectively

5 Application of GPS ReceiversAccelerometers and Smartphones in SHM

As previously documented several are the advantages anddisadvantages of GPS receivers accelerometers and smart-phones in SHM Both GPS receivers and commercial accel-erometers have limitations that can be addressed using certainmethodologies For example integrating GPS and commer-cial accelerometer the combined result may recover lost dataeg missing measurements from GPS +is integration maysolve the lack of semistatic and static components +e re-dundancy of both systems (GPS and accelerometers) providesa good performance in SHM [6] In general it is reported inthe literature that high-frequency displacements may beobtained from accelerometers and low frequency ones fromGPS +e reason of considering low frequencies from GPSand high ones from accelerometers is that some scholars [14]demonstrated that GPS is a helpful tool for measuring staticand semistatic behaviors (low-frequency displacements) butit experiences problems related to collecting data of dynamicdisplacements from structures vibrating with high frequen-cies On the other hand accelerometer records dynamicresponses very accurately without low-frequency compo-nents +erefore it is reported in the literature that accel-erometers and GPS make a reasonable complement to eachother [2 6 14] Unfortunately GPS receivers and acceler-ometers are expensive instruments Commonly the accel-eration is recorded using commercial accelerometersHowever the improved quality of acceleration sensors in-tegrated in smartphones provides a feasible alternativeGenerally their level of accuracy may be lower than acommercial accelerometer but it is reported in the literaturethat their results are very similar between them for smallstructural systems [15 16] +e contribution of this article tothe literature is that accelerometer sensors of smartphones areused together with other devices to form a smart sensor to beused in the SHM of a reinforced concrete bridge in real time+en a parametric analysis is performed with respect to GPSreceivers and commercial accelerometers

6 Dynamic Characteristics of Structures

Before proceeding further with the discussion dynamiccharacteristic of structures must be introduced Structuresare designed to resist different types of loading conditions+ey must have the capacity to reliably transmit loadingsfrom the applied point to their foundations when this ishappening movements may be produced in structures as aresponse Movements are generated following several fre-quencies depending mainly on the degrees of freedom of thestructure one of the most important frequencies is thenatural or fundamental frequency [33] To analyze fre-quencies is necessary to change the analysis from time tofrequency domain generally using the Fast Fourier Trans-formation (FFT) In general sense FFT is a computationalcomplexity improvement of DFT that separates a complexsignal into a set of sinusoidal ones where each signal has acorresponded frequency and potential Similarly there areother techniques such as periodogram It is well-documented that both FFT and periodogram algorithmscan be utilized for analyzing movements of bridge structureson frequency domain [34] Because FFT is a well-knownalgorithm its mathematical formulation is not describedhere On the other hand periodogram equation can beobserved in Equation (10) where x(t) represents the timedomain input signal X(f) is its Fourier transform and Xp(f )is the output periodogram which is calculated as the Fouriertransform of the autocorrelation function from x(t) It alsocan be alternatively calculated as the multiplication betweenX(f ) and its conjugated counterpart as [34]

Xp(f) F x(t)lowastx(minust) X(f) middot Xlowast(f) (10)

7 StructuralDescriptionof theAnalyzedBridge

+e analyzed structure in this paper using the fused smartsensor is the Juarez Bridge (see Figure 5) Such a structure islocated in Culiacan Mexico It is a reinforced concretebridge with an approximate length of 200meters +is is animportant civil infrastructure of the city Some importantfindings about the structural performance of this bridge weredocumented earlier by Vazquez-Becerra et al [26] It wasreported that the bridge presented deformations in averageof about 8 cm in the vertical direction representing anundesirable structural behavior In the above researchmaximum displacements were recorded at the ZEN2 station(see Figure 5(b)) which was reported to be the most affectedlocation over the bridge mainly because several vehicles arestopped by a traffic light close to it Considering displace-ments with respect to the vertical direction the calculatedaverage of probability of failure based on comparingsemistatic displacement with AASHTO (American Associ-ation of State Highway and Transportation Officials) de-flection limits was about 40 In addition according toVazquez-Becerra et al [26] the Juarez Bridgersquos probability offailure and its maximum displacements were calculated bythe average displacements extracted via GPS receivers lo-cated at critical points of the bridge However they only usedGPS receivers to extract the performance of the structure

6 Advances in Civil Engineering

under normal loading conditions On the other hand for theinvestigation presented in this paper the GPS performancewas improved by adding acceleration data using commercialaccelerometer and smartphone respectively +e combi-nation between GPS and commercial accelerometers hasbeen developed and reported in the literature [2 14 23]However very few information is documented in the lit-erature about using sensors embedded in smartphones torecord acceleration of real-scale bridges Consequently ifresults obtained via smartphone for bridge structures areviable this will provide a great advancement in SHM ofbridges because price and time would be drastically reducedin extracting their structural performance in real time

8 Measurement Campaign Using FusedSmart Sensor

For this study the measurement campaign was developed atthe ZEN2 station (see Figure 5(b)) because it was previouslydocumented by Vazquez-Becerra et al [26] that this locationpresented the worst performance with respect to averagedisplacements and probability of failure Additionally thesemeasurements were developed on Monday since this daypresented the highest probability of inappropriate structuralbehavior +e structural assessment of the bridge was de-veloped simultaneously in terms of the smart sensor in-tegrating GPS receivers accelerometer and smartphone+e process is comprehensively described in this section

81 GPS Receivers Campaign +e GPS measurementmethodology employed was the kinematic relative posi-tioning which involves two stations measuring simulta-neously one inmovement (the one located on the bridge) andthe other one static (outside of the bridge as static reference)[10] +e stable-based reference station used is the 373Hstation this station is located near to the bridge It is part ofthe active national geodetic network from the Mexican Na-tional Institute of Statistics and Geography (INEGI inSpanish) Figure 6(a) shows the position of station 373H withrespect to the ZEN2 station +e length of the baseline(from ZEN2 to 373H) is approximately 120meters +e

displacement measurement lasted 500 seconds using asampling rate of 100Hz when the Juarez Bridge was subjectedto in-service traffic loading conditions Figures 6(b) and 6(c)illustrate the ZEN2 and 373H stations respectively+ey weremonitored during the GPS displacements campaign

Previously to postprocessing the collected data raw GPSdata was debugged by Teqc (translation editing and qualitycheck) [35] +e preprocessing consists in insulate GPS dataand perform a quality check+e latest step is very interestingconsidering that it provides an understanding of the condi-tion of the GPS receiverrsquos position in terms of Multipath(MP) visible satellites and signal to noise ratio (S) For thiscase the results demonstrate that the rover reports multipathvalues of MP1RMS 00478m MP2RMS 00119m 19 GPSsatellites with observations S1 701 and S2 449 for thereference station MP1RMS 00428m MP2RMS 00114S1 712 and S2 449 the values of both receivers are verysimilar due to the short distance between them On the otherhand because relative positioning with double differencing(DD) over a short baseline (less than 10 kilometers of length)was used some error sources were either reduced or droppedfor example clock bias from the satellite-receiver and at-mospheric effects (ionospheric and tropospheric bias) be-cause both receivers are generally affected by the sameconditions [36] Additionally this positioning techniquemakes possible to achieve a good performance using com-mercial software for processing raw GPS data which can beseen in most of investigations related to the use of GPS re-ceivers in SHM [7 21 28 31 32] However due to the ex-cessive amount of data the software considered to process allthe information is the GAMITGLOBK TRACK developed bythe MIT (Massachusetts Institute of Technology) +is soft-ware processes GPS data in kinematic mode +e mainspecifications of GPS data processing included 15deg cut-offangle precise final orbits disseminated by IGS (InternationalGNSS Service) mode air (used for high sample) NationalGeodetic survey antenna calibration parameters and others

82 Accelerometer Campaign As previously discussedduring the 500 seconds measurement session one com-mercial accelerometer was utilized +is device was

(a) (b)

Figure 5 Juarez Bridge (a) lateral view (b) ZEN2 GPS location

Advances in Civil Engineering 7

recording acceleration at a rate of 100Hz+e accelerometerwas perfectly fixed to present the same physical center as theother devices to record the behavior of ZEN2 stationFigure 6(b) illustrates the smart sensor that is holding thethree devices One more important fact in the accelerometercampaign is that information was recorded at the same timethan GPS and smartphone providing a correct synchroni-zation [37]

83 SmartphoneCampaign As illustrated earlier in Figure 4the smartphone was fixed at the lowest plate of the mechanicstructure for the smart sensor +e measured informationfrom the accelerometer embedded in the smartphone wascollected by an open access App [24] +is application canrecord acceleration and then send it to the user in the form oftext file+e smartphone collected data at the same time thanthe GPS receiver and commercial accelerometer during500 seconds with a sampling rate of 400Hz

84 Flow Chart Diagram of the Measurement Process Insummary the data acquisition process of the bridge usingthe smart sensor is illustrated in Figure 7 Firstly theaccelerometer smartphone and GPS antenna are perfectlyfixed to the location of ZEN2 station +e particularissues related to this location were previously discussedSecondly acceleration time histories are extracted fromaccelerometer and smartphone respectively and dis-placements from GPS receivers As documented earlier thedata acquisition process is performed simultaneously usingthe three instruments +irdly the apparent displacementin three directions (N-S E-W and vertical) is calculatedusing GPS +en semistatic and dynamic displacementsare extracted using low- and high-pass filters respectivelyFourthly in order to study the dynamic characteristics ofthe Juarez Bridge FFTand periodogram are plotted for thelongitudinal transversal and vertical directions using rawacceleration of smartphone and accelerometer re-spectively Finally to compare the extracting responsecapabilities of the three instruments dynamic displace-ments reported by every device are used for the calculationof periodograms +e above discussion is clarified in thenext section

9 Structural Response Collected by theSmart Sensor

+e structural performance of the Juarez Bridge is studiedunder normal service loading conditions in real time ie thestructure is evaluated during a loading condition that is ex-perienced daily by users In fact a nonpermissible vibration inthe vertical direction has been reported by users and it hasbeen documented in the literature by Vazquez-Becerra et al[26] To verify the undesirable behavior of this bridge in thisstudy the authors decided to implement other devices duringthe SHM Hence the collected structural response is obtainedsimultaneously using GPS accelerometer and smartphone inreal time as a fused smart sensor

91 Acceleration TimeHistory One measurement session of500 seconds was performed using the three instrumentsunder consideration In terms of collected accelerationFigure 8 illustrates the recorded response of the smartphoneand accelerometer respectively Figures 8(a)ndash8(c) illustratethe longitudinal transversal and vertical acceleration re-spectively Several observations can be made correspondingto Figure 8 It can be observed that acceleration in thevertical direction is greater than longitudinal and transversaldirections In terms of displacements the same observationwas reported by [26] In addition it can be noted inFigure 8(a) that in the longitudinal direction the acceler-ation reported by smartphone and accelerometer is pre-senting some discrepancies +is issue may be justifiedbecause of the precision of commercial accelerometers insuch a direction However this structural response does notrepresent a threat of safety because of the stiffness of thebridge in the longitudinal direction In addition it can beobserved in Figure 8(c) that the critical component is as-sociated to accelerations acting in the vertical directionConversely in terms of accuracy in Figures 8(b) and 8(c) isshown that acceleration for both smartphone and acceler-ometer is matching very closely In general terms Figure 8demonstrates that acceleration recorded by smartphone andaccelerometer can be positively comparable

92 Semistatic Displacements Using GPS Considering dis-placements obtained via GPS semistatic displacements are

(a) (b) (c)

Figure 6 GPS campaign (a) location of ZEN2 and 373H (b) station ZEN2 (c) station 373H

8 Advances in Civil Engineering

extracted by applying a low-pass 13lter Semistatic dis-placements corresponding to the bridge under considerationrepresent slowmovements of the structure they are reportedin the N-S E-W and vertical direction in Figures 9(a)ndash9(c)respectively

It can be observed in Figures 9(a) and 9(b) that smallerdisplacements are present in both N-S and E-W horizontaldirections than those in the vertical one is is illustrated inFigure 9(c) it is noted that considerable great displacementsare reported in the vertical direction Hence in terms ofsemistatic displacements it can be justi13ed that the verticalcomponent is presenting critical behavior ere exist casesof bridge studies with longer dimensions such as the SanFrancisco-Oakland bay bridge that presented small dis-placements [38] However the analyzed bridge presented inthis paper is a very particular case of study whichreports unusual behavior with respect to the vertical di-rection comparing semistatic displacement with AASHTO

deection limits atrsquos why the authors of this paper de-cided to use it for the evaluation of the sensing capabilities ofthe proposed smart sensor

93 Dynamic Displacements Using GPS Movements relatedto short period component are represented by dynamicdisplacements (see Figure 10)

Dynamic displacements are obtained applying a high-pass 13lter to GPS apparent displacements Figure 10 illus-trates dynamic displacements corresponding to the bridge Itcan be noted in Figures 10(a) and 10(b) that dynamicdisplacements are relatively small for both N-S and E-Whorizontal components On the contrary the vertical di-rection is presenting excessive displacements and this isillustrated in Figure 10(c) One more time critical dis-placements are appearing for the vertical component but inthis case with respect to dynamic displacements

Longitudinal direction

100 200 300 400 5000Time (sec)

Smartphone Accelerometer

ndash1

0

1

Acc

eler

atio

n (m

s2 )

(a)

Transversal directionSmartphone Accelerometer

ndash1

0

1A

ccel

erat

ion

(ms

2 )

100 200 300 400 5000

Time (sec)

(b)

Vertical directionSmartphone Accelerometer

100 200 300 400 5000

Time (sec)

ndash2

0

2

Acc

eler

atio

n (m

s2 )

(c)

Figure 8 Acceleration time history

Accelerometer

Acceleration

High-pass filterFFTPeriodogram

Detrend

t

0

Ri ndash (1n)sumni=1 Ri

t

0

Dynamicdisplacement

Semistaticdisplacement

Low-pass filterBand-Pass filter

Smartphone Displacementfrom GPS

Figure 7 Flow chart of measurement process

Advances in Civil Engineering 9

Apparent displacementSemistatic displacement

ndash001

0

001D

ispla

cem

ent (

m)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(a)

Apparent displacementSemistatic displacement

ndash005

0

005

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(b)

Apparent displacementSemistatic displacement

ndash001

0

001

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(c)

Figure 9 Semistatic displacements in three directions (a) N-S (b) E-W and (c) vertical

NndashS

ndash001

0

001

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(a)

EndashW

ndash005

0

005

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(b)

Figure 10 Continued

10 Advances in Civil Engineering

94 FFT and Periodogram Using GPS DynamicDisplacements To identify structural dynamic character-istics of the structure under consideration FFT and perio-dograms are calculated and plotted in frequency domainusing dynamic displacements of GPS

Figure 11 illustrates FFT and periodogram in the threeGPS directions N-S (Figures 11(a) and 11(b)) E-W(Figures 11(c) and 11(d)) and vertical (Figures 11(e) and11(f)) To clarify the results they are normalized with respectto their maximum amplitude value

95 Acceleration Time History Using Accelerometer andSmartphone Acceleration time histories corresponding tothe longitudinal transversal and vertical direction of thebridge were earlier presented in Figure 8 considering bothsmartphone and accelerometer In the longitudinal direction(see Figure 8(a)) it can be noted that there are some dif-ferences in the acceleration reported by smartphone andaccelerometer such discrepancies were justi13ed earlier inSection 91 In the transversal (see Figure 8(b)) and verticaldirections (see Figure 8(c)) it is observed that smartphone

Vertical

ndash001

0

001D

ispla

cem

ent (

m)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(c)

Figure 10 Dynamic displacements in three directions (a) N-S (b) E-W and (c) vertical

NndashSFFT

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(a)

NndashSPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(b)

EndashWFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(c)

EndashWPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

VerticalFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(e)

VerticalPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 11 Dynamic frequencies in terms of normalized FFT and periodogram

Advances in Civil Engineering 11

and accelerometer are recording very similar results In timedomain these results are very promising However tocomplement them results are changed to the frequencydomain in terms of FFT and periodogram is is discussedin the next section

96 FFT and Periodogram Using Accelerometer and Smart-phone Accelerations In this section results are studied infrequency domain ey are presented in terms of FFT andperiodogram (see Figure 12)

Figure 12 summarizes the results related to FFT andperiodogram corresponding to longitudinal transversal andvertical directions It can be observed in Figure 12 that thefrequency that contributes the most to the structural re-sponse may be located approximately at 9Hz In addition itis noted that FFT in the longitudinal and transversal di-rection (see Figures 12(a) and 12(c)) is presenting the 13rstfrequency of interest at about 35Hz Based on the resultsplotted in Figure 12 it can be stated that both smartphoneand accelerometer are providing very accurate results with

respect to each other Hence it may result practical to usesmartphones for extracting dynamic characteristics of real-scale bridges

97 Comparison of Periodogram of GPS Accelerometers andSmartphones Using Dynamic Displacements To comparedata acquisition capabilities of smartphones accelerometersand GPS the periodogram of them with respect to dynamicdisplacements in the vertical direction is illustrated inFigure 13 Only vertical direction is studied because it wasdemonstrated that this direction presented a critical be-havior It is noted as reported in Figure 12 that the fre-quency that contributes the most to the structural responseis located approximately at 937Hz which is related toprevious acceleration spectra is frequency is well detectedby the three instruments Periodogram reported in Figure 13is plotted considering dynamic displacements of smart-phone accelerometer and GPS Hence the accelerationrecorded by the smartphone and accelerometer respectivelyis converted to dynamic displacement by carrying out a

SmartphoneAccelerometer

FFTmdashlongitudinal direction

0

05

1A

mpl

itude

5 10 15 200Frequency (Hz)

(a)

SmartphoneAccelerometer

Periodogrammdashlongitudinal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(b)

SmartphoneAccelerometer

FFTmdashtransversal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(c)

SmartphoneAccelerometer

Periodogrammdashtransversal direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

SmartphoneAccelerometer

FFTmdashvertical direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(e)

SmartphoneAccelerometer

Periodogrammdashvertical direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 12 Dynamic frequencies in terms of FFT and periodogram

12 Advances in Civil Engineering

double integration which may involve some discrepancies inobtaining displacements In addition to numerical in-tegration some of the dierences observed in Figure 13 maybe associated with the applied 13ltering techniques to obtaindynamic displacements On the other hand GPS reportsdisplacements directly improving considerably data pro-cessing and interpretations However it is considered im-portant to analyze the three periodograms from Figure 13 tohave a better understanding about the coincidences of somefrequencies that appear in two or more periodograms and itmay be useful to have a better idea about de dynamic be-havior of a bridge

10 Conclusions

An alternative methodology to perform SHM on real-scalebridges is presented in this paper using GPS an acceler-ometer and a smartphone integrated in a fused smartsensor Such sensor is validated for the SHM of complexstructures as bridges under normal serviceability conditionsin real time A reinforced concrete bridge was studied byusing the proposed smart sensor because it presented ex-cessive vibration problems e accuracy of this smart sensorwas demonstrated comparing the gathered data from GPSaccelerometers and smartphones in both time and fre-quency domain Semistatic and dynamic displacements aswell as accelerations demonstrated that there are criticalperformance problems on the analyzed bridge particularlyin the vertical direction Since this critical behavior wasreported by other scholars that performed the SHM withdierent devices it can be stated that this smart sensor canbe implemented for SHM on bridges in real time In ad-dition it was demonstrated that the proposed smart sensor isadequate to perform SHM in real-scale bridges Conse-quently for the bridge under consideration it was calculated

that the fundamental frequency in the horizontal direction islocated at 9Hz Furthermore it was detected the 13rst fre-quency of interest around 35Hz In order to improvefrequency domain analysis the proposed smart sensormethodology performs a periodogram algorithm over thevertical direction data to obtain a more accurate SHM It wasobserved that there is a dierence between GPS displace-ment data and accelerometer-smartphone based data It canbe justi13ed due to the nature of each measurement meth-odology where GPS base data comes from atmosphere andaccelerometer-based data comes from a vibration sensorBecause of this the noise nature of the aforementionedmeasurement devices is not the same due to its dierentmeasurement principle On the other hand it was observed acoincidence of the three instruments approximately at937Hz for the vertical component us it can explain theusefulness of fusing GPS accelerometer and smartphoneinformation in order to discriminate undesired noise thatcan be generated by atmosphere disturbances of doubleintegration error from acceleration data when it appearsonly in one device data Finally it can be stated that thispaper contributed in understanding how GPS accelerom-eters and smartphones can be successfully fused in a smartsensor to perform SHM on real-scale bridges

Data Availability

e data used to support the 13ndings of this study areavailable from the corresponding author upon request

Disclosure

Any opinions 13ndings conclusions or recommendationsexpressed in this paper are those of the authors and do notnecessarily reect the views of the sponsors

Conflicts of Interest

e authors do not have any conicts of interest to declareabout this research

Acknowledgments

e authors would like to thank Septentrio Inc for pro-viding the AsteRx-U Receiver and the PolaNt-x MF antennaused during the GPS data collection is study is based onwork partly supported by the Consejo Nacional de Ciencia yTecnologıa (CONACYT-Mexico) as scholarships for the 13rstand seventh author respectively In addition the work isalso supported by the Universidad Autonoma de Sinaloa(UAS)

References

[1] V Ashkenazi and G W Roberts ldquoExperimental monitoringof the Humber bridge using GPSrdquo Proceedings of the In-stitution of Civil Engineers-Civil Engineering vol 120 no 4pp 177ndash182 1997

[2] W S Chan Y L Xu X L Ding andW J Dai ldquoAn integratedGPSndashaccelerometer data processing technique for structural

1

09

08

07

06

05

04

03

02

01

00 5 10

Frequency (Hz)15 20

SmartphoneAccelerometerGPS (vertical)

Figure 13 Dynamic frequencies in terms of periodogram (verticaldirection)

Advances in Civil Engineering 13

deformation monitoringrdquo Journal of Geodesy vol 80 no 12pp 705ndash719 2006

[3] S B Im S Hurlebaus and Y J Kang ldquoSummary review ofGPS technology for structural health monitoringrdquo Journal ofStructural Engineering vol 139 no 10 pp 1653ndash1664 2011

[4] M R Kaloop and H Li ldquoMonitoring of bridge deformationusing GPS techniquerdquo KSCE Journal of Civil Engineeringvol 13 no 6 pp 423ndash431 2009

[5] M Kaloop E Elbeltagi J Hu and A Elrefai ldquoRecent advancesof structures monitoring and evaluation using GPS-time seriesmonitoring systems a reviewrdquo ISPRS International Journal ofGeo-Information vol 6 no 12 p 382 2017

[6] X Li C Rizos L Ge Y Tamura and A Yoshida ldquo+ecomplementary characteristics of GPS and accelerometer inmonitoring structural deformationrdquo in Proceedings of the ION2005 National Technical Meeting Institute of Navigation(ION) San Diego CA USA January 2005

[7] P Psimoulis S Pytharouli D Karambalis and S StirosldquoPotential of Global Positioning System (GPS) to measurefrequencies of oscillations of engineering structuresrdquo Journalof Sound and Vibration vol 318 no 3 pp 606ndash623 2008

[8] C Watson T Watson and R Coleman ldquoStructural moni-toring of cable-stayed bridge analysis of GPS versus modeleddeflectionsrdquo Journal of Surveying Engineering vol 133 no 1pp 23ndash28 2007

[9] C Xiong and Y Niu ldquoInvestigation of the dynamic behavior ofa super high-rise structure using RTK-GNSS techniquerdquo KSCEJournal of Civil Engineering vol 23 no 2 pp 654ndash665 2018

[10] B Hofmann-Wellenhof H Lichtenegger and E WasleGNSSndashGlobal Navigation Satellite Systems GPS GLONASSGalileo and More Springer Science amp Business MediaVienna Austria 2007

[11] M Celebi ldquoGPS in dynamic monitoring of long-periodstructuresrdquo Soil Dynamics and Earthquake Engineeringvol 20 no 5ndash8 pp 477ndash483 2000

[12] D K Kim J I Kim and M Q Feng ldquoInstrumentation ofbridges for structural health monitoringrdquo KSCE Journal ofCivil Engineering vol 5 no 3 pp 231ndash242 2001

[13] G Q Wang D M Boore H Igel and X Y Zhou ldquoSomeobservations on colocated and closely spaced strong ground-motion records of the 1999 Chi-Chi Taiwan earthquakerdquoBulletin of the Seismological Society of America vol 93 no 2pp 674ndash693 2003

[14] J Hwang H Yun S K Park D Lee and S Hong ldquoOptimalmethods of RTK-GPSaccelerometer integration to monitorthe displacement of structuresrdquo Sensors vol 12 no 1pp 1014ndash1034 2012

[15] S Dashti J D Bray J Reilly S Glaser A Bayen and E MarildquoEvaluating the reliability of phones as seismic monitoringinstrumentsrdquo Earthquake Spectra vol 30 no 2 pp 721ndash7422014

[16] M Feng Y Fukuda M Mizuta and E Ozer ldquoCitizen sensorsfor SHM use of accelerometer data from smartphonesrdquoSensors vol 15 no 2 pp 2980ndash2998 2015

[17] X Zhao R Han Y Ding et al ldquoPortable and convenient cableforce measurement using smartphonerdquo Journal of CivilStructural Health Monitoring vol 5 no 4 pp 481ndash491 2015

[18] X Zhao K Ri R Han Y Yu M Li and J Ou ldquoExperimentalresearch on quick structural health monitoring technique forbridges using smartphonerdquoAdvances inMaterials Science andEngineering vol 2016 no 1 pp 1ndash14 2016

[19] Y Ding R Han H Liu S Li X Zhao and Y Yu ldquoBridgeinspection and management system based on smart phonerdquoin Proceedings of the ASME 2016 Conference on Smart

Materials Adaptive Structures and Intelligent SystemsAmerican Society of Mechanical Engineers Stowe VT USASeptember 2016

[20] X Zhao R Han K J Loh B Xie J Li and J Ou ldquoShakingtable tests for evaluating the damage features under earth-quake excitations using smartphonesrdquo in Proceedings of theHealth Monitoring of Structural and Biological Systems XIIInternational Society for Optics and Photonics Denver COUSA May 2018

[21] F Moschas and S Stiros ldquoMeasurement of the dynamicdisplacements and of the modal frequencies of a short-spanpedestrian bridge using GPS and an accelerometerrdquo Engi-neering Structures vol 33 no 1 pp 10ndash17 2011

[22] M R Kaloop ldquoBridge safety monitoring based-GPS tech-nique case study Zhujiang Huangpu bridgerdquo Smart Struc-tures and Systems vol 9 no 6 pp 473ndash487 2012

[23] G W Roberts X Meng and A H Dodson ldquoIntegrating aglobal positioning system and accelerometers to monitor thedeflection of bridgesrdquo Journal of Surveying Engineeringvol 130 no 2 pp 65ndash72 2004

[24] Accelerometer Analyzer App 2017 httpsplaygooglecomstoreappsdetailsidcomlulaccelerometeramphles

[25] X Meng ldquoReal-time deformationmonitoring of bridges usingGPSaccelerometersrdquo Doctoral dissertation University ofNottingham UK 2002

[26] G E Vazquez-Becerra J R Gaxiola-Camacho R BennettG M Guzman-Acevedo and I E Gaxiola-CamacholdquoStructural evaluation of dynamic and semi-static displace-ments of the Juarez bridge using GPS technologyrdquo Mea-surement vol 110 no 1 pp 146ndash153 2017

[27] P Psimoulis I Peppa L Bonenberg S Ince and X MengldquoCombination of GPS and RTS measurements for themonitoring of semi-static and dynamic motion of pedestrianbridgerdquo in Proceedings of the 3rd Joint International Sym-posium on Deformation Monitoring Vienna Austria April2016

[28] F Moschas and S Stiros ldquoNoise characteristics of high-frequencyshort-duration GPS records from analysis ofidentical collocated instrumentsrdquo Measurement vol 46no 4 pp 1488ndash1506 2013

[29] S W Smithgte Scientist amp Engineerrsquos Guide to Digital SignalProcessing California Technical Publishing San Diego CAUSA 1997

[30] J Wang X Meng C Qin and J Yi ldquoVibration frequenciesextraction of the forth road bridge using high sampling GPSdatardquo Shock and Vibration vol 2016 Article ID 980786118 pages 2016

[31] J Yu XMeng X Shao B Yan and L Yang ldquoIdentification ofdynamic displacements and modal frequencies of a medium-span suspension bridge using multimode GNSS processingrdquoEngineering Structures vol 81 no 1 pp 432ndash443 2014

[32] E Elbeltagi M Kaloop and M Elnabwy ldquoMonitoring andassessment of Mansoura railway steel bridge using RTK-GPStechniquerdquo in Proceedings of the 2nd Joint InternationalSymposium on Deformation Monitoring pp 9ndash11 Notting-ham England September 2013

[33] N Raziq ldquoGPS structural deformation monitoring the mid-height problemrdquo Doctoral dissertation University ofMelbourneAustralia 2008

[34] S S HaykinAdaptive Filtergteory Pearson Education India5th edition 2014

[35] L H Estey and C M Meertens ldquoTEQC the multi-purposetoolkit for GPSGLONASS datardquo GPS Solutions vol 3 no 1pp 42ndash49 1999

14 Advances in Civil Engineering

the input signal +e resulted set of functions from EMD isdefined as [14]

a(t) 1113944N

n1Cn(t) + rn(t) (7)

where a(t) is the input signal rn(t) is the residue and Cn(t)

is the nth function which is named intrinsic mode functions(IMF) +e EMD filter is expressed as [14]

ab(t) 1113944r

nq

Cn(t) (8)

where

1lt qlt rlt n (9)

Hence the above expressions can be used to calculatedynamic displacements from acceleration time historiesrecorded by smartphone and accelerometer respectively

5 Application of GPS ReceiversAccelerometers and Smartphones in SHM

As previously documented several are the advantages anddisadvantages of GPS receivers accelerometers and smart-phones in SHM Both GPS receivers and commercial accel-erometers have limitations that can be addressed using certainmethodologies For example integrating GPS and commer-cial accelerometer the combined result may recover lost dataeg missing measurements from GPS +is integration maysolve the lack of semistatic and static components +e re-dundancy of both systems (GPS and accelerometers) providesa good performance in SHM [6] In general it is reported inthe literature that high-frequency displacements may beobtained from accelerometers and low frequency ones fromGPS +e reason of considering low frequencies from GPSand high ones from accelerometers is that some scholars [14]demonstrated that GPS is a helpful tool for measuring staticand semistatic behaviors (low-frequency displacements) butit experiences problems related to collecting data of dynamicdisplacements from structures vibrating with high frequen-cies On the other hand accelerometer records dynamicresponses very accurately without low-frequency compo-nents +erefore it is reported in the literature that accel-erometers and GPS make a reasonable complement to eachother [2 6 14] Unfortunately GPS receivers and acceler-ometers are expensive instruments Commonly the accel-eration is recorded using commercial accelerometersHowever the improved quality of acceleration sensors in-tegrated in smartphones provides a feasible alternativeGenerally their level of accuracy may be lower than acommercial accelerometer but it is reported in the literaturethat their results are very similar between them for smallstructural systems [15 16] +e contribution of this article tothe literature is that accelerometer sensors of smartphones areused together with other devices to form a smart sensor to beused in the SHM of a reinforced concrete bridge in real time+en a parametric analysis is performed with respect to GPSreceivers and commercial accelerometers

6 Dynamic Characteristics of Structures

Before proceeding further with the discussion dynamiccharacteristic of structures must be introduced Structuresare designed to resist different types of loading conditions+ey must have the capacity to reliably transmit loadingsfrom the applied point to their foundations when this ishappening movements may be produced in structures as aresponse Movements are generated following several fre-quencies depending mainly on the degrees of freedom of thestructure one of the most important frequencies is thenatural or fundamental frequency [33] To analyze fre-quencies is necessary to change the analysis from time tofrequency domain generally using the Fast Fourier Trans-formation (FFT) In general sense FFT is a computationalcomplexity improvement of DFT that separates a complexsignal into a set of sinusoidal ones where each signal has acorresponded frequency and potential Similarly there areother techniques such as periodogram It is well-documented that both FFT and periodogram algorithmscan be utilized for analyzing movements of bridge structureson frequency domain [34] Because FFT is a well-knownalgorithm its mathematical formulation is not describedhere On the other hand periodogram equation can beobserved in Equation (10) where x(t) represents the timedomain input signal X(f) is its Fourier transform and Xp(f )is the output periodogram which is calculated as the Fouriertransform of the autocorrelation function from x(t) It alsocan be alternatively calculated as the multiplication betweenX(f ) and its conjugated counterpart as [34]

Xp(f) F x(t)lowastx(minust) X(f) middot Xlowast(f) (10)

7 StructuralDescriptionof theAnalyzedBridge

+e analyzed structure in this paper using the fused smartsensor is the Juarez Bridge (see Figure 5) Such a structure islocated in Culiacan Mexico It is a reinforced concretebridge with an approximate length of 200meters +is is animportant civil infrastructure of the city Some importantfindings about the structural performance of this bridge weredocumented earlier by Vazquez-Becerra et al [26] It wasreported that the bridge presented deformations in averageof about 8 cm in the vertical direction representing anundesirable structural behavior In the above researchmaximum displacements were recorded at the ZEN2 station(see Figure 5(b)) which was reported to be the most affectedlocation over the bridge mainly because several vehicles arestopped by a traffic light close to it Considering displace-ments with respect to the vertical direction the calculatedaverage of probability of failure based on comparingsemistatic displacement with AASHTO (American Associ-ation of State Highway and Transportation Officials) de-flection limits was about 40 In addition according toVazquez-Becerra et al [26] the Juarez Bridgersquos probability offailure and its maximum displacements were calculated bythe average displacements extracted via GPS receivers lo-cated at critical points of the bridge However they only usedGPS receivers to extract the performance of the structure

6 Advances in Civil Engineering

under normal loading conditions On the other hand for theinvestigation presented in this paper the GPS performancewas improved by adding acceleration data using commercialaccelerometer and smartphone respectively +e combi-nation between GPS and commercial accelerometers hasbeen developed and reported in the literature [2 14 23]However very few information is documented in the lit-erature about using sensors embedded in smartphones torecord acceleration of real-scale bridges Consequently ifresults obtained via smartphone for bridge structures areviable this will provide a great advancement in SHM ofbridges because price and time would be drastically reducedin extracting their structural performance in real time

8 Measurement Campaign Using FusedSmart Sensor

For this study the measurement campaign was developed atthe ZEN2 station (see Figure 5(b)) because it was previouslydocumented by Vazquez-Becerra et al [26] that this locationpresented the worst performance with respect to averagedisplacements and probability of failure Additionally thesemeasurements were developed on Monday since this daypresented the highest probability of inappropriate structuralbehavior +e structural assessment of the bridge was de-veloped simultaneously in terms of the smart sensor in-tegrating GPS receivers accelerometer and smartphone+e process is comprehensively described in this section

81 GPS Receivers Campaign +e GPS measurementmethodology employed was the kinematic relative posi-tioning which involves two stations measuring simulta-neously one inmovement (the one located on the bridge) andthe other one static (outside of the bridge as static reference)[10] +e stable-based reference station used is the 373Hstation this station is located near to the bridge It is part ofthe active national geodetic network from the Mexican Na-tional Institute of Statistics and Geography (INEGI inSpanish) Figure 6(a) shows the position of station 373H withrespect to the ZEN2 station +e length of the baseline(from ZEN2 to 373H) is approximately 120meters +e

displacement measurement lasted 500 seconds using asampling rate of 100Hz when the Juarez Bridge was subjectedto in-service traffic loading conditions Figures 6(b) and 6(c)illustrate the ZEN2 and 373H stations respectively+ey weremonitored during the GPS displacements campaign

Previously to postprocessing the collected data raw GPSdata was debugged by Teqc (translation editing and qualitycheck) [35] +e preprocessing consists in insulate GPS dataand perform a quality check+e latest step is very interestingconsidering that it provides an understanding of the condi-tion of the GPS receiverrsquos position in terms of Multipath(MP) visible satellites and signal to noise ratio (S) For thiscase the results demonstrate that the rover reports multipathvalues of MP1RMS 00478m MP2RMS 00119m 19 GPSsatellites with observations S1 701 and S2 449 for thereference station MP1RMS 00428m MP2RMS 00114S1 712 and S2 449 the values of both receivers are verysimilar due to the short distance between them On the otherhand because relative positioning with double differencing(DD) over a short baseline (less than 10 kilometers of length)was used some error sources were either reduced or droppedfor example clock bias from the satellite-receiver and at-mospheric effects (ionospheric and tropospheric bias) be-cause both receivers are generally affected by the sameconditions [36] Additionally this positioning techniquemakes possible to achieve a good performance using com-mercial software for processing raw GPS data which can beseen in most of investigations related to the use of GPS re-ceivers in SHM [7 21 28 31 32] However due to the ex-cessive amount of data the software considered to process allthe information is the GAMITGLOBK TRACK developed bythe MIT (Massachusetts Institute of Technology) +is soft-ware processes GPS data in kinematic mode +e mainspecifications of GPS data processing included 15deg cut-offangle precise final orbits disseminated by IGS (InternationalGNSS Service) mode air (used for high sample) NationalGeodetic survey antenna calibration parameters and others

82 Accelerometer Campaign As previously discussedduring the 500 seconds measurement session one com-mercial accelerometer was utilized +is device was

(a) (b)

Figure 5 Juarez Bridge (a) lateral view (b) ZEN2 GPS location

Advances in Civil Engineering 7

recording acceleration at a rate of 100Hz+e accelerometerwas perfectly fixed to present the same physical center as theother devices to record the behavior of ZEN2 stationFigure 6(b) illustrates the smart sensor that is holding thethree devices One more important fact in the accelerometercampaign is that information was recorded at the same timethan GPS and smartphone providing a correct synchroni-zation [37]

83 SmartphoneCampaign As illustrated earlier in Figure 4the smartphone was fixed at the lowest plate of the mechanicstructure for the smart sensor +e measured informationfrom the accelerometer embedded in the smartphone wascollected by an open access App [24] +is application canrecord acceleration and then send it to the user in the form oftext file+e smartphone collected data at the same time thanthe GPS receiver and commercial accelerometer during500 seconds with a sampling rate of 400Hz

84 Flow Chart Diagram of the Measurement Process Insummary the data acquisition process of the bridge usingthe smart sensor is illustrated in Figure 7 Firstly theaccelerometer smartphone and GPS antenna are perfectlyfixed to the location of ZEN2 station +e particularissues related to this location were previously discussedSecondly acceleration time histories are extracted fromaccelerometer and smartphone respectively and dis-placements from GPS receivers As documented earlier thedata acquisition process is performed simultaneously usingthe three instruments +irdly the apparent displacementin three directions (N-S E-W and vertical) is calculatedusing GPS +en semistatic and dynamic displacementsare extracted using low- and high-pass filters respectivelyFourthly in order to study the dynamic characteristics ofthe Juarez Bridge FFTand periodogram are plotted for thelongitudinal transversal and vertical directions using rawacceleration of smartphone and accelerometer re-spectively Finally to compare the extracting responsecapabilities of the three instruments dynamic displace-ments reported by every device are used for the calculationof periodograms +e above discussion is clarified in thenext section

9 Structural Response Collected by theSmart Sensor

+e structural performance of the Juarez Bridge is studiedunder normal service loading conditions in real time ie thestructure is evaluated during a loading condition that is ex-perienced daily by users In fact a nonpermissible vibration inthe vertical direction has been reported by users and it hasbeen documented in the literature by Vazquez-Becerra et al[26] To verify the undesirable behavior of this bridge in thisstudy the authors decided to implement other devices duringthe SHM Hence the collected structural response is obtainedsimultaneously using GPS accelerometer and smartphone inreal time as a fused smart sensor

91 Acceleration TimeHistory One measurement session of500 seconds was performed using the three instrumentsunder consideration In terms of collected accelerationFigure 8 illustrates the recorded response of the smartphoneand accelerometer respectively Figures 8(a)ndash8(c) illustratethe longitudinal transversal and vertical acceleration re-spectively Several observations can be made correspondingto Figure 8 It can be observed that acceleration in thevertical direction is greater than longitudinal and transversaldirections In terms of displacements the same observationwas reported by [26] In addition it can be noted inFigure 8(a) that in the longitudinal direction the acceler-ation reported by smartphone and accelerometer is pre-senting some discrepancies +is issue may be justifiedbecause of the precision of commercial accelerometers insuch a direction However this structural response does notrepresent a threat of safety because of the stiffness of thebridge in the longitudinal direction In addition it can beobserved in Figure 8(c) that the critical component is as-sociated to accelerations acting in the vertical directionConversely in terms of accuracy in Figures 8(b) and 8(c) isshown that acceleration for both smartphone and acceler-ometer is matching very closely In general terms Figure 8demonstrates that acceleration recorded by smartphone andaccelerometer can be positively comparable

92 Semistatic Displacements Using GPS Considering dis-placements obtained via GPS semistatic displacements are

(a) (b) (c)

Figure 6 GPS campaign (a) location of ZEN2 and 373H (b) station ZEN2 (c) station 373H

8 Advances in Civil Engineering

extracted by applying a low-pass 13lter Semistatic dis-placements corresponding to the bridge under considerationrepresent slowmovements of the structure they are reportedin the N-S E-W and vertical direction in Figures 9(a)ndash9(c)respectively

It can be observed in Figures 9(a) and 9(b) that smallerdisplacements are present in both N-S and E-W horizontaldirections than those in the vertical one is is illustrated inFigure 9(c) it is noted that considerable great displacementsare reported in the vertical direction Hence in terms ofsemistatic displacements it can be justi13ed that the verticalcomponent is presenting critical behavior ere exist casesof bridge studies with longer dimensions such as the SanFrancisco-Oakland bay bridge that presented small dis-placements [38] However the analyzed bridge presented inthis paper is a very particular case of study whichreports unusual behavior with respect to the vertical di-rection comparing semistatic displacement with AASHTO

deection limits atrsquos why the authors of this paper de-cided to use it for the evaluation of the sensing capabilities ofthe proposed smart sensor

93 Dynamic Displacements Using GPS Movements relatedto short period component are represented by dynamicdisplacements (see Figure 10)

Dynamic displacements are obtained applying a high-pass 13lter to GPS apparent displacements Figure 10 illus-trates dynamic displacements corresponding to the bridge Itcan be noted in Figures 10(a) and 10(b) that dynamicdisplacements are relatively small for both N-S and E-Whorizontal components On the contrary the vertical di-rection is presenting excessive displacements and this isillustrated in Figure 10(c) One more time critical dis-placements are appearing for the vertical component but inthis case with respect to dynamic displacements

Longitudinal direction

100 200 300 400 5000Time (sec)

Smartphone Accelerometer

ndash1

0

1

Acc

eler

atio

n (m

s2 )

(a)

Transversal directionSmartphone Accelerometer

ndash1

0

1A

ccel

erat

ion

(ms

2 )

100 200 300 400 5000

Time (sec)

(b)

Vertical directionSmartphone Accelerometer

100 200 300 400 5000

Time (sec)

ndash2

0

2

Acc

eler

atio

n (m

s2 )

(c)

Figure 8 Acceleration time history

Accelerometer

Acceleration

High-pass filterFFTPeriodogram

Detrend

t

0

Ri ndash (1n)sumni=1 Ri

t

0

Dynamicdisplacement

Semistaticdisplacement

Low-pass filterBand-Pass filter

Smartphone Displacementfrom GPS

Figure 7 Flow chart of measurement process

Advances in Civil Engineering 9

Apparent displacementSemistatic displacement

ndash001

0

001D

ispla

cem

ent (

m)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(a)

Apparent displacementSemistatic displacement

ndash005

0

005

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(b)

Apparent displacementSemistatic displacement

ndash001

0

001

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(c)

Figure 9 Semistatic displacements in three directions (a) N-S (b) E-W and (c) vertical

NndashS

ndash001

0

001

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(a)

EndashW

ndash005

0

005

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(b)

Figure 10 Continued

10 Advances in Civil Engineering

94 FFT and Periodogram Using GPS DynamicDisplacements To identify structural dynamic character-istics of the structure under consideration FFT and perio-dograms are calculated and plotted in frequency domainusing dynamic displacements of GPS

Figure 11 illustrates FFT and periodogram in the threeGPS directions N-S (Figures 11(a) and 11(b)) E-W(Figures 11(c) and 11(d)) and vertical (Figures 11(e) and11(f)) To clarify the results they are normalized with respectto their maximum amplitude value

95 Acceleration Time History Using Accelerometer andSmartphone Acceleration time histories corresponding tothe longitudinal transversal and vertical direction of thebridge were earlier presented in Figure 8 considering bothsmartphone and accelerometer In the longitudinal direction(see Figure 8(a)) it can be noted that there are some dif-ferences in the acceleration reported by smartphone andaccelerometer such discrepancies were justi13ed earlier inSection 91 In the transversal (see Figure 8(b)) and verticaldirections (see Figure 8(c)) it is observed that smartphone

Vertical

ndash001

0

001D

ispla

cem

ent (

m)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(c)

Figure 10 Dynamic displacements in three directions (a) N-S (b) E-W and (c) vertical

NndashSFFT

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(a)

NndashSPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(b)

EndashWFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(c)

EndashWPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

VerticalFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(e)

VerticalPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 11 Dynamic frequencies in terms of normalized FFT and periodogram

Advances in Civil Engineering 11

and accelerometer are recording very similar results In timedomain these results are very promising However tocomplement them results are changed to the frequencydomain in terms of FFT and periodogram is is discussedin the next section

96 FFT and Periodogram Using Accelerometer and Smart-phone Accelerations In this section results are studied infrequency domain ey are presented in terms of FFT andperiodogram (see Figure 12)

Figure 12 summarizes the results related to FFT andperiodogram corresponding to longitudinal transversal andvertical directions It can be observed in Figure 12 that thefrequency that contributes the most to the structural re-sponse may be located approximately at 9Hz In addition itis noted that FFT in the longitudinal and transversal di-rection (see Figures 12(a) and 12(c)) is presenting the 13rstfrequency of interest at about 35Hz Based on the resultsplotted in Figure 12 it can be stated that both smartphoneand accelerometer are providing very accurate results with

respect to each other Hence it may result practical to usesmartphones for extracting dynamic characteristics of real-scale bridges

97 Comparison of Periodogram of GPS Accelerometers andSmartphones Using Dynamic Displacements To comparedata acquisition capabilities of smartphones accelerometersand GPS the periodogram of them with respect to dynamicdisplacements in the vertical direction is illustrated inFigure 13 Only vertical direction is studied because it wasdemonstrated that this direction presented a critical be-havior It is noted as reported in Figure 12 that the fre-quency that contributes the most to the structural responseis located approximately at 937Hz which is related toprevious acceleration spectra is frequency is well detectedby the three instruments Periodogram reported in Figure 13is plotted considering dynamic displacements of smart-phone accelerometer and GPS Hence the accelerationrecorded by the smartphone and accelerometer respectivelyis converted to dynamic displacement by carrying out a

SmartphoneAccelerometer

FFTmdashlongitudinal direction

0

05

1A

mpl

itude

5 10 15 200Frequency (Hz)

(a)

SmartphoneAccelerometer

Periodogrammdashlongitudinal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(b)

SmartphoneAccelerometer

FFTmdashtransversal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(c)

SmartphoneAccelerometer

Periodogrammdashtransversal direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

SmartphoneAccelerometer

FFTmdashvertical direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(e)

SmartphoneAccelerometer

Periodogrammdashvertical direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 12 Dynamic frequencies in terms of FFT and periodogram

12 Advances in Civil Engineering

double integration which may involve some discrepancies inobtaining displacements In addition to numerical in-tegration some of the dierences observed in Figure 13 maybe associated with the applied 13ltering techniques to obtaindynamic displacements On the other hand GPS reportsdisplacements directly improving considerably data pro-cessing and interpretations However it is considered im-portant to analyze the three periodograms from Figure 13 tohave a better understanding about the coincidences of somefrequencies that appear in two or more periodograms and itmay be useful to have a better idea about de dynamic be-havior of a bridge

10 Conclusions

An alternative methodology to perform SHM on real-scalebridges is presented in this paper using GPS an acceler-ometer and a smartphone integrated in a fused smartsensor Such sensor is validated for the SHM of complexstructures as bridges under normal serviceability conditionsin real time A reinforced concrete bridge was studied byusing the proposed smart sensor because it presented ex-cessive vibration problems e accuracy of this smart sensorwas demonstrated comparing the gathered data from GPSaccelerometers and smartphones in both time and fre-quency domain Semistatic and dynamic displacements aswell as accelerations demonstrated that there are criticalperformance problems on the analyzed bridge particularlyin the vertical direction Since this critical behavior wasreported by other scholars that performed the SHM withdierent devices it can be stated that this smart sensor canbe implemented for SHM on bridges in real time In ad-dition it was demonstrated that the proposed smart sensor isadequate to perform SHM in real-scale bridges Conse-quently for the bridge under consideration it was calculated

that the fundamental frequency in the horizontal direction islocated at 9Hz Furthermore it was detected the 13rst fre-quency of interest around 35Hz In order to improvefrequency domain analysis the proposed smart sensormethodology performs a periodogram algorithm over thevertical direction data to obtain a more accurate SHM It wasobserved that there is a dierence between GPS displace-ment data and accelerometer-smartphone based data It canbe justi13ed due to the nature of each measurement meth-odology where GPS base data comes from atmosphere andaccelerometer-based data comes from a vibration sensorBecause of this the noise nature of the aforementionedmeasurement devices is not the same due to its dierentmeasurement principle On the other hand it was observed acoincidence of the three instruments approximately at937Hz for the vertical component us it can explain theusefulness of fusing GPS accelerometer and smartphoneinformation in order to discriminate undesired noise thatcan be generated by atmosphere disturbances of doubleintegration error from acceleration data when it appearsonly in one device data Finally it can be stated that thispaper contributed in understanding how GPS accelerom-eters and smartphones can be successfully fused in a smartsensor to perform SHM on real-scale bridges

Data Availability

e data used to support the 13ndings of this study areavailable from the corresponding author upon request

Disclosure

Any opinions 13ndings conclusions or recommendationsexpressed in this paper are those of the authors and do notnecessarily reect the views of the sponsors

Conflicts of Interest

e authors do not have any conicts of interest to declareabout this research

Acknowledgments

e authors would like to thank Septentrio Inc for pro-viding the AsteRx-U Receiver and the PolaNt-x MF antennaused during the GPS data collection is study is based onwork partly supported by the Consejo Nacional de Ciencia yTecnologıa (CONACYT-Mexico) as scholarships for the 13rstand seventh author respectively In addition the work isalso supported by the Universidad Autonoma de Sinaloa(UAS)

References

[1] V Ashkenazi and G W Roberts ldquoExperimental monitoringof the Humber bridge using GPSrdquo Proceedings of the In-stitution of Civil Engineers-Civil Engineering vol 120 no 4pp 177ndash182 1997

[2] W S Chan Y L Xu X L Ding andW J Dai ldquoAn integratedGPSndashaccelerometer data processing technique for structural

1

09

08

07

06

05

04

03

02

01

00 5 10

Frequency (Hz)15 20

SmartphoneAccelerometerGPS (vertical)

Figure 13 Dynamic frequencies in terms of periodogram (verticaldirection)

Advances in Civil Engineering 13

deformation monitoringrdquo Journal of Geodesy vol 80 no 12pp 705ndash719 2006

[3] S B Im S Hurlebaus and Y J Kang ldquoSummary review ofGPS technology for structural health monitoringrdquo Journal ofStructural Engineering vol 139 no 10 pp 1653ndash1664 2011

[4] M R Kaloop and H Li ldquoMonitoring of bridge deformationusing GPS techniquerdquo KSCE Journal of Civil Engineeringvol 13 no 6 pp 423ndash431 2009

[5] M Kaloop E Elbeltagi J Hu and A Elrefai ldquoRecent advancesof structures monitoring and evaluation using GPS-time seriesmonitoring systems a reviewrdquo ISPRS International Journal ofGeo-Information vol 6 no 12 p 382 2017

[6] X Li C Rizos L Ge Y Tamura and A Yoshida ldquo+ecomplementary characteristics of GPS and accelerometer inmonitoring structural deformationrdquo in Proceedings of the ION2005 National Technical Meeting Institute of Navigation(ION) San Diego CA USA January 2005

[7] P Psimoulis S Pytharouli D Karambalis and S StirosldquoPotential of Global Positioning System (GPS) to measurefrequencies of oscillations of engineering structuresrdquo Journalof Sound and Vibration vol 318 no 3 pp 606ndash623 2008

[8] C Watson T Watson and R Coleman ldquoStructural moni-toring of cable-stayed bridge analysis of GPS versus modeleddeflectionsrdquo Journal of Surveying Engineering vol 133 no 1pp 23ndash28 2007

[9] C Xiong and Y Niu ldquoInvestigation of the dynamic behavior ofa super high-rise structure using RTK-GNSS techniquerdquo KSCEJournal of Civil Engineering vol 23 no 2 pp 654ndash665 2018

[10] B Hofmann-Wellenhof H Lichtenegger and E WasleGNSSndashGlobal Navigation Satellite Systems GPS GLONASSGalileo and More Springer Science amp Business MediaVienna Austria 2007

[11] M Celebi ldquoGPS in dynamic monitoring of long-periodstructuresrdquo Soil Dynamics and Earthquake Engineeringvol 20 no 5ndash8 pp 477ndash483 2000

[12] D K Kim J I Kim and M Q Feng ldquoInstrumentation ofbridges for structural health monitoringrdquo KSCE Journal ofCivil Engineering vol 5 no 3 pp 231ndash242 2001

[13] G Q Wang D M Boore H Igel and X Y Zhou ldquoSomeobservations on colocated and closely spaced strong ground-motion records of the 1999 Chi-Chi Taiwan earthquakerdquoBulletin of the Seismological Society of America vol 93 no 2pp 674ndash693 2003

[14] J Hwang H Yun S K Park D Lee and S Hong ldquoOptimalmethods of RTK-GPSaccelerometer integration to monitorthe displacement of structuresrdquo Sensors vol 12 no 1pp 1014ndash1034 2012

[15] S Dashti J D Bray J Reilly S Glaser A Bayen and E MarildquoEvaluating the reliability of phones as seismic monitoringinstrumentsrdquo Earthquake Spectra vol 30 no 2 pp 721ndash7422014

[16] M Feng Y Fukuda M Mizuta and E Ozer ldquoCitizen sensorsfor SHM use of accelerometer data from smartphonesrdquoSensors vol 15 no 2 pp 2980ndash2998 2015

[17] X Zhao R Han Y Ding et al ldquoPortable and convenient cableforce measurement using smartphonerdquo Journal of CivilStructural Health Monitoring vol 5 no 4 pp 481ndash491 2015

[18] X Zhao K Ri R Han Y Yu M Li and J Ou ldquoExperimentalresearch on quick structural health monitoring technique forbridges using smartphonerdquoAdvances inMaterials Science andEngineering vol 2016 no 1 pp 1ndash14 2016

[19] Y Ding R Han H Liu S Li X Zhao and Y Yu ldquoBridgeinspection and management system based on smart phonerdquoin Proceedings of the ASME 2016 Conference on Smart

Materials Adaptive Structures and Intelligent SystemsAmerican Society of Mechanical Engineers Stowe VT USASeptember 2016

[20] X Zhao R Han K J Loh B Xie J Li and J Ou ldquoShakingtable tests for evaluating the damage features under earth-quake excitations using smartphonesrdquo in Proceedings of theHealth Monitoring of Structural and Biological Systems XIIInternational Society for Optics and Photonics Denver COUSA May 2018

[21] F Moschas and S Stiros ldquoMeasurement of the dynamicdisplacements and of the modal frequencies of a short-spanpedestrian bridge using GPS and an accelerometerrdquo Engi-neering Structures vol 33 no 1 pp 10ndash17 2011

[22] M R Kaloop ldquoBridge safety monitoring based-GPS tech-nique case study Zhujiang Huangpu bridgerdquo Smart Struc-tures and Systems vol 9 no 6 pp 473ndash487 2012

[23] G W Roberts X Meng and A H Dodson ldquoIntegrating aglobal positioning system and accelerometers to monitor thedeflection of bridgesrdquo Journal of Surveying Engineeringvol 130 no 2 pp 65ndash72 2004

[24] Accelerometer Analyzer App 2017 httpsplaygooglecomstoreappsdetailsidcomlulaccelerometeramphles

[25] X Meng ldquoReal-time deformationmonitoring of bridges usingGPSaccelerometersrdquo Doctoral dissertation University ofNottingham UK 2002

[26] G E Vazquez-Becerra J R Gaxiola-Camacho R BennettG M Guzman-Acevedo and I E Gaxiola-CamacholdquoStructural evaluation of dynamic and semi-static displace-ments of the Juarez bridge using GPS technologyrdquo Mea-surement vol 110 no 1 pp 146ndash153 2017

[27] P Psimoulis I Peppa L Bonenberg S Ince and X MengldquoCombination of GPS and RTS measurements for themonitoring of semi-static and dynamic motion of pedestrianbridgerdquo in Proceedings of the 3rd Joint International Sym-posium on Deformation Monitoring Vienna Austria April2016

[28] F Moschas and S Stiros ldquoNoise characteristics of high-frequencyshort-duration GPS records from analysis ofidentical collocated instrumentsrdquo Measurement vol 46no 4 pp 1488ndash1506 2013

[29] S W Smithgte Scientist amp Engineerrsquos Guide to Digital SignalProcessing California Technical Publishing San Diego CAUSA 1997

[30] J Wang X Meng C Qin and J Yi ldquoVibration frequenciesextraction of the forth road bridge using high sampling GPSdatardquo Shock and Vibration vol 2016 Article ID 980786118 pages 2016

[31] J Yu XMeng X Shao B Yan and L Yang ldquoIdentification ofdynamic displacements and modal frequencies of a medium-span suspension bridge using multimode GNSS processingrdquoEngineering Structures vol 81 no 1 pp 432ndash443 2014

[32] E Elbeltagi M Kaloop and M Elnabwy ldquoMonitoring andassessment of Mansoura railway steel bridge using RTK-GPStechniquerdquo in Proceedings of the 2nd Joint InternationalSymposium on Deformation Monitoring pp 9ndash11 Notting-ham England September 2013

[33] N Raziq ldquoGPS structural deformation monitoring the mid-height problemrdquo Doctoral dissertation University ofMelbourneAustralia 2008

[34] S S HaykinAdaptive Filtergteory Pearson Education India5th edition 2014

[35] L H Estey and C M Meertens ldquoTEQC the multi-purposetoolkit for GPSGLONASS datardquo GPS Solutions vol 3 no 1pp 42ndash49 1999

14 Advances in Civil Engineering

under normal loading conditions On the other hand for theinvestigation presented in this paper the GPS performancewas improved by adding acceleration data using commercialaccelerometer and smartphone respectively +e combi-nation between GPS and commercial accelerometers hasbeen developed and reported in the literature [2 14 23]However very few information is documented in the lit-erature about using sensors embedded in smartphones torecord acceleration of real-scale bridges Consequently ifresults obtained via smartphone for bridge structures areviable this will provide a great advancement in SHM ofbridges because price and time would be drastically reducedin extracting their structural performance in real time

8 Measurement Campaign Using FusedSmart Sensor

For this study the measurement campaign was developed atthe ZEN2 station (see Figure 5(b)) because it was previouslydocumented by Vazquez-Becerra et al [26] that this locationpresented the worst performance with respect to averagedisplacements and probability of failure Additionally thesemeasurements were developed on Monday since this daypresented the highest probability of inappropriate structuralbehavior +e structural assessment of the bridge was de-veloped simultaneously in terms of the smart sensor in-tegrating GPS receivers accelerometer and smartphone+e process is comprehensively described in this section

81 GPS Receivers Campaign +e GPS measurementmethodology employed was the kinematic relative posi-tioning which involves two stations measuring simulta-neously one inmovement (the one located on the bridge) andthe other one static (outside of the bridge as static reference)[10] +e stable-based reference station used is the 373Hstation this station is located near to the bridge It is part ofthe active national geodetic network from the Mexican Na-tional Institute of Statistics and Geography (INEGI inSpanish) Figure 6(a) shows the position of station 373H withrespect to the ZEN2 station +e length of the baseline(from ZEN2 to 373H) is approximately 120meters +e

displacement measurement lasted 500 seconds using asampling rate of 100Hz when the Juarez Bridge was subjectedto in-service traffic loading conditions Figures 6(b) and 6(c)illustrate the ZEN2 and 373H stations respectively+ey weremonitored during the GPS displacements campaign

Previously to postprocessing the collected data raw GPSdata was debugged by Teqc (translation editing and qualitycheck) [35] +e preprocessing consists in insulate GPS dataand perform a quality check+e latest step is very interestingconsidering that it provides an understanding of the condi-tion of the GPS receiverrsquos position in terms of Multipath(MP) visible satellites and signal to noise ratio (S) For thiscase the results demonstrate that the rover reports multipathvalues of MP1RMS 00478m MP2RMS 00119m 19 GPSsatellites with observations S1 701 and S2 449 for thereference station MP1RMS 00428m MP2RMS 00114S1 712 and S2 449 the values of both receivers are verysimilar due to the short distance between them On the otherhand because relative positioning with double differencing(DD) over a short baseline (less than 10 kilometers of length)was used some error sources were either reduced or droppedfor example clock bias from the satellite-receiver and at-mospheric effects (ionospheric and tropospheric bias) be-cause both receivers are generally affected by the sameconditions [36] Additionally this positioning techniquemakes possible to achieve a good performance using com-mercial software for processing raw GPS data which can beseen in most of investigations related to the use of GPS re-ceivers in SHM [7 21 28 31 32] However due to the ex-cessive amount of data the software considered to process allthe information is the GAMITGLOBK TRACK developed bythe MIT (Massachusetts Institute of Technology) +is soft-ware processes GPS data in kinematic mode +e mainspecifications of GPS data processing included 15deg cut-offangle precise final orbits disseminated by IGS (InternationalGNSS Service) mode air (used for high sample) NationalGeodetic survey antenna calibration parameters and others

82 Accelerometer Campaign As previously discussedduring the 500 seconds measurement session one com-mercial accelerometer was utilized +is device was

(a) (b)

Figure 5 Juarez Bridge (a) lateral view (b) ZEN2 GPS location

Advances in Civil Engineering 7

recording acceleration at a rate of 100Hz+e accelerometerwas perfectly fixed to present the same physical center as theother devices to record the behavior of ZEN2 stationFigure 6(b) illustrates the smart sensor that is holding thethree devices One more important fact in the accelerometercampaign is that information was recorded at the same timethan GPS and smartphone providing a correct synchroni-zation [37]

83 SmartphoneCampaign As illustrated earlier in Figure 4the smartphone was fixed at the lowest plate of the mechanicstructure for the smart sensor +e measured informationfrom the accelerometer embedded in the smartphone wascollected by an open access App [24] +is application canrecord acceleration and then send it to the user in the form oftext file+e smartphone collected data at the same time thanthe GPS receiver and commercial accelerometer during500 seconds with a sampling rate of 400Hz

84 Flow Chart Diagram of the Measurement Process Insummary the data acquisition process of the bridge usingthe smart sensor is illustrated in Figure 7 Firstly theaccelerometer smartphone and GPS antenna are perfectlyfixed to the location of ZEN2 station +e particularissues related to this location were previously discussedSecondly acceleration time histories are extracted fromaccelerometer and smartphone respectively and dis-placements from GPS receivers As documented earlier thedata acquisition process is performed simultaneously usingthe three instruments +irdly the apparent displacementin three directions (N-S E-W and vertical) is calculatedusing GPS +en semistatic and dynamic displacementsare extracted using low- and high-pass filters respectivelyFourthly in order to study the dynamic characteristics ofthe Juarez Bridge FFTand periodogram are plotted for thelongitudinal transversal and vertical directions using rawacceleration of smartphone and accelerometer re-spectively Finally to compare the extracting responsecapabilities of the three instruments dynamic displace-ments reported by every device are used for the calculationof periodograms +e above discussion is clarified in thenext section

9 Structural Response Collected by theSmart Sensor

+e structural performance of the Juarez Bridge is studiedunder normal service loading conditions in real time ie thestructure is evaluated during a loading condition that is ex-perienced daily by users In fact a nonpermissible vibration inthe vertical direction has been reported by users and it hasbeen documented in the literature by Vazquez-Becerra et al[26] To verify the undesirable behavior of this bridge in thisstudy the authors decided to implement other devices duringthe SHM Hence the collected structural response is obtainedsimultaneously using GPS accelerometer and smartphone inreal time as a fused smart sensor

91 Acceleration TimeHistory One measurement session of500 seconds was performed using the three instrumentsunder consideration In terms of collected accelerationFigure 8 illustrates the recorded response of the smartphoneand accelerometer respectively Figures 8(a)ndash8(c) illustratethe longitudinal transversal and vertical acceleration re-spectively Several observations can be made correspondingto Figure 8 It can be observed that acceleration in thevertical direction is greater than longitudinal and transversaldirections In terms of displacements the same observationwas reported by [26] In addition it can be noted inFigure 8(a) that in the longitudinal direction the acceler-ation reported by smartphone and accelerometer is pre-senting some discrepancies +is issue may be justifiedbecause of the precision of commercial accelerometers insuch a direction However this structural response does notrepresent a threat of safety because of the stiffness of thebridge in the longitudinal direction In addition it can beobserved in Figure 8(c) that the critical component is as-sociated to accelerations acting in the vertical directionConversely in terms of accuracy in Figures 8(b) and 8(c) isshown that acceleration for both smartphone and acceler-ometer is matching very closely In general terms Figure 8demonstrates that acceleration recorded by smartphone andaccelerometer can be positively comparable

92 Semistatic Displacements Using GPS Considering dis-placements obtained via GPS semistatic displacements are

(a) (b) (c)

Figure 6 GPS campaign (a) location of ZEN2 and 373H (b) station ZEN2 (c) station 373H

8 Advances in Civil Engineering

extracted by applying a low-pass 13lter Semistatic dis-placements corresponding to the bridge under considerationrepresent slowmovements of the structure they are reportedin the N-S E-W and vertical direction in Figures 9(a)ndash9(c)respectively

It can be observed in Figures 9(a) and 9(b) that smallerdisplacements are present in both N-S and E-W horizontaldirections than those in the vertical one is is illustrated inFigure 9(c) it is noted that considerable great displacementsare reported in the vertical direction Hence in terms ofsemistatic displacements it can be justi13ed that the verticalcomponent is presenting critical behavior ere exist casesof bridge studies with longer dimensions such as the SanFrancisco-Oakland bay bridge that presented small dis-placements [38] However the analyzed bridge presented inthis paper is a very particular case of study whichreports unusual behavior with respect to the vertical di-rection comparing semistatic displacement with AASHTO

deection limits atrsquos why the authors of this paper de-cided to use it for the evaluation of the sensing capabilities ofthe proposed smart sensor

93 Dynamic Displacements Using GPS Movements relatedto short period component are represented by dynamicdisplacements (see Figure 10)

Dynamic displacements are obtained applying a high-pass 13lter to GPS apparent displacements Figure 10 illus-trates dynamic displacements corresponding to the bridge Itcan be noted in Figures 10(a) and 10(b) that dynamicdisplacements are relatively small for both N-S and E-Whorizontal components On the contrary the vertical di-rection is presenting excessive displacements and this isillustrated in Figure 10(c) One more time critical dis-placements are appearing for the vertical component but inthis case with respect to dynamic displacements

Longitudinal direction

100 200 300 400 5000Time (sec)

Smartphone Accelerometer

ndash1

0

1

Acc

eler

atio

n (m

s2 )

(a)

Transversal directionSmartphone Accelerometer

ndash1

0

1A

ccel

erat

ion

(ms

2 )

100 200 300 400 5000

Time (sec)

(b)

Vertical directionSmartphone Accelerometer

100 200 300 400 5000

Time (sec)

ndash2

0

2

Acc

eler

atio

n (m

s2 )

(c)

Figure 8 Acceleration time history

Accelerometer

Acceleration

High-pass filterFFTPeriodogram

Detrend

t

0

Ri ndash (1n)sumni=1 Ri

t

0

Dynamicdisplacement

Semistaticdisplacement

Low-pass filterBand-Pass filter

Smartphone Displacementfrom GPS

Figure 7 Flow chart of measurement process

Advances in Civil Engineering 9

Apparent displacementSemistatic displacement

ndash001

0

001D

ispla

cem

ent (

m)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(a)

Apparent displacementSemistatic displacement

ndash005

0

005

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(b)

Apparent displacementSemistatic displacement

ndash001

0

001

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(c)

Figure 9 Semistatic displacements in three directions (a) N-S (b) E-W and (c) vertical

NndashS

ndash001

0

001

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(a)

EndashW

ndash005

0

005

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(b)

Figure 10 Continued

10 Advances in Civil Engineering

94 FFT and Periodogram Using GPS DynamicDisplacements To identify structural dynamic character-istics of the structure under consideration FFT and perio-dograms are calculated and plotted in frequency domainusing dynamic displacements of GPS

Figure 11 illustrates FFT and periodogram in the threeGPS directions N-S (Figures 11(a) and 11(b)) E-W(Figures 11(c) and 11(d)) and vertical (Figures 11(e) and11(f)) To clarify the results they are normalized with respectto their maximum amplitude value

95 Acceleration Time History Using Accelerometer andSmartphone Acceleration time histories corresponding tothe longitudinal transversal and vertical direction of thebridge were earlier presented in Figure 8 considering bothsmartphone and accelerometer In the longitudinal direction(see Figure 8(a)) it can be noted that there are some dif-ferences in the acceleration reported by smartphone andaccelerometer such discrepancies were justi13ed earlier inSection 91 In the transversal (see Figure 8(b)) and verticaldirections (see Figure 8(c)) it is observed that smartphone

Vertical

ndash001

0

001D

ispla

cem

ent (

m)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(c)

Figure 10 Dynamic displacements in three directions (a) N-S (b) E-W and (c) vertical

NndashSFFT

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(a)

NndashSPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(b)

EndashWFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(c)

EndashWPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

VerticalFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(e)

VerticalPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 11 Dynamic frequencies in terms of normalized FFT and periodogram

Advances in Civil Engineering 11

and accelerometer are recording very similar results In timedomain these results are very promising However tocomplement them results are changed to the frequencydomain in terms of FFT and periodogram is is discussedin the next section

96 FFT and Periodogram Using Accelerometer and Smart-phone Accelerations In this section results are studied infrequency domain ey are presented in terms of FFT andperiodogram (see Figure 12)

Figure 12 summarizes the results related to FFT andperiodogram corresponding to longitudinal transversal andvertical directions It can be observed in Figure 12 that thefrequency that contributes the most to the structural re-sponse may be located approximately at 9Hz In addition itis noted that FFT in the longitudinal and transversal di-rection (see Figures 12(a) and 12(c)) is presenting the 13rstfrequency of interest at about 35Hz Based on the resultsplotted in Figure 12 it can be stated that both smartphoneand accelerometer are providing very accurate results with

respect to each other Hence it may result practical to usesmartphones for extracting dynamic characteristics of real-scale bridges

97 Comparison of Periodogram of GPS Accelerometers andSmartphones Using Dynamic Displacements To comparedata acquisition capabilities of smartphones accelerometersand GPS the periodogram of them with respect to dynamicdisplacements in the vertical direction is illustrated inFigure 13 Only vertical direction is studied because it wasdemonstrated that this direction presented a critical be-havior It is noted as reported in Figure 12 that the fre-quency that contributes the most to the structural responseis located approximately at 937Hz which is related toprevious acceleration spectra is frequency is well detectedby the three instruments Periodogram reported in Figure 13is plotted considering dynamic displacements of smart-phone accelerometer and GPS Hence the accelerationrecorded by the smartphone and accelerometer respectivelyis converted to dynamic displacement by carrying out a

SmartphoneAccelerometer

FFTmdashlongitudinal direction

0

05

1A

mpl

itude

5 10 15 200Frequency (Hz)

(a)

SmartphoneAccelerometer

Periodogrammdashlongitudinal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(b)

SmartphoneAccelerometer

FFTmdashtransversal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(c)

SmartphoneAccelerometer

Periodogrammdashtransversal direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

SmartphoneAccelerometer

FFTmdashvertical direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(e)

SmartphoneAccelerometer

Periodogrammdashvertical direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 12 Dynamic frequencies in terms of FFT and periodogram

12 Advances in Civil Engineering

double integration which may involve some discrepancies inobtaining displacements In addition to numerical in-tegration some of the dierences observed in Figure 13 maybe associated with the applied 13ltering techniques to obtaindynamic displacements On the other hand GPS reportsdisplacements directly improving considerably data pro-cessing and interpretations However it is considered im-portant to analyze the three periodograms from Figure 13 tohave a better understanding about the coincidences of somefrequencies that appear in two or more periodograms and itmay be useful to have a better idea about de dynamic be-havior of a bridge

10 Conclusions

An alternative methodology to perform SHM on real-scalebridges is presented in this paper using GPS an acceler-ometer and a smartphone integrated in a fused smartsensor Such sensor is validated for the SHM of complexstructures as bridges under normal serviceability conditionsin real time A reinforced concrete bridge was studied byusing the proposed smart sensor because it presented ex-cessive vibration problems e accuracy of this smart sensorwas demonstrated comparing the gathered data from GPSaccelerometers and smartphones in both time and fre-quency domain Semistatic and dynamic displacements aswell as accelerations demonstrated that there are criticalperformance problems on the analyzed bridge particularlyin the vertical direction Since this critical behavior wasreported by other scholars that performed the SHM withdierent devices it can be stated that this smart sensor canbe implemented for SHM on bridges in real time In ad-dition it was demonstrated that the proposed smart sensor isadequate to perform SHM in real-scale bridges Conse-quently for the bridge under consideration it was calculated

that the fundamental frequency in the horizontal direction islocated at 9Hz Furthermore it was detected the 13rst fre-quency of interest around 35Hz In order to improvefrequency domain analysis the proposed smart sensormethodology performs a periodogram algorithm over thevertical direction data to obtain a more accurate SHM It wasobserved that there is a dierence between GPS displace-ment data and accelerometer-smartphone based data It canbe justi13ed due to the nature of each measurement meth-odology where GPS base data comes from atmosphere andaccelerometer-based data comes from a vibration sensorBecause of this the noise nature of the aforementionedmeasurement devices is not the same due to its dierentmeasurement principle On the other hand it was observed acoincidence of the three instruments approximately at937Hz for the vertical component us it can explain theusefulness of fusing GPS accelerometer and smartphoneinformation in order to discriminate undesired noise thatcan be generated by atmosphere disturbances of doubleintegration error from acceleration data when it appearsonly in one device data Finally it can be stated that thispaper contributed in understanding how GPS accelerom-eters and smartphones can be successfully fused in a smartsensor to perform SHM on real-scale bridges

Data Availability

e data used to support the 13ndings of this study areavailable from the corresponding author upon request

Disclosure

Any opinions 13ndings conclusions or recommendationsexpressed in this paper are those of the authors and do notnecessarily reect the views of the sponsors

Conflicts of Interest

e authors do not have any conicts of interest to declareabout this research

Acknowledgments

e authors would like to thank Septentrio Inc for pro-viding the AsteRx-U Receiver and the PolaNt-x MF antennaused during the GPS data collection is study is based onwork partly supported by the Consejo Nacional de Ciencia yTecnologıa (CONACYT-Mexico) as scholarships for the 13rstand seventh author respectively In addition the work isalso supported by the Universidad Autonoma de Sinaloa(UAS)

References

[1] V Ashkenazi and G W Roberts ldquoExperimental monitoringof the Humber bridge using GPSrdquo Proceedings of the In-stitution of Civil Engineers-Civil Engineering vol 120 no 4pp 177ndash182 1997

[2] W S Chan Y L Xu X L Ding andW J Dai ldquoAn integratedGPSndashaccelerometer data processing technique for structural

1

09

08

07

06

05

04

03

02

01

00 5 10

Frequency (Hz)15 20

SmartphoneAccelerometerGPS (vertical)

Figure 13 Dynamic frequencies in terms of periodogram (verticaldirection)

Advances in Civil Engineering 13

deformation monitoringrdquo Journal of Geodesy vol 80 no 12pp 705ndash719 2006

[3] S B Im S Hurlebaus and Y J Kang ldquoSummary review ofGPS technology for structural health monitoringrdquo Journal ofStructural Engineering vol 139 no 10 pp 1653ndash1664 2011

[4] M R Kaloop and H Li ldquoMonitoring of bridge deformationusing GPS techniquerdquo KSCE Journal of Civil Engineeringvol 13 no 6 pp 423ndash431 2009

[5] M Kaloop E Elbeltagi J Hu and A Elrefai ldquoRecent advancesof structures monitoring and evaluation using GPS-time seriesmonitoring systems a reviewrdquo ISPRS International Journal ofGeo-Information vol 6 no 12 p 382 2017

[6] X Li C Rizos L Ge Y Tamura and A Yoshida ldquo+ecomplementary characteristics of GPS and accelerometer inmonitoring structural deformationrdquo in Proceedings of the ION2005 National Technical Meeting Institute of Navigation(ION) San Diego CA USA January 2005

[7] P Psimoulis S Pytharouli D Karambalis and S StirosldquoPotential of Global Positioning System (GPS) to measurefrequencies of oscillations of engineering structuresrdquo Journalof Sound and Vibration vol 318 no 3 pp 606ndash623 2008

[8] C Watson T Watson and R Coleman ldquoStructural moni-toring of cable-stayed bridge analysis of GPS versus modeleddeflectionsrdquo Journal of Surveying Engineering vol 133 no 1pp 23ndash28 2007

[9] C Xiong and Y Niu ldquoInvestigation of the dynamic behavior ofa super high-rise structure using RTK-GNSS techniquerdquo KSCEJournal of Civil Engineering vol 23 no 2 pp 654ndash665 2018

[10] B Hofmann-Wellenhof H Lichtenegger and E WasleGNSSndashGlobal Navigation Satellite Systems GPS GLONASSGalileo and More Springer Science amp Business MediaVienna Austria 2007

[11] M Celebi ldquoGPS in dynamic monitoring of long-periodstructuresrdquo Soil Dynamics and Earthquake Engineeringvol 20 no 5ndash8 pp 477ndash483 2000

[12] D K Kim J I Kim and M Q Feng ldquoInstrumentation ofbridges for structural health monitoringrdquo KSCE Journal ofCivil Engineering vol 5 no 3 pp 231ndash242 2001

[13] G Q Wang D M Boore H Igel and X Y Zhou ldquoSomeobservations on colocated and closely spaced strong ground-motion records of the 1999 Chi-Chi Taiwan earthquakerdquoBulletin of the Seismological Society of America vol 93 no 2pp 674ndash693 2003

[14] J Hwang H Yun S K Park D Lee and S Hong ldquoOptimalmethods of RTK-GPSaccelerometer integration to monitorthe displacement of structuresrdquo Sensors vol 12 no 1pp 1014ndash1034 2012

[15] S Dashti J D Bray J Reilly S Glaser A Bayen and E MarildquoEvaluating the reliability of phones as seismic monitoringinstrumentsrdquo Earthquake Spectra vol 30 no 2 pp 721ndash7422014

[16] M Feng Y Fukuda M Mizuta and E Ozer ldquoCitizen sensorsfor SHM use of accelerometer data from smartphonesrdquoSensors vol 15 no 2 pp 2980ndash2998 2015

[17] X Zhao R Han Y Ding et al ldquoPortable and convenient cableforce measurement using smartphonerdquo Journal of CivilStructural Health Monitoring vol 5 no 4 pp 481ndash491 2015

[18] X Zhao K Ri R Han Y Yu M Li and J Ou ldquoExperimentalresearch on quick structural health monitoring technique forbridges using smartphonerdquoAdvances inMaterials Science andEngineering vol 2016 no 1 pp 1ndash14 2016

[19] Y Ding R Han H Liu S Li X Zhao and Y Yu ldquoBridgeinspection and management system based on smart phonerdquoin Proceedings of the ASME 2016 Conference on Smart

Materials Adaptive Structures and Intelligent SystemsAmerican Society of Mechanical Engineers Stowe VT USASeptember 2016

[20] X Zhao R Han K J Loh B Xie J Li and J Ou ldquoShakingtable tests for evaluating the damage features under earth-quake excitations using smartphonesrdquo in Proceedings of theHealth Monitoring of Structural and Biological Systems XIIInternational Society for Optics and Photonics Denver COUSA May 2018

[21] F Moschas and S Stiros ldquoMeasurement of the dynamicdisplacements and of the modal frequencies of a short-spanpedestrian bridge using GPS and an accelerometerrdquo Engi-neering Structures vol 33 no 1 pp 10ndash17 2011

[22] M R Kaloop ldquoBridge safety monitoring based-GPS tech-nique case study Zhujiang Huangpu bridgerdquo Smart Struc-tures and Systems vol 9 no 6 pp 473ndash487 2012

[23] G W Roberts X Meng and A H Dodson ldquoIntegrating aglobal positioning system and accelerometers to monitor thedeflection of bridgesrdquo Journal of Surveying Engineeringvol 130 no 2 pp 65ndash72 2004

[24] Accelerometer Analyzer App 2017 httpsplaygooglecomstoreappsdetailsidcomlulaccelerometeramphles

[25] X Meng ldquoReal-time deformationmonitoring of bridges usingGPSaccelerometersrdquo Doctoral dissertation University ofNottingham UK 2002

[26] G E Vazquez-Becerra J R Gaxiola-Camacho R BennettG M Guzman-Acevedo and I E Gaxiola-CamacholdquoStructural evaluation of dynamic and semi-static displace-ments of the Juarez bridge using GPS technologyrdquo Mea-surement vol 110 no 1 pp 146ndash153 2017

[27] P Psimoulis I Peppa L Bonenberg S Ince and X MengldquoCombination of GPS and RTS measurements for themonitoring of semi-static and dynamic motion of pedestrianbridgerdquo in Proceedings of the 3rd Joint International Sym-posium on Deformation Monitoring Vienna Austria April2016

[28] F Moschas and S Stiros ldquoNoise characteristics of high-frequencyshort-duration GPS records from analysis ofidentical collocated instrumentsrdquo Measurement vol 46no 4 pp 1488ndash1506 2013

[29] S W Smithgte Scientist amp Engineerrsquos Guide to Digital SignalProcessing California Technical Publishing San Diego CAUSA 1997

[30] J Wang X Meng C Qin and J Yi ldquoVibration frequenciesextraction of the forth road bridge using high sampling GPSdatardquo Shock and Vibration vol 2016 Article ID 980786118 pages 2016

[31] J Yu XMeng X Shao B Yan and L Yang ldquoIdentification ofdynamic displacements and modal frequencies of a medium-span suspension bridge using multimode GNSS processingrdquoEngineering Structures vol 81 no 1 pp 432ndash443 2014

[32] E Elbeltagi M Kaloop and M Elnabwy ldquoMonitoring andassessment of Mansoura railway steel bridge using RTK-GPStechniquerdquo in Proceedings of the 2nd Joint InternationalSymposium on Deformation Monitoring pp 9ndash11 Notting-ham England September 2013

[33] N Raziq ldquoGPS structural deformation monitoring the mid-height problemrdquo Doctoral dissertation University ofMelbourneAustralia 2008

[34] S S HaykinAdaptive Filtergteory Pearson Education India5th edition 2014

[35] L H Estey and C M Meertens ldquoTEQC the multi-purposetoolkit for GPSGLONASS datardquo GPS Solutions vol 3 no 1pp 42ndash49 1999

14 Advances in Civil Engineering

recording acceleration at a rate of 100Hz+e accelerometerwas perfectly fixed to present the same physical center as theother devices to record the behavior of ZEN2 stationFigure 6(b) illustrates the smart sensor that is holding thethree devices One more important fact in the accelerometercampaign is that information was recorded at the same timethan GPS and smartphone providing a correct synchroni-zation [37]

83 SmartphoneCampaign As illustrated earlier in Figure 4the smartphone was fixed at the lowest plate of the mechanicstructure for the smart sensor +e measured informationfrom the accelerometer embedded in the smartphone wascollected by an open access App [24] +is application canrecord acceleration and then send it to the user in the form oftext file+e smartphone collected data at the same time thanthe GPS receiver and commercial accelerometer during500 seconds with a sampling rate of 400Hz

84 Flow Chart Diagram of the Measurement Process Insummary the data acquisition process of the bridge usingthe smart sensor is illustrated in Figure 7 Firstly theaccelerometer smartphone and GPS antenna are perfectlyfixed to the location of ZEN2 station +e particularissues related to this location were previously discussedSecondly acceleration time histories are extracted fromaccelerometer and smartphone respectively and dis-placements from GPS receivers As documented earlier thedata acquisition process is performed simultaneously usingthe three instruments +irdly the apparent displacementin three directions (N-S E-W and vertical) is calculatedusing GPS +en semistatic and dynamic displacementsare extracted using low- and high-pass filters respectivelyFourthly in order to study the dynamic characteristics ofthe Juarez Bridge FFTand periodogram are plotted for thelongitudinal transversal and vertical directions using rawacceleration of smartphone and accelerometer re-spectively Finally to compare the extracting responsecapabilities of the three instruments dynamic displace-ments reported by every device are used for the calculationof periodograms +e above discussion is clarified in thenext section

9 Structural Response Collected by theSmart Sensor

+e structural performance of the Juarez Bridge is studiedunder normal service loading conditions in real time ie thestructure is evaluated during a loading condition that is ex-perienced daily by users In fact a nonpermissible vibration inthe vertical direction has been reported by users and it hasbeen documented in the literature by Vazquez-Becerra et al[26] To verify the undesirable behavior of this bridge in thisstudy the authors decided to implement other devices duringthe SHM Hence the collected structural response is obtainedsimultaneously using GPS accelerometer and smartphone inreal time as a fused smart sensor

91 Acceleration TimeHistory One measurement session of500 seconds was performed using the three instrumentsunder consideration In terms of collected accelerationFigure 8 illustrates the recorded response of the smartphoneand accelerometer respectively Figures 8(a)ndash8(c) illustratethe longitudinal transversal and vertical acceleration re-spectively Several observations can be made correspondingto Figure 8 It can be observed that acceleration in thevertical direction is greater than longitudinal and transversaldirections In terms of displacements the same observationwas reported by [26] In addition it can be noted inFigure 8(a) that in the longitudinal direction the acceler-ation reported by smartphone and accelerometer is pre-senting some discrepancies +is issue may be justifiedbecause of the precision of commercial accelerometers insuch a direction However this structural response does notrepresent a threat of safety because of the stiffness of thebridge in the longitudinal direction In addition it can beobserved in Figure 8(c) that the critical component is as-sociated to accelerations acting in the vertical directionConversely in terms of accuracy in Figures 8(b) and 8(c) isshown that acceleration for both smartphone and acceler-ometer is matching very closely In general terms Figure 8demonstrates that acceleration recorded by smartphone andaccelerometer can be positively comparable

92 Semistatic Displacements Using GPS Considering dis-placements obtained via GPS semistatic displacements are

(a) (b) (c)

Figure 6 GPS campaign (a) location of ZEN2 and 373H (b) station ZEN2 (c) station 373H

8 Advances in Civil Engineering

extracted by applying a low-pass 13lter Semistatic dis-placements corresponding to the bridge under considerationrepresent slowmovements of the structure they are reportedin the N-S E-W and vertical direction in Figures 9(a)ndash9(c)respectively

It can be observed in Figures 9(a) and 9(b) that smallerdisplacements are present in both N-S and E-W horizontaldirections than those in the vertical one is is illustrated inFigure 9(c) it is noted that considerable great displacementsare reported in the vertical direction Hence in terms ofsemistatic displacements it can be justi13ed that the verticalcomponent is presenting critical behavior ere exist casesof bridge studies with longer dimensions such as the SanFrancisco-Oakland bay bridge that presented small dis-placements [38] However the analyzed bridge presented inthis paper is a very particular case of study whichreports unusual behavior with respect to the vertical di-rection comparing semistatic displacement with AASHTO

deection limits atrsquos why the authors of this paper de-cided to use it for the evaluation of the sensing capabilities ofthe proposed smart sensor

93 Dynamic Displacements Using GPS Movements relatedto short period component are represented by dynamicdisplacements (see Figure 10)

Dynamic displacements are obtained applying a high-pass 13lter to GPS apparent displacements Figure 10 illus-trates dynamic displacements corresponding to the bridge Itcan be noted in Figures 10(a) and 10(b) that dynamicdisplacements are relatively small for both N-S and E-Whorizontal components On the contrary the vertical di-rection is presenting excessive displacements and this isillustrated in Figure 10(c) One more time critical dis-placements are appearing for the vertical component but inthis case with respect to dynamic displacements

Longitudinal direction

100 200 300 400 5000Time (sec)

Smartphone Accelerometer

ndash1

0

1

Acc

eler

atio

n (m

s2 )

(a)

Transversal directionSmartphone Accelerometer

ndash1

0

1A

ccel

erat

ion

(ms

2 )

100 200 300 400 5000

Time (sec)

(b)

Vertical directionSmartphone Accelerometer

100 200 300 400 5000

Time (sec)

ndash2

0

2

Acc

eler

atio

n (m

s2 )

(c)

Figure 8 Acceleration time history

Accelerometer

Acceleration

High-pass filterFFTPeriodogram

Detrend

t

0

Ri ndash (1n)sumni=1 Ri

t

0

Dynamicdisplacement

Semistaticdisplacement

Low-pass filterBand-Pass filter

Smartphone Displacementfrom GPS

Figure 7 Flow chart of measurement process

Advances in Civil Engineering 9

Apparent displacementSemistatic displacement

ndash001

0

001D

ispla

cem

ent (

m)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(a)

Apparent displacementSemistatic displacement

ndash005

0

005

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(b)

Apparent displacementSemistatic displacement

ndash001

0

001

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(c)

Figure 9 Semistatic displacements in three directions (a) N-S (b) E-W and (c) vertical

NndashS

ndash001

0

001

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(a)

EndashW

ndash005

0

005

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(b)

Figure 10 Continued

10 Advances in Civil Engineering

94 FFT and Periodogram Using GPS DynamicDisplacements To identify structural dynamic character-istics of the structure under consideration FFT and perio-dograms are calculated and plotted in frequency domainusing dynamic displacements of GPS

Figure 11 illustrates FFT and periodogram in the threeGPS directions N-S (Figures 11(a) and 11(b)) E-W(Figures 11(c) and 11(d)) and vertical (Figures 11(e) and11(f)) To clarify the results they are normalized with respectto their maximum amplitude value

95 Acceleration Time History Using Accelerometer andSmartphone Acceleration time histories corresponding tothe longitudinal transversal and vertical direction of thebridge were earlier presented in Figure 8 considering bothsmartphone and accelerometer In the longitudinal direction(see Figure 8(a)) it can be noted that there are some dif-ferences in the acceleration reported by smartphone andaccelerometer such discrepancies were justi13ed earlier inSection 91 In the transversal (see Figure 8(b)) and verticaldirections (see Figure 8(c)) it is observed that smartphone

Vertical

ndash001

0

001D

ispla

cem

ent (

m)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(c)

Figure 10 Dynamic displacements in three directions (a) N-S (b) E-W and (c) vertical

NndashSFFT

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(a)

NndashSPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(b)

EndashWFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(c)

EndashWPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

VerticalFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(e)

VerticalPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 11 Dynamic frequencies in terms of normalized FFT and periodogram

Advances in Civil Engineering 11

and accelerometer are recording very similar results In timedomain these results are very promising However tocomplement them results are changed to the frequencydomain in terms of FFT and periodogram is is discussedin the next section

96 FFT and Periodogram Using Accelerometer and Smart-phone Accelerations In this section results are studied infrequency domain ey are presented in terms of FFT andperiodogram (see Figure 12)

Figure 12 summarizes the results related to FFT andperiodogram corresponding to longitudinal transversal andvertical directions It can be observed in Figure 12 that thefrequency that contributes the most to the structural re-sponse may be located approximately at 9Hz In addition itis noted that FFT in the longitudinal and transversal di-rection (see Figures 12(a) and 12(c)) is presenting the 13rstfrequency of interest at about 35Hz Based on the resultsplotted in Figure 12 it can be stated that both smartphoneand accelerometer are providing very accurate results with

respect to each other Hence it may result practical to usesmartphones for extracting dynamic characteristics of real-scale bridges

97 Comparison of Periodogram of GPS Accelerometers andSmartphones Using Dynamic Displacements To comparedata acquisition capabilities of smartphones accelerometersand GPS the periodogram of them with respect to dynamicdisplacements in the vertical direction is illustrated inFigure 13 Only vertical direction is studied because it wasdemonstrated that this direction presented a critical be-havior It is noted as reported in Figure 12 that the fre-quency that contributes the most to the structural responseis located approximately at 937Hz which is related toprevious acceleration spectra is frequency is well detectedby the three instruments Periodogram reported in Figure 13is plotted considering dynamic displacements of smart-phone accelerometer and GPS Hence the accelerationrecorded by the smartphone and accelerometer respectivelyis converted to dynamic displacement by carrying out a

SmartphoneAccelerometer

FFTmdashlongitudinal direction

0

05

1A

mpl

itude

5 10 15 200Frequency (Hz)

(a)

SmartphoneAccelerometer

Periodogrammdashlongitudinal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(b)

SmartphoneAccelerometer

FFTmdashtransversal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(c)

SmartphoneAccelerometer

Periodogrammdashtransversal direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

SmartphoneAccelerometer

FFTmdashvertical direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(e)

SmartphoneAccelerometer

Periodogrammdashvertical direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 12 Dynamic frequencies in terms of FFT and periodogram

12 Advances in Civil Engineering

double integration which may involve some discrepancies inobtaining displacements In addition to numerical in-tegration some of the dierences observed in Figure 13 maybe associated with the applied 13ltering techniques to obtaindynamic displacements On the other hand GPS reportsdisplacements directly improving considerably data pro-cessing and interpretations However it is considered im-portant to analyze the three periodograms from Figure 13 tohave a better understanding about the coincidences of somefrequencies that appear in two or more periodograms and itmay be useful to have a better idea about de dynamic be-havior of a bridge

10 Conclusions

An alternative methodology to perform SHM on real-scalebridges is presented in this paper using GPS an acceler-ometer and a smartphone integrated in a fused smartsensor Such sensor is validated for the SHM of complexstructures as bridges under normal serviceability conditionsin real time A reinforced concrete bridge was studied byusing the proposed smart sensor because it presented ex-cessive vibration problems e accuracy of this smart sensorwas demonstrated comparing the gathered data from GPSaccelerometers and smartphones in both time and fre-quency domain Semistatic and dynamic displacements aswell as accelerations demonstrated that there are criticalperformance problems on the analyzed bridge particularlyin the vertical direction Since this critical behavior wasreported by other scholars that performed the SHM withdierent devices it can be stated that this smart sensor canbe implemented for SHM on bridges in real time In ad-dition it was demonstrated that the proposed smart sensor isadequate to perform SHM in real-scale bridges Conse-quently for the bridge under consideration it was calculated

that the fundamental frequency in the horizontal direction islocated at 9Hz Furthermore it was detected the 13rst fre-quency of interest around 35Hz In order to improvefrequency domain analysis the proposed smart sensormethodology performs a periodogram algorithm over thevertical direction data to obtain a more accurate SHM It wasobserved that there is a dierence between GPS displace-ment data and accelerometer-smartphone based data It canbe justi13ed due to the nature of each measurement meth-odology where GPS base data comes from atmosphere andaccelerometer-based data comes from a vibration sensorBecause of this the noise nature of the aforementionedmeasurement devices is not the same due to its dierentmeasurement principle On the other hand it was observed acoincidence of the three instruments approximately at937Hz for the vertical component us it can explain theusefulness of fusing GPS accelerometer and smartphoneinformation in order to discriminate undesired noise thatcan be generated by atmosphere disturbances of doubleintegration error from acceleration data when it appearsonly in one device data Finally it can be stated that thispaper contributed in understanding how GPS accelerom-eters and smartphones can be successfully fused in a smartsensor to perform SHM on real-scale bridges

Data Availability

e data used to support the 13ndings of this study areavailable from the corresponding author upon request

Disclosure

Any opinions 13ndings conclusions or recommendationsexpressed in this paper are those of the authors and do notnecessarily reect the views of the sponsors

Conflicts of Interest

e authors do not have any conicts of interest to declareabout this research

Acknowledgments

e authors would like to thank Septentrio Inc for pro-viding the AsteRx-U Receiver and the PolaNt-x MF antennaused during the GPS data collection is study is based onwork partly supported by the Consejo Nacional de Ciencia yTecnologıa (CONACYT-Mexico) as scholarships for the 13rstand seventh author respectively In addition the work isalso supported by the Universidad Autonoma de Sinaloa(UAS)

References

[1] V Ashkenazi and G W Roberts ldquoExperimental monitoringof the Humber bridge using GPSrdquo Proceedings of the In-stitution of Civil Engineers-Civil Engineering vol 120 no 4pp 177ndash182 1997

[2] W S Chan Y L Xu X L Ding andW J Dai ldquoAn integratedGPSndashaccelerometer data processing technique for structural

1

09

08

07

06

05

04

03

02

01

00 5 10

Frequency (Hz)15 20

SmartphoneAccelerometerGPS (vertical)

Figure 13 Dynamic frequencies in terms of periodogram (verticaldirection)

Advances in Civil Engineering 13

deformation monitoringrdquo Journal of Geodesy vol 80 no 12pp 705ndash719 2006

[3] S B Im S Hurlebaus and Y J Kang ldquoSummary review ofGPS technology for structural health monitoringrdquo Journal ofStructural Engineering vol 139 no 10 pp 1653ndash1664 2011

[4] M R Kaloop and H Li ldquoMonitoring of bridge deformationusing GPS techniquerdquo KSCE Journal of Civil Engineeringvol 13 no 6 pp 423ndash431 2009

[5] M Kaloop E Elbeltagi J Hu and A Elrefai ldquoRecent advancesof structures monitoring and evaluation using GPS-time seriesmonitoring systems a reviewrdquo ISPRS International Journal ofGeo-Information vol 6 no 12 p 382 2017

[6] X Li C Rizos L Ge Y Tamura and A Yoshida ldquo+ecomplementary characteristics of GPS and accelerometer inmonitoring structural deformationrdquo in Proceedings of the ION2005 National Technical Meeting Institute of Navigation(ION) San Diego CA USA January 2005

[7] P Psimoulis S Pytharouli D Karambalis and S StirosldquoPotential of Global Positioning System (GPS) to measurefrequencies of oscillations of engineering structuresrdquo Journalof Sound and Vibration vol 318 no 3 pp 606ndash623 2008

[8] C Watson T Watson and R Coleman ldquoStructural moni-toring of cable-stayed bridge analysis of GPS versus modeleddeflectionsrdquo Journal of Surveying Engineering vol 133 no 1pp 23ndash28 2007

[9] C Xiong and Y Niu ldquoInvestigation of the dynamic behavior ofa super high-rise structure using RTK-GNSS techniquerdquo KSCEJournal of Civil Engineering vol 23 no 2 pp 654ndash665 2018

[10] B Hofmann-Wellenhof H Lichtenegger and E WasleGNSSndashGlobal Navigation Satellite Systems GPS GLONASSGalileo and More Springer Science amp Business MediaVienna Austria 2007

[11] M Celebi ldquoGPS in dynamic monitoring of long-periodstructuresrdquo Soil Dynamics and Earthquake Engineeringvol 20 no 5ndash8 pp 477ndash483 2000

[12] D K Kim J I Kim and M Q Feng ldquoInstrumentation ofbridges for structural health monitoringrdquo KSCE Journal ofCivil Engineering vol 5 no 3 pp 231ndash242 2001

[13] G Q Wang D M Boore H Igel and X Y Zhou ldquoSomeobservations on colocated and closely spaced strong ground-motion records of the 1999 Chi-Chi Taiwan earthquakerdquoBulletin of the Seismological Society of America vol 93 no 2pp 674ndash693 2003

[14] J Hwang H Yun S K Park D Lee and S Hong ldquoOptimalmethods of RTK-GPSaccelerometer integration to monitorthe displacement of structuresrdquo Sensors vol 12 no 1pp 1014ndash1034 2012

[15] S Dashti J D Bray J Reilly S Glaser A Bayen and E MarildquoEvaluating the reliability of phones as seismic monitoringinstrumentsrdquo Earthquake Spectra vol 30 no 2 pp 721ndash7422014

[16] M Feng Y Fukuda M Mizuta and E Ozer ldquoCitizen sensorsfor SHM use of accelerometer data from smartphonesrdquoSensors vol 15 no 2 pp 2980ndash2998 2015

[17] X Zhao R Han Y Ding et al ldquoPortable and convenient cableforce measurement using smartphonerdquo Journal of CivilStructural Health Monitoring vol 5 no 4 pp 481ndash491 2015

[18] X Zhao K Ri R Han Y Yu M Li and J Ou ldquoExperimentalresearch on quick structural health monitoring technique forbridges using smartphonerdquoAdvances inMaterials Science andEngineering vol 2016 no 1 pp 1ndash14 2016

[19] Y Ding R Han H Liu S Li X Zhao and Y Yu ldquoBridgeinspection and management system based on smart phonerdquoin Proceedings of the ASME 2016 Conference on Smart

Materials Adaptive Structures and Intelligent SystemsAmerican Society of Mechanical Engineers Stowe VT USASeptember 2016

[20] X Zhao R Han K J Loh B Xie J Li and J Ou ldquoShakingtable tests for evaluating the damage features under earth-quake excitations using smartphonesrdquo in Proceedings of theHealth Monitoring of Structural and Biological Systems XIIInternational Society for Optics and Photonics Denver COUSA May 2018

[21] F Moschas and S Stiros ldquoMeasurement of the dynamicdisplacements and of the modal frequencies of a short-spanpedestrian bridge using GPS and an accelerometerrdquo Engi-neering Structures vol 33 no 1 pp 10ndash17 2011

[22] M R Kaloop ldquoBridge safety monitoring based-GPS tech-nique case study Zhujiang Huangpu bridgerdquo Smart Struc-tures and Systems vol 9 no 6 pp 473ndash487 2012

[23] G W Roberts X Meng and A H Dodson ldquoIntegrating aglobal positioning system and accelerometers to monitor thedeflection of bridgesrdquo Journal of Surveying Engineeringvol 130 no 2 pp 65ndash72 2004

[24] Accelerometer Analyzer App 2017 httpsplaygooglecomstoreappsdetailsidcomlulaccelerometeramphles

[25] X Meng ldquoReal-time deformationmonitoring of bridges usingGPSaccelerometersrdquo Doctoral dissertation University ofNottingham UK 2002

[26] G E Vazquez-Becerra J R Gaxiola-Camacho R BennettG M Guzman-Acevedo and I E Gaxiola-CamacholdquoStructural evaluation of dynamic and semi-static displace-ments of the Juarez bridge using GPS technologyrdquo Mea-surement vol 110 no 1 pp 146ndash153 2017

[27] P Psimoulis I Peppa L Bonenberg S Ince and X MengldquoCombination of GPS and RTS measurements for themonitoring of semi-static and dynamic motion of pedestrianbridgerdquo in Proceedings of the 3rd Joint International Sym-posium on Deformation Monitoring Vienna Austria April2016

[28] F Moschas and S Stiros ldquoNoise characteristics of high-frequencyshort-duration GPS records from analysis ofidentical collocated instrumentsrdquo Measurement vol 46no 4 pp 1488ndash1506 2013

[29] S W Smithgte Scientist amp Engineerrsquos Guide to Digital SignalProcessing California Technical Publishing San Diego CAUSA 1997

[30] J Wang X Meng C Qin and J Yi ldquoVibration frequenciesextraction of the forth road bridge using high sampling GPSdatardquo Shock and Vibration vol 2016 Article ID 980786118 pages 2016

[31] J Yu XMeng X Shao B Yan and L Yang ldquoIdentification ofdynamic displacements and modal frequencies of a medium-span suspension bridge using multimode GNSS processingrdquoEngineering Structures vol 81 no 1 pp 432ndash443 2014

[32] E Elbeltagi M Kaloop and M Elnabwy ldquoMonitoring andassessment of Mansoura railway steel bridge using RTK-GPStechniquerdquo in Proceedings of the 2nd Joint InternationalSymposium on Deformation Monitoring pp 9ndash11 Notting-ham England September 2013

[33] N Raziq ldquoGPS structural deformation monitoring the mid-height problemrdquo Doctoral dissertation University ofMelbourneAustralia 2008

[34] S S HaykinAdaptive Filtergteory Pearson Education India5th edition 2014

[35] L H Estey and C M Meertens ldquoTEQC the multi-purposetoolkit for GPSGLONASS datardquo GPS Solutions vol 3 no 1pp 42ndash49 1999

14 Advances in Civil Engineering

extracted by applying a low-pass 13lter Semistatic dis-placements corresponding to the bridge under considerationrepresent slowmovements of the structure they are reportedin the N-S E-W and vertical direction in Figures 9(a)ndash9(c)respectively

It can be observed in Figures 9(a) and 9(b) that smallerdisplacements are present in both N-S and E-W horizontaldirections than those in the vertical one is is illustrated inFigure 9(c) it is noted that considerable great displacementsare reported in the vertical direction Hence in terms ofsemistatic displacements it can be justi13ed that the verticalcomponent is presenting critical behavior ere exist casesof bridge studies with longer dimensions such as the SanFrancisco-Oakland bay bridge that presented small dis-placements [38] However the analyzed bridge presented inthis paper is a very particular case of study whichreports unusual behavior with respect to the vertical di-rection comparing semistatic displacement with AASHTO

deection limits atrsquos why the authors of this paper de-cided to use it for the evaluation of the sensing capabilities ofthe proposed smart sensor

93 Dynamic Displacements Using GPS Movements relatedto short period component are represented by dynamicdisplacements (see Figure 10)

Dynamic displacements are obtained applying a high-pass 13lter to GPS apparent displacements Figure 10 illus-trates dynamic displacements corresponding to the bridge Itcan be noted in Figures 10(a) and 10(b) that dynamicdisplacements are relatively small for both N-S and E-Whorizontal components On the contrary the vertical di-rection is presenting excessive displacements and this isillustrated in Figure 10(c) One more time critical dis-placements are appearing for the vertical component but inthis case with respect to dynamic displacements

Longitudinal direction

100 200 300 400 5000Time (sec)

Smartphone Accelerometer

ndash1

0

1

Acc

eler

atio

n (m

s2 )

(a)

Transversal directionSmartphone Accelerometer

ndash1

0

1A

ccel

erat

ion

(ms

2 )

100 200 300 400 5000

Time (sec)

(b)

Vertical directionSmartphone Accelerometer

100 200 300 400 5000

Time (sec)

ndash2

0

2

Acc

eler

atio

n (m

s2 )

(c)

Figure 8 Acceleration time history

Accelerometer

Acceleration

High-pass filterFFTPeriodogram

Detrend

t

0

Ri ndash (1n)sumni=1 Ri

t

0

Dynamicdisplacement

Semistaticdisplacement

Low-pass filterBand-Pass filter

Smartphone Displacementfrom GPS

Figure 7 Flow chart of measurement process

Advances in Civil Engineering 9

Apparent displacementSemistatic displacement

ndash001

0

001D

ispla

cem

ent (

m)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(a)

Apparent displacementSemistatic displacement

ndash005

0

005

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(b)

Apparent displacementSemistatic displacement

ndash001

0

001

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(c)

Figure 9 Semistatic displacements in three directions (a) N-S (b) E-W and (c) vertical

NndashS

ndash001

0

001

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(a)

EndashW

ndash005

0

005

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(b)

Figure 10 Continued

10 Advances in Civil Engineering

94 FFT and Periodogram Using GPS DynamicDisplacements To identify structural dynamic character-istics of the structure under consideration FFT and perio-dograms are calculated and plotted in frequency domainusing dynamic displacements of GPS

Figure 11 illustrates FFT and periodogram in the threeGPS directions N-S (Figures 11(a) and 11(b)) E-W(Figures 11(c) and 11(d)) and vertical (Figures 11(e) and11(f)) To clarify the results they are normalized with respectto their maximum amplitude value

95 Acceleration Time History Using Accelerometer andSmartphone Acceleration time histories corresponding tothe longitudinal transversal and vertical direction of thebridge were earlier presented in Figure 8 considering bothsmartphone and accelerometer In the longitudinal direction(see Figure 8(a)) it can be noted that there are some dif-ferences in the acceleration reported by smartphone andaccelerometer such discrepancies were justi13ed earlier inSection 91 In the transversal (see Figure 8(b)) and verticaldirections (see Figure 8(c)) it is observed that smartphone

Vertical

ndash001

0

001D

ispla

cem

ent (

m)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(c)

Figure 10 Dynamic displacements in three directions (a) N-S (b) E-W and (c) vertical

NndashSFFT

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(a)

NndashSPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(b)

EndashWFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(c)

EndashWPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

VerticalFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(e)

VerticalPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 11 Dynamic frequencies in terms of normalized FFT and periodogram

Advances in Civil Engineering 11

and accelerometer are recording very similar results In timedomain these results are very promising However tocomplement them results are changed to the frequencydomain in terms of FFT and periodogram is is discussedin the next section

96 FFT and Periodogram Using Accelerometer and Smart-phone Accelerations In this section results are studied infrequency domain ey are presented in terms of FFT andperiodogram (see Figure 12)

Figure 12 summarizes the results related to FFT andperiodogram corresponding to longitudinal transversal andvertical directions It can be observed in Figure 12 that thefrequency that contributes the most to the structural re-sponse may be located approximately at 9Hz In addition itis noted that FFT in the longitudinal and transversal di-rection (see Figures 12(a) and 12(c)) is presenting the 13rstfrequency of interest at about 35Hz Based on the resultsplotted in Figure 12 it can be stated that both smartphoneand accelerometer are providing very accurate results with

respect to each other Hence it may result practical to usesmartphones for extracting dynamic characteristics of real-scale bridges

97 Comparison of Periodogram of GPS Accelerometers andSmartphones Using Dynamic Displacements To comparedata acquisition capabilities of smartphones accelerometersand GPS the periodogram of them with respect to dynamicdisplacements in the vertical direction is illustrated inFigure 13 Only vertical direction is studied because it wasdemonstrated that this direction presented a critical be-havior It is noted as reported in Figure 12 that the fre-quency that contributes the most to the structural responseis located approximately at 937Hz which is related toprevious acceleration spectra is frequency is well detectedby the three instruments Periodogram reported in Figure 13is plotted considering dynamic displacements of smart-phone accelerometer and GPS Hence the accelerationrecorded by the smartphone and accelerometer respectivelyis converted to dynamic displacement by carrying out a

SmartphoneAccelerometer

FFTmdashlongitudinal direction

0

05

1A

mpl

itude

5 10 15 200Frequency (Hz)

(a)

SmartphoneAccelerometer

Periodogrammdashlongitudinal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(b)

SmartphoneAccelerometer

FFTmdashtransversal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(c)

SmartphoneAccelerometer

Periodogrammdashtransversal direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

SmartphoneAccelerometer

FFTmdashvertical direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(e)

SmartphoneAccelerometer

Periodogrammdashvertical direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 12 Dynamic frequencies in terms of FFT and periodogram

12 Advances in Civil Engineering

double integration which may involve some discrepancies inobtaining displacements In addition to numerical in-tegration some of the dierences observed in Figure 13 maybe associated with the applied 13ltering techniques to obtaindynamic displacements On the other hand GPS reportsdisplacements directly improving considerably data pro-cessing and interpretations However it is considered im-portant to analyze the three periodograms from Figure 13 tohave a better understanding about the coincidences of somefrequencies that appear in two or more periodograms and itmay be useful to have a better idea about de dynamic be-havior of a bridge

10 Conclusions

An alternative methodology to perform SHM on real-scalebridges is presented in this paper using GPS an acceler-ometer and a smartphone integrated in a fused smartsensor Such sensor is validated for the SHM of complexstructures as bridges under normal serviceability conditionsin real time A reinforced concrete bridge was studied byusing the proposed smart sensor because it presented ex-cessive vibration problems e accuracy of this smart sensorwas demonstrated comparing the gathered data from GPSaccelerometers and smartphones in both time and fre-quency domain Semistatic and dynamic displacements aswell as accelerations demonstrated that there are criticalperformance problems on the analyzed bridge particularlyin the vertical direction Since this critical behavior wasreported by other scholars that performed the SHM withdierent devices it can be stated that this smart sensor canbe implemented for SHM on bridges in real time In ad-dition it was demonstrated that the proposed smart sensor isadequate to perform SHM in real-scale bridges Conse-quently for the bridge under consideration it was calculated

that the fundamental frequency in the horizontal direction islocated at 9Hz Furthermore it was detected the 13rst fre-quency of interest around 35Hz In order to improvefrequency domain analysis the proposed smart sensormethodology performs a periodogram algorithm over thevertical direction data to obtain a more accurate SHM It wasobserved that there is a dierence between GPS displace-ment data and accelerometer-smartphone based data It canbe justi13ed due to the nature of each measurement meth-odology where GPS base data comes from atmosphere andaccelerometer-based data comes from a vibration sensorBecause of this the noise nature of the aforementionedmeasurement devices is not the same due to its dierentmeasurement principle On the other hand it was observed acoincidence of the three instruments approximately at937Hz for the vertical component us it can explain theusefulness of fusing GPS accelerometer and smartphoneinformation in order to discriminate undesired noise thatcan be generated by atmosphere disturbances of doubleintegration error from acceleration data when it appearsonly in one device data Finally it can be stated that thispaper contributed in understanding how GPS accelerom-eters and smartphones can be successfully fused in a smartsensor to perform SHM on real-scale bridges

Data Availability

e data used to support the 13ndings of this study areavailable from the corresponding author upon request

Disclosure

Any opinions 13ndings conclusions or recommendationsexpressed in this paper are those of the authors and do notnecessarily reect the views of the sponsors

Conflicts of Interest

e authors do not have any conicts of interest to declareabout this research

Acknowledgments

e authors would like to thank Septentrio Inc for pro-viding the AsteRx-U Receiver and the PolaNt-x MF antennaused during the GPS data collection is study is based onwork partly supported by the Consejo Nacional de Ciencia yTecnologıa (CONACYT-Mexico) as scholarships for the 13rstand seventh author respectively In addition the work isalso supported by the Universidad Autonoma de Sinaloa(UAS)

References

[1] V Ashkenazi and G W Roberts ldquoExperimental monitoringof the Humber bridge using GPSrdquo Proceedings of the In-stitution of Civil Engineers-Civil Engineering vol 120 no 4pp 177ndash182 1997

[2] W S Chan Y L Xu X L Ding andW J Dai ldquoAn integratedGPSndashaccelerometer data processing technique for structural

1

09

08

07

06

05

04

03

02

01

00 5 10

Frequency (Hz)15 20

SmartphoneAccelerometerGPS (vertical)

Figure 13 Dynamic frequencies in terms of periodogram (verticaldirection)

Advances in Civil Engineering 13

deformation monitoringrdquo Journal of Geodesy vol 80 no 12pp 705ndash719 2006

[3] S B Im S Hurlebaus and Y J Kang ldquoSummary review ofGPS technology for structural health monitoringrdquo Journal ofStructural Engineering vol 139 no 10 pp 1653ndash1664 2011

[4] M R Kaloop and H Li ldquoMonitoring of bridge deformationusing GPS techniquerdquo KSCE Journal of Civil Engineeringvol 13 no 6 pp 423ndash431 2009

[5] M Kaloop E Elbeltagi J Hu and A Elrefai ldquoRecent advancesof structures monitoring and evaluation using GPS-time seriesmonitoring systems a reviewrdquo ISPRS International Journal ofGeo-Information vol 6 no 12 p 382 2017

[6] X Li C Rizos L Ge Y Tamura and A Yoshida ldquo+ecomplementary characteristics of GPS and accelerometer inmonitoring structural deformationrdquo in Proceedings of the ION2005 National Technical Meeting Institute of Navigation(ION) San Diego CA USA January 2005

[7] P Psimoulis S Pytharouli D Karambalis and S StirosldquoPotential of Global Positioning System (GPS) to measurefrequencies of oscillations of engineering structuresrdquo Journalof Sound and Vibration vol 318 no 3 pp 606ndash623 2008

[8] C Watson T Watson and R Coleman ldquoStructural moni-toring of cable-stayed bridge analysis of GPS versus modeleddeflectionsrdquo Journal of Surveying Engineering vol 133 no 1pp 23ndash28 2007

[9] C Xiong and Y Niu ldquoInvestigation of the dynamic behavior ofa super high-rise structure using RTK-GNSS techniquerdquo KSCEJournal of Civil Engineering vol 23 no 2 pp 654ndash665 2018

[10] B Hofmann-Wellenhof H Lichtenegger and E WasleGNSSndashGlobal Navigation Satellite Systems GPS GLONASSGalileo and More Springer Science amp Business MediaVienna Austria 2007

[11] M Celebi ldquoGPS in dynamic monitoring of long-periodstructuresrdquo Soil Dynamics and Earthquake Engineeringvol 20 no 5ndash8 pp 477ndash483 2000

[12] D K Kim J I Kim and M Q Feng ldquoInstrumentation ofbridges for structural health monitoringrdquo KSCE Journal ofCivil Engineering vol 5 no 3 pp 231ndash242 2001

[13] G Q Wang D M Boore H Igel and X Y Zhou ldquoSomeobservations on colocated and closely spaced strong ground-motion records of the 1999 Chi-Chi Taiwan earthquakerdquoBulletin of the Seismological Society of America vol 93 no 2pp 674ndash693 2003

[14] J Hwang H Yun S K Park D Lee and S Hong ldquoOptimalmethods of RTK-GPSaccelerometer integration to monitorthe displacement of structuresrdquo Sensors vol 12 no 1pp 1014ndash1034 2012

[15] S Dashti J D Bray J Reilly S Glaser A Bayen and E MarildquoEvaluating the reliability of phones as seismic monitoringinstrumentsrdquo Earthquake Spectra vol 30 no 2 pp 721ndash7422014

[16] M Feng Y Fukuda M Mizuta and E Ozer ldquoCitizen sensorsfor SHM use of accelerometer data from smartphonesrdquoSensors vol 15 no 2 pp 2980ndash2998 2015

[17] X Zhao R Han Y Ding et al ldquoPortable and convenient cableforce measurement using smartphonerdquo Journal of CivilStructural Health Monitoring vol 5 no 4 pp 481ndash491 2015

[18] X Zhao K Ri R Han Y Yu M Li and J Ou ldquoExperimentalresearch on quick structural health monitoring technique forbridges using smartphonerdquoAdvances inMaterials Science andEngineering vol 2016 no 1 pp 1ndash14 2016

[19] Y Ding R Han H Liu S Li X Zhao and Y Yu ldquoBridgeinspection and management system based on smart phonerdquoin Proceedings of the ASME 2016 Conference on Smart

Materials Adaptive Structures and Intelligent SystemsAmerican Society of Mechanical Engineers Stowe VT USASeptember 2016

[20] X Zhao R Han K J Loh B Xie J Li and J Ou ldquoShakingtable tests for evaluating the damage features under earth-quake excitations using smartphonesrdquo in Proceedings of theHealth Monitoring of Structural and Biological Systems XIIInternational Society for Optics and Photonics Denver COUSA May 2018

[21] F Moschas and S Stiros ldquoMeasurement of the dynamicdisplacements and of the modal frequencies of a short-spanpedestrian bridge using GPS and an accelerometerrdquo Engi-neering Structures vol 33 no 1 pp 10ndash17 2011

[22] M R Kaloop ldquoBridge safety monitoring based-GPS tech-nique case study Zhujiang Huangpu bridgerdquo Smart Struc-tures and Systems vol 9 no 6 pp 473ndash487 2012

[23] G W Roberts X Meng and A H Dodson ldquoIntegrating aglobal positioning system and accelerometers to monitor thedeflection of bridgesrdquo Journal of Surveying Engineeringvol 130 no 2 pp 65ndash72 2004

[24] Accelerometer Analyzer App 2017 httpsplaygooglecomstoreappsdetailsidcomlulaccelerometeramphles

[25] X Meng ldquoReal-time deformationmonitoring of bridges usingGPSaccelerometersrdquo Doctoral dissertation University ofNottingham UK 2002

[26] G E Vazquez-Becerra J R Gaxiola-Camacho R BennettG M Guzman-Acevedo and I E Gaxiola-CamacholdquoStructural evaluation of dynamic and semi-static displace-ments of the Juarez bridge using GPS technologyrdquo Mea-surement vol 110 no 1 pp 146ndash153 2017

[27] P Psimoulis I Peppa L Bonenberg S Ince and X MengldquoCombination of GPS and RTS measurements for themonitoring of semi-static and dynamic motion of pedestrianbridgerdquo in Proceedings of the 3rd Joint International Sym-posium on Deformation Monitoring Vienna Austria April2016

[28] F Moschas and S Stiros ldquoNoise characteristics of high-frequencyshort-duration GPS records from analysis ofidentical collocated instrumentsrdquo Measurement vol 46no 4 pp 1488ndash1506 2013

[29] S W Smithgte Scientist amp Engineerrsquos Guide to Digital SignalProcessing California Technical Publishing San Diego CAUSA 1997

[30] J Wang X Meng C Qin and J Yi ldquoVibration frequenciesextraction of the forth road bridge using high sampling GPSdatardquo Shock and Vibration vol 2016 Article ID 980786118 pages 2016

[31] J Yu XMeng X Shao B Yan and L Yang ldquoIdentification ofdynamic displacements and modal frequencies of a medium-span suspension bridge using multimode GNSS processingrdquoEngineering Structures vol 81 no 1 pp 432ndash443 2014

[32] E Elbeltagi M Kaloop and M Elnabwy ldquoMonitoring andassessment of Mansoura railway steel bridge using RTK-GPStechniquerdquo in Proceedings of the 2nd Joint InternationalSymposium on Deformation Monitoring pp 9ndash11 Notting-ham England September 2013

[33] N Raziq ldquoGPS structural deformation monitoring the mid-height problemrdquo Doctoral dissertation University ofMelbourneAustralia 2008

[34] S S HaykinAdaptive Filtergteory Pearson Education India5th edition 2014

[35] L H Estey and C M Meertens ldquoTEQC the multi-purposetoolkit for GPSGLONASS datardquo GPS Solutions vol 3 no 1pp 42ndash49 1999

14 Advances in Civil Engineering

Apparent displacementSemistatic displacement

ndash001

0

001D

ispla

cem

ent (

m)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(a)

Apparent displacementSemistatic displacement

ndash005

0

005

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(b)

Apparent displacementSemistatic displacement

ndash001

0

001

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(c)

Figure 9 Semistatic displacements in three directions (a) N-S (b) E-W and (c) vertical

NndashS

ndash001

0

001

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(a)

EndashW

ndash005

0

005

Disp

lace

men

t (m

)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(b)

Figure 10 Continued

10 Advances in Civil Engineering

94 FFT and Periodogram Using GPS DynamicDisplacements To identify structural dynamic character-istics of the structure under consideration FFT and perio-dograms are calculated and plotted in frequency domainusing dynamic displacements of GPS

Figure 11 illustrates FFT and periodogram in the threeGPS directions N-S (Figures 11(a) and 11(b)) E-W(Figures 11(c) and 11(d)) and vertical (Figures 11(e) and11(f)) To clarify the results they are normalized with respectto their maximum amplitude value

95 Acceleration Time History Using Accelerometer andSmartphone Acceleration time histories corresponding tothe longitudinal transversal and vertical direction of thebridge were earlier presented in Figure 8 considering bothsmartphone and accelerometer In the longitudinal direction(see Figure 8(a)) it can be noted that there are some dif-ferences in the acceleration reported by smartphone andaccelerometer such discrepancies were justi13ed earlier inSection 91 In the transversal (see Figure 8(b)) and verticaldirections (see Figure 8(c)) it is observed that smartphone

Vertical

ndash001

0

001D

ispla

cem

ent (

m)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(c)

Figure 10 Dynamic displacements in three directions (a) N-S (b) E-W and (c) vertical

NndashSFFT

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(a)

NndashSPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(b)

EndashWFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(c)

EndashWPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

VerticalFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(e)

VerticalPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 11 Dynamic frequencies in terms of normalized FFT and periodogram

Advances in Civil Engineering 11

and accelerometer are recording very similar results In timedomain these results are very promising However tocomplement them results are changed to the frequencydomain in terms of FFT and periodogram is is discussedin the next section

96 FFT and Periodogram Using Accelerometer and Smart-phone Accelerations In this section results are studied infrequency domain ey are presented in terms of FFT andperiodogram (see Figure 12)

Figure 12 summarizes the results related to FFT andperiodogram corresponding to longitudinal transversal andvertical directions It can be observed in Figure 12 that thefrequency that contributes the most to the structural re-sponse may be located approximately at 9Hz In addition itis noted that FFT in the longitudinal and transversal di-rection (see Figures 12(a) and 12(c)) is presenting the 13rstfrequency of interest at about 35Hz Based on the resultsplotted in Figure 12 it can be stated that both smartphoneand accelerometer are providing very accurate results with

respect to each other Hence it may result practical to usesmartphones for extracting dynamic characteristics of real-scale bridges

97 Comparison of Periodogram of GPS Accelerometers andSmartphones Using Dynamic Displacements To comparedata acquisition capabilities of smartphones accelerometersand GPS the periodogram of them with respect to dynamicdisplacements in the vertical direction is illustrated inFigure 13 Only vertical direction is studied because it wasdemonstrated that this direction presented a critical be-havior It is noted as reported in Figure 12 that the fre-quency that contributes the most to the structural responseis located approximately at 937Hz which is related toprevious acceleration spectra is frequency is well detectedby the three instruments Periodogram reported in Figure 13is plotted considering dynamic displacements of smart-phone accelerometer and GPS Hence the accelerationrecorded by the smartphone and accelerometer respectivelyis converted to dynamic displacement by carrying out a

SmartphoneAccelerometer

FFTmdashlongitudinal direction

0

05

1A

mpl

itude

5 10 15 200Frequency (Hz)

(a)

SmartphoneAccelerometer

Periodogrammdashlongitudinal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(b)

SmartphoneAccelerometer

FFTmdashtransversal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(c)

SmartphoneAccelerometer

Periodogrammdashtransversal direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

SmartphoneAccelerometer

FFTmdashvertical direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(e)

SmartphoneAccelerometer

Periodogrammdashvertical direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 12 Dynamic frequencies in terms of FFT and periodogram

12 Advances in Civil Engineering

double integration which may involve some discrepancies inobtaining displacements In addition to numerical in-tegration some of the dierences observed in Figure 13 maybe associated with the applied 13ltering techniques to obtaindynamic displacements On the other hand GPS reportsdisplacements directly improving considerably data pro-cessing and interpretations However it is considered im-portant to analyze the three periodograms from Figure 13 tohave a better understanding about the coincidences of somefrequencies that appear in two or more periodograms and itmay be useful to have a better idea about de dynamic be-havior of a bridge

10 Conclusions

An alternative methodology to perform SHM on real-scalebridges is presented in this paper using GPS an acceler-ometer and a smartphone integrated in a fused smartsensor Such sensor is validated for the SHM of complexstructures as bridges under normal serviceability conditionsin real time A reinforced concrete bridge was studied byusing the proposed smart sensor because it presented ex-cessive vibration problems e accuracy of this smart sensorwas demonstrated comparing the gathered data from GPSaccelerometers and smartphones in both time and fre-quency domain Semistatic and dynamic displacements aswell as accelerations demonstrated that there are criticalperformance problems on the analyzed bridge particularlyin the vertical direction Since this critical behavior wasreported by other scholars that performed the SHM withdierent devices it can be stated that this smart sensor canbe implemented for SHM on bridges in real time In ad-dition it was demonstrated that the proposed smart sensor isadequate to perform SHM in real-scale bridges Conse-quently for the bridge under consideration it was calculated

that the fundamental frequency in the horizontal direction islocated at 9Hz Furthermore it was detected the 13rst fre-quency of interest around 35Hz In order to improvefrequency domain analysis the proposed smart sensormethodology performs a periodogram algorithm over thevertical direction data to obtain a more accurate SHM It wasobserved that there is a dierence between GPS displace-ment data and accelerometer-smartphone based data It canbe justi13ed due to the nature of each measurement meth-odology where GPS base data comes from atmosphere andaccelerometer-based data comes from a vibration sensorBecause of this the noise nature of the aforementionedmeasurement devices is not the same due to its dierentmeasurement principle On the other hand it was observed acoincidence of the three instruments approximately at937Hz for the vertical component us it can explain theusefulness of fusing GPS accelerometer and smartphoneinformation in order to discriminate undesired noise thatcan be generated by atmosphere disturbances of doubleintegration error from acceleration data when it appearsonly in one device data Finally it can be stated that thispaper contributed in understanding how GPS accelerom-eters and smartphones can be successfully fused in a smartsensor to perform SHM on real-scale bridges

Data Availability

e data used to support the 13ndings of this study areavailable from the corresponding author upon request

Disclosure

Any opinions 13ndings conclusions or recommendationsexpressed in this paper are those of the authors and do notnecessarily reect the views of the sponsors

Conflicts of Interest

e authors do not have any conicts of interest to declareabout this research

Acknowledgments

e authors would like to thank Septentrio Inc for pro-viding the AsteRx-U Receiver and the PolaNt-x MF antennaused during the GPS data collection is study is based onwork partly supported by the Consejo Nacional de Ciencia yTecnologıa (CONACYT-Mexico) as scholarships for the 13rstand seventh author respectively In addition the work isalso supported by the Universidad Autonoma de Sinaloa(UAS)

References

[1] V Ashkenazi and G W Roberts ldquoExperimental monitoringof the Humber bridge using GPSrdquo Proceedings of the In-stitution of Civil Engineers-Civil Engineering vol 120 no 4pp 177ndash182 1997

[2] W S Chan Y L Xu X L Ding andW J Dai ldquoAn integratedGPSndashaccelerometer data processing technique for structural

1

09

08

07

06

05

04

03

02

01

00 5 10

Frequency (Hz)15 20

SmartphoneAccelerometerGPS (vertical)

Figure 13 Dynamic frequencies in terms of periodogram (verticaldirection)

Advances in Civil Engineering 13

deformation monitoringrdquo Journal of Geodesy vol 80 no 12pp 705ndash719 2006

[3] S B Im S Hurlebaus and Y J Kang ldquoSummary review ofGPS technology for structural health monitoringrdquo Journal ofStructural Engineering vol 139 no 10 pp 1653ndash1664 2011

[4] M R Kaloop and H Li ldquoMonitoring of bridge deformationusing GPS techniquerdquo KSCE Journal of Civil Engineeringvol 13 no 6 pp 423ndash431 2009

[5] M Kaloop E Elbeltagi J Hu and A Elrefai ldquoRecent advancesof structures monitoring and evaluation using GPS-time seriesmonitoring systems a reviewrdquo ISPRS International Journal ofGeo-Information vol 6 no 12 p 382 2017

[6] X Li C Rizos L Ge Y Tamura and A Yoshida ldquo+ecomplementary characteristics of GPS and accelerometer inmonitoring structural deformationrdquo in Proceedings of the ION2005 National Technical Meeting Institute of Navigation(ION) San Diego CA USA January 2005

[7] P Psimoulis S Pytharouli D Karambalis and S StirosldquoPotential of Global Positioning System (GPS) to measurefrequencies of oscillations of engineering structuresrdquo Journalof Sound and Vibration vol 318 no 3 pp 606ndash623 2008

[8] C Watson T Watson and R Coleman ldquoStructural moni-toring of cable-stayed bridge analysis of GPS versus modeleddeflectionsrdquo Journal of Surveying Engineering vol 133 no 1pp 23ndash28 2007

[9] C Xiong and Y Niu ldquoInvestigation of the dynamic behavior ofa super high-rise structure using RTK-GNSS techniquerdquo KSCEJournal of Civil Engineering vol 23 no 2 pp 654ndash665 2018

[10] B Hofmann-Wellenhof H Lichtenegger and E WasleGNSSndashGlobal Navigation Satellite Systems GPS GLONASSGalileo and More Springer Science amp Business MediaVienna Austria 2007

[11] M Celebi ldquoGPS in dynamic monitoring of long-periodstructuresrdquo Soil Dynamics and Earthquake Engineeringvol 20 no 5ndash8 pp 477ndash483 2000

[12] D K Kim J I Kim and M Q Feng ldquoInstrumentation ofbridges for structural health monitoringrdquo KSCE Journal ofCivil Engineering vol 5 no 3 pp 231ndash242 2001

[13] G Q Wang D M Boore H Igel and X Y Zhou ldquoSomeobservations on colocated and closely spaced strong ground-motion records of the 1999 Chi-Chi Taiwan earthquakerdquoBulletin of the Seismological Society of America vol 93 no 2pp 674ndash693 2003

[14] J Hwang H Yun S K Park D Lee and S Hong ldquoOptimalmethods of RTK-GPSaccelerometer integration to monitorthe displacement of structuresrdquo Sensors vol 12 no 1pp 1014ndash1034 2012

[15] S Dashti J D Bray J Reilly S Glaser A Bayen and E MarildquoEvaluating the reliability of phones as seismic monitoringinstrumentsrdquo Earthquake Spectra vol 30 no 2 pp 721ndash7422014

[16] M Feng Y Fukuda M Mizuta and E Ozer ldquoCitizen sensorsfor SHM use of accelerometer data from smartphonesrdquoSensors vol 15 no 2 pp 2980ndash2998 2015

[17] X Zhao R Han Y Ding et al ldquoPortable and convenient cableforce measurement using smartphonerdquo Journal of CivilStructural Health Monitoring vol 5 no 4 pp 481ndash491 2015

[18] X Zhao K Ri R Han Y Yu M Li and J Ou ldquoExperimentalresearch on quick structural health monitoring technique forbridges using smartphonerdquoAdvances inMaterials Science andEngineering vol 2016 no 1 pp 1ndash14 2016

[19] Y Ding R Han H Liu S Li X Zhao and Y Yu ldquoBridgeinspection and management system based on smart phonerdquoin Proceedings of the ASME 2016 Conference on Smart

Materials Adaptive Structures and Intelligent SystemsAmerican Society of Mechanical Engineers Stowe VT USASeptember 2016

[20] X Zhao R Han K J Loh B Xie J Li and J Ou ldquoShakingtable tests for evaluating the damage features under earth-quake excitations using smartphonesrdquo in Proceedings of theHealth Monitoring of Structural and Biological Systems XIIInternational Society for Optics and Photonics Denver COUSA May 2018

[21] F Moschas and S Stiros ldquoMeasurement of the dynamicdisplacements and of the modal frequencies of a short-spanpedestrian bridge using GPS and an accelerometerrdquo Engi-neering Structures vol 33 no 1 pp 10ndash17 2011

[22] M R Kaloop ldquoBridge safety monitoring based-GPS tech-nique case study Zhujiang Huangpu bridgerdquo Smart Struc-tures and Systems vol 9 no 6 pp 473ndash487 2012

[23] G W Roberts X Meng and A H Dodson ldquoIntegrating aglobal positioning system and accelerometers to monitor thedeflection of bridgesrdquo Journal of Surveying Engineeringvol 130 no 2 pp 65ndash72 2004

[24] Accelerometer Analyzer App 2017 httpsplaygooglecomstoreappsdetailsidcomlulaccelerometeramphles

[25] X Meng ldquoReal-time deformationmonitoring of bridges usingGPSaccelerometersrdquo Doctoral dissertation University ofNottingham UK 2002

[26] G E Vazquez-Becerra J R Gaxiola-Camacho R BennettG M Guzman-Acevedo and I E Gaxiola-CamacholdquoStructural evaluation of dynamic and semi-static displace-ments of the Juarez bridge using GPS technologyrdquo Mea-surement vol 110 no 1 pp 146ndash153 2017

[27] P Psimoulis I Peppa L Bonenberg S Ince and X MengldquoCombination of GPS and RTS measurements for themonitoring of semi-static and dynamic motion of pedestrianbridgerdquo in Proceedings of the 3rd Joint International Sym-posium on Deformation Monitoring Vienna Austria April2016

[28] F Moschas and S Stiros ldquoNoise characteristics of high-frequencyshort-duration GPS records from analysis ofidentical collocated instrumentsrdquo Measurement vol 46no 4 pp 1488ndash1506 2013

[29] S W Smithgte Scientist amp Engineerrsquos Guide to Digital SignalProcessing California Technical Publishing San Diego CAUSA 1997

[30] J Wang X Meng C Qin and J Yi ldquoVibration frequenciesextraction of the forth road bridge using high sampling GPSdatardquo Shock and Vibration vol 2016 Article ID 980786118 pages 2016

[31] J Yu XMeng X Shao B Yan and L Yang ldquoIdentification ofdynamic displacements and modal frequencies of a medium-span suspension bridge using multimode GNSS processingrdquoEngineering Structures vol 81 no 1 pp 432ndash443 2014

[32] E Elbeltagi M Kaloop and M Elnabwy ldquoMonitoring andassessment of Mansoura railway steel bridge using RTK-GPStechniquerdquo in Proceedings of the 2nd Joint InternationalSymposium on Deformation Monitoring pp 9ndash11 Notting-ham England September 2013

[33] N Raziq ldquoGPS structural deformation monitoring the mid-height problemrdquo Doctoral dissertation University ofMelbourneAustralia 2008

[34] S S HaykinAdaptive Filtergteory Pearson Education India5th edition 2014

[35] L H Estey and C M Meertens ldquoTEQC the multi-purposetoolkit for GPSGLONASS datardquo GPS Solutions vol 3 no 1pp 42ndash49 1999

14 Advances in Civil Engineering

94 FFT and Periodogram Using GPS DynamicDisplacements To identify structural dynamic character-istics of the structure under consideration FFT and perio-dograms are calculated and plotted in frequency domainusing dynamic displacements of GPS

Figure 11 illustrates FFT and periodogram in the threeGPS directions N-S (Figures 11(a) and 11(b)) E-W(Figures 11(c) and 11(d)) and vertical (Figures 11(e) and11(f)) To clarify the results they are normalized with respectto their maximum amplitude value

95 Acceleration Time History Using Accelerometer andSmartphone Acceleration time histories corresponding tothe longitudinal transversal and vertical direction of thebridge were earlier presented in Figure 8 considering bothsmartphone and accelerometer In the longitudinal direction(see Figure 8(a)) it can be noted that there are some dif-ferences in the acceleration reported by smartphone andaccelerometer such discrepancies were justi13ed earlier inSection 91 In the transversal (see Figure 8(b)) and verticaldirections (see Figure 8(c)) it is observed that smartphone

Vertical

ndash001

0

001D

ispla

cem

ent (

m)

50 100 150 200 250 300 350 400 450 5000Time (sec)

(c)

Figure 10 Dynamic displacements in three directions (a) N-S (b) E-W and (c) vertical

NndashSFFT

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(a)

NndashSPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(b)

EndashWFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(c)

EndashWPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

VerticalFFT

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(e)

VerticalPeriodogram

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 11 Dynamic frequencies in terms of normalized FFT and periodogram

Advances in Civil Engineering 11

and accelerometer are recording very similar results In timedomain these results are very promising However tocomplement them results are changed to the frequencydomain in terms of FFT and periodogram is is discussedin the next section

96 FFT and Periodogram Using Accelerometer and Smart-phone Accelerations In this section results are studied infrequency domain ey are presented in terms of FFT andperiodogram (see Figure 12)

Figure 12 summarizes the results related to FFT andperiodogram corresponding to longitudinal transversal andvertical directions It can be observed in Figure 12 that thefrequency that contributes the most to the structural re-sponse may be located approximately at 9Hz In addition itis noted that FFT in the longitudinal and transversal di-rection (see Figures 12(a) and 12(c)) is presenting the 13rstfrequency of interest at about 35Hz Based on the resultsplotted in Figure 12 it can be stated that both smartphoneand accelerometer are providing very accurate results with

respect to each other Hence it may result practical to usesmartphones for extracting dynamic characteristics of real-scale bridges

97 Comparison of Periodogram of GPS Accelerometers andSmartphones Using Dynamic Displacements To comparedata acquisition capabilities of smartphones accelerometersand GPS the periodogram of them with respect to dynamicdisplacements in the vertical direction is illustrated inFigure 13 Only vertical direction is studied because it wasdemonstrated that this direction presented a critical be-havior It is noted as reported in Figure 12 that the fre-quency that contributes the most to the structural responseis located approximately at 937Hz which is related toprevious acceleration spectra is frequency is well detectedby the three instruments Periodogram reported in Figure 13is plotted considering dynamic displacements of smart-phone accelerometer and GPS Hence the accelerationrecorded by the smartphone and accelerometer respectivelyis converted to dynamic displacement by carrying out a

SmartphoneAccelerometer

FFTmdashlongitudinal direction

0

05

1A

mpl

itude

5 10 15 200Frequency (Hz)

(a)

SmartphoneAccelerometer

Periodogrammdashlongitudinal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(b)

SmartphoneAccelerometer

FFTmdashtransversal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(c)

SmartphoneAccelerometer

Periodogrammdashtransversal direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

SmartphoneAccelerometer

FFTmdashvertical direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(e)

SmartphoneAccelerometer

Periodogrammdashvertical direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 12 Dynamic frequencies in terms of FFT and periodogram

12 Advances in Civil Engineering

double integration which may involve some discrepancies inobtaining displacements In addition to numerical in-tegration some of the dierences observed in Figure 13 maybe associated with the applied 13ltering techniques to obtaindynamic displacements On the other hand GPS reportsdisplacements directly improving considerably data pro-cessing and interpretations However it is considered im-portant to analyze the three periodograms from Figure 13 tohave a better understanding about the coincidences of somefrequencies that appear in two or more periodograms and itmay be useful to have a better idea about de dynamic be-havior of a bridge

10 Conclusions

An alternative methodology to perform SHM on real-scalebridges is presented in this paper using GPS an acceler-ometer and a smartphone integrated in a fused smartsensor Such sensor is validated for the SHM of complexstructures as bridges under normal serviceability conditionsin real time A reinforced concrete bridge was studied byusing the proposed smart sensor because it presented ex-cessive vibration problems e accuracy of this smart sensorwas demonstrated comparing the gathered data from GPSaccelerometers and smartphones in both time and fre-quency domain Semistatic and dynamic displacements aswell as accelerations demonstrated that there are criticalperformance problems on the analyzed bridge particularlyin the vertical direction Since this critical behavior wasreported by other scholars that performed the SHM withdierent devices it can be stated that this smart sensor canbe implemented for SHM on bridges in real time In ad-dition it was demonstrated that the proposed smart sensor isadequate to perform SHM in real-scale bridges Conse-quently for the bridge under consideration it was calculated

that the fundamental frequency in the horizontal direction islocated at 9Hz Furthermore it was detected the 13rst fre-quency of interest around 35Hz In order to improvefrequency domain analysis the proposed smart sensormethodology performs a periodogram algorithm over thevertical direction data to obtain a more accurate SHM It wasobserved that there is a dierence between GPS displace-ment data and accelerometer-smartphone based data It canbe justi13ed due to the nature of each measurement meth-odology where GPS base data comes from atmosphere andaccelerometer-based data comes from a vibration sensorBecause of this the noise nature of the aforementionedmeasurement devices is not the same due to its dierentmeasurement principle On the other hand it was observed acoincidence of the three instruments approximately at937Hz for the vertical component us it can explain theusefulness of fusing GPS accelerometer and smartphoneinformation in order to discriminate undesired noise thatcan be generated by atmosphere disturbances of doubleintegration error from acceleration data when it appearsonly in one device data Finally it can be stated that thispaper contributed in understanding how GPS accelerom-eters and smartphones can be successfully fused in a smartsensor to perform SHM on real-scale bridges

Data Availability

e data used to support the 13ndings of this study areavailable from the corresponding author upon request

Disclosure

Any opinions 13ndings conclusions or recommendationsexpressed in this paper are those of the authors and do notnecessarily reect the views of the sponsors

Conflicts of Interest

e authors do not have any conicts of interest to declareabout this research

Acknowledgments

e authors would like to thank Septentrio Inc for pro-viding the AsteRx-U Receiver and the PolaNt-x MF antennaused during the GPS data collection is study is based onwork partly supported by the Consejo Nacional de Ciencia yTecnologıa (CONACYT-Mexico) as scholarships for the 13rstand seventh author respectively In addition the work isalso supported by the Universidad Autonoma de Sinaloa(UAS)

References

[1] V Ashkenazi and G W Roberts ldquoExperimental monitoringof the Humber bridge using GPSrdquo Proceedings of the In-stitution of Civil Engineers-Civil Engineering vol 120 no 4pp 177ndash182 1997

[2] W S Chan Y L Xu X L Ding andW J Dai ldquoAn integratedGPSndashaccelerometer data processing technique for structural

1

09

08

07

06

05

04

03

02

01

00 5 10

Frequency (Hz)15 20

SmartphoneAccelerometerGPS (vertical)

Figure 13 Dynamic frequencies in terms of periodogram (verticaldirection)

Advances in Civil Engineering 13

deformation monitoringrdquo Journal of Geodesy vol 80 no 12pp 705ndash719 2006

[3] S B Im S Hurlebaus and Y J Kang ldquoSummary review ofGPS technology for structural health monitoringrdquo Journal ofStructural Engineering vol 139 no 10 pp 1653ndash1664 2011

[4] M R Kaloop and H Li ldquoMonitoring of bridge deformationusing GPS techniquerdquo KSCE Journal of Civil Engineeringvol 13 no 6 pp 423ndash431 2009

[5] M Kaloop E Elbeltagi J Hu and A Elrefai ldquoRecent advancesof structures monitoring and evaluation using GPS-time seriesmonitoring systems a reviewrdquo ISPRS International Journal ofGeo-Information vol 6 no 12 p 382 2017

[6] X Li C Rizos L Ge Y Tamura and A Yoshida ldquo+ecomplementary characteristics of GPS and accelerometer inmonitoring structural deformationrdquo in Proceedings of the ION2005 National Technical Meeting Institute of Navigation(ION) San Diego CA USA January 2005

[7] P Psimoulis S Pytharouli D Karambalis and S StirosldquoPotential of Global Positioning System (GPS) to measurefrequencies of oscillations of engineering structuresrdquo Journalof Sound and Vibration vol 318 no 3 pp 606ndash623 2008

[8] C Watson T Watson and R Coleman ldquoStructural moni-toring of cable-stayed bridge analysis of GPS versus modeleddeflectionsrdquo Journal of Surveying Engineering vol 133 no 1pp 23ndash28 2007

[9] C Xiong and Y Niu ldquoInvestigation of the dynamic behavior ofa super high-rise structure using RTK-GNSS techniquerdquo KSCEJournal of Civil Engineering vol 23 no 2 pp 654ndash665 2018

[10] B Hofmann-Wellenhof H Lichtenegger and E WasleGNSSndashGlobal Navigation Satellite Systems GPS GLONASSGalileo and More Springer Science amp Business MediaVienna Austria 2007

[11] M Celebi ldquoGPS in dynamic monitoring of long-periodstructuresrdquo Soil Dynamics and Earthquake Engineeringvol 20 no 5ndash8 pp 477ndash483 2000

[12] D K Kim J I Kim and M Q Feng ldquoInstrumentation ofbridges for structural health monitoringrdquo KSCE Journal ofCivil Engineering vol 5 no 3 pp 231ndash242 2001

[13] G Q Wang D M Boore H Igel and X Y Zhou ldquoSomeobservations on colocated and closely spaced strong ground-motion records of the 1999 Chi-Chi Taiwan earthquakerdquoBulletin of the Seismological Society of America vol 93 no 2pp 674ndash693 2003

[14] J Hwang H Yun S K Park D Lee and S Hong ldquoOptimalmethods of RTK-GPSaccelerometer integration to monitorthe displacement of structuresrdquo Sensors vol 12 no 1pp 1014ndash1034 2012

[15] S Dashti J D Bray J Reilly S Glaser A Bayen and E MarildquoEvaluating the reliability of phones as seismic monitoringinstrumentsrdquo Earthquake Spectra vol 30 no 2 pp 721ndash7422014

[16] M Feng Y Fukuda M Mizuta and E Ozer ldquoCitizen sensorsfor SHM use of accelerometer data from smartphonesrdquoSensors vol 15 no 2 pp 2980ndash2998 2015

[17] X Zhao R Han Y Ding et al ldquoPortable and convenient cableforce measurement using smartphonerdquo Journal of CivilStructural Health Monitoring vol 5 no 4 pp 481ndash491 2015

[18] X Zhao K Ri R Han Y Yu M Li and J Ou ldquoExperimentalresearch on quick structural health monitoring technique forbridges using smartphonerdquoAdvances inMaterials Science andEngineering vol 2016 no 1 pp 1ndash14 2016

[19] Y Ding R Han H Liu S Li X Zhao and Y Yu ldquoBridgeinspection and management system based on smart phonerdquoin Proceedings of the ASME 2016 Conference on Smart

Materials Adaptive Structures and Intelligent SystemsAmerican Society of Mechanical Engineers Stowe VT USASeptember 2016

[20] X Zhao R Han K J Loh B Xie J Li and J Ou ldquoShakingtable tests for evaluating the damage features under earth-quake excitations using smartphonesrdquo in Proceedings of theHealth Monitoring of Structural and Biological Systems XIIInternational Society for Optics and Photonics Denver COUSA May 2018

[21] F Moschas and S Stiros ldquoMeasurement of the dynamicdisplacements and of the modal frequencies of a short-spanpedestrian bridge using GPS and an accelerometerrdquo Engi-neering Structures vol 33 no 1 pp 10ndash17 2011

[22] M R Kaloop ldquoBridge safety monitoring based-GPS tech-nique case study Zhujiang Huangpu bridgerdquo Smart Struc-tures and Systems vol 9 no 6 pp 473ndash487 2012

[23] G W Roberts X Meng and A H Dodson ldquoIntegrating aglobal positioning system and accelerometers to monitor thedeflection of bridgesrdquo Journal of Surveying Engineeringvol 130 no 2 pp 65ndash72 2004

[24] Accelerometer Analyzer App 2017 httpsplaygooglecomstoreappsdetailsidcomlulaccelerometeramphles

[25] X Meng ldquoReal-time deformationmonitoring of bridges usingGPSaccelerometersrdquo Doctoral dissertation University ofNottingham UK 2002

[26] G E Vazquez-Becerra J R Gaxiola-Camacho R BennettG M Guzman-Acevedo and I E Gaxiola-CamacholdquoStructural evaluation of dynamic and semi-static displace-ments of the Juarez bridge using GPS technologyrdquo Mea-surement vol 110 no 1 pp 146ndash153 2017

[27] P Psimoulis I Peppa L Bonenberg S Ince and X MengldquoCombination of GPS and RTS measurements for themonitoring of semi-static and dynamic motion of pedestrianbridgerdquo in Proceedings of the 3rd Joint International Sym-posium on Deformation Monitoring Vienna Austria April2016

[28] F Moschas and S Stiros ldquoNoise characteristics of high-frequencyshort-duration GPS records from analysis ofidentical collocated instrumentsrdquo Measurement vol 46no 4 pp 1488ndash1506 2013

[29] S W Smithgte Scientist amp Engineerrsquos Guide to Digital SignalProcessing California Technical Publishing San Diego CAUSA 1997

[30] J Wang X Meng C Qin and J Yi ldquoVibration frequenciesextraction of the forth road bridge using high sampling GPSdatardquo Shock and Vibration vol 2016 Article ID 980786118 pages 2016

[31] J Yu XMeng X Shao B Yan and L Yang ldquoIdentification ofdynamic displacements and modal frequencies of a medium-span suspension bridge using multimode GNSS processingrdquoEngineering Structures vol 81 no 1 pp 432ndash443 2014

[32] E Elbeltagi M Kaloop and M Elnabwy ldquoMonitoring andassessment of Mansoura railway steel bridge using RTK-GPStechniquerdquo in Proceedings of the 2nd Joint InternationalSymposium on Deformation Monitoring pp 9ndash11 Notting-ham England September 2013

[33] N Raziq ldquoGPS structural deformation monitoring the mid-height problemrdquo Doctoral dissertation University ofMelbourneAustralia 2008

[34] S S HaykinAdaptive Filtergteory Pearson Education India5th edition 2014

[35] L H Estey and C M Meertens ldquoTEQC the multi-purposetoolkit for GPSGLONASS datardquo GPS Solutions vol 3 no 1pp 42ndash49 1999

14 Advances in Civil Engineering

and accelerometer are recording very similar results In timedomain these results are very promising However tocomplement them results are changed to the frequencydomain in terms of FFT and periodogram is is discussedin the next section

96 FFT and Periodogram Using Accelerometer and Smart-phone Accelerations In this section results are studied infrequency domain ey are presented in terms of FFT andperiodogram (see Figure 12)

Figure 12 summarizes the results related to FFT andperiodogram corresponding to longitudinal transversal andvertical directions It can be observed in Figure 12 that thefrequency that contributes the most to the structural re-sponse may be located approximately at 9Hz In addition itis noted that FFT in the longitudinal and transversal di-rection (see Figures 12(a) and 12(c)) is presenting the 13rstfrequency of interest at about 35Hz Based on the resultsplotted in Figure 12 it can be stated that both smartphoneand accelerometer are providing very accurate results with

respect to each other Hence it may result practical to usesmartphones for extracting dynamic characteristics of real-scale bridges

97 Comparison of Periodogram of GPS Accelerometers andSmartphones Using Dynamic Displacements To comparedata acquisition capabilities of smartphones accelerometersand GPS the periodogram of them with respect to dynamicdisplacements in the vertical direction is illustrated inFigure 13 Only vertical direction is studied because it wasdemonstrated that this direction presented a critical be-havior It is noted as reported in Figure 12 that the fre-quency that contributes the most to the structural responseis located approximately at 937Hz which is related toprevious acceleration spectra is frequency is well detectedby the three instruments Periodogram reported in Figure 13is plotted considering dynamic displacements of smart-phone accelerometer and GPS Hence the accelerationrecorded by the smartphone and accelerometer respectivelyis converted to dynamic displacement by carrying out a

SmartphoneAccelerometer

FFTmdashlongitudinal direction

0

05

1A

mpl

itude

5 10 15 200Frequency (Hz)

(a)

SmartphoneAccelerometer

Periodogrammdashlongitudinal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(b)

SmartphoneAccelerometer

FFTmdashtransversal direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(c)

SmartphoneAccelerometer

Periodogrammdashtransversal direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(d)

SmartphoneAccelerometer

FFTmdashvertical direction

0

05

1

Am

plitu

de

5 10 15 200Frequency (Hz)

(e)

SmartphoneAccelerometer

Periodogrammdashvertical direction

5 10 15 200Frequency (Hz)

0

05

1

Am

plitu

de

(f )

Figure 12 Dynamic frequencies in terms of FFT and periodogram

12 Advances in Civil Engineering

double integration which may involve some discrepancies inobtaining displacements In addition to numerical in-tegration some of the dierences observed in Figure 13 maybe associated with the applied 13ltering techniques to obtaindynamic displacements On the other hand GPS reportsdisplacements directly improving considerably data pro-cessing and interpretations However it is considered im-portant to analyze the three periodograms from Figure 13 tohave a better understanding about the coincidences of somefrequencies that appear in two or more periodograms and itmay be useful to have a better idea about de dynamic be-havior of a bridge

10 Conclusions

An alternative methodology to perform SHM on real-scalebridges is presented in this paper using GPS an acceler-ometer and a smartphone integrated in a fused smartsensor Such sensor is validated for the SHM of complexstructures as bridges under normal serviceability conditionsin real time A reinforced concrete bridge was studied byusing the proposed smart sensor because it presented ex-cessive vibration problems e accuracy of this smart sensorwas demonstrated comparing the gathered data from GPSaccelerometers and smartphones in both time and fre-quency domain Semistatic and dynamic displacements aswell as accelerations demonstrated that there are criticalperformance problems on the analyzed bridge particularlyin the vertical direction Since this critical behavior wasreported by other scholars that performed the SHM withdierent devices it can be stated that this smart sensor canbe implemented for SHM on bridges in real time In ad-dition it was demonstrated that the proposed smart sensor isadequate to perform SHM in real-scale bridges Conse-quently for the bridge under consideration it was calculated

that the fundamental frequency in the horizontal direction islocated at 9Hz Furthermore it was detected the 13rst fre-quency of interest around 35Hz In order to improvefrequency domain analysis the proposed smart sensormethodology performs a periodogram algorithm over thevertical direction data to obtain a more accurate SHM It wasobserved that there is a dierence between GPS displace-ment data and accelerometer-smartphone based data It canbe justi13ed due to the nature of each measurement meth-odology where GPS base data comes from atmosphere andaccelerometer-based data comes from a vibration sensorBecause of this the noise nature of the aforementionedmeasurement devices is not the same due to its dierentmeasurement principle On the other hand it was observed acoincidence of the three instruments approximately at937Hz for the vertical component us it can explain theusefulness of fusing GPS accelerometer and smartphoneinformation in order to discriminate undesired noise thatcan be generated by atmosphere disturbances of doubleintegration error from acceleration data when it appearsonly in one device data Finally it can be stated that thispaper contributed in understanding how GPS accelerom-eters and smartphones can be successfully fused in a smartsensor to perform SHM on real-scale bridges

Data Availability

e data used to support the 13ndings of this study areavailable from the corresponding author upon request

Disclosure

Any opinions 13ndings conclusions or recommendationsexpressed in this paper are those of the authors and do notnecessarily reect the views of the sponsors

Conflicts of Interest

e authors do not have any conicts of interest to declareabout this research

Acknowledgments

e authors would like to thank Septentrio Inc for pro-viding the AsteRx-U Receiver and the PolaNt-x MF antennaused during the GPS data collection is study is based onwork partly supported by the Consejo Nacional de Ciencia yTecnologıa (CONACYT-Mexico) as scholarships for the 13rstand seventh author respectively In addition the work isalso supported by the Universidad Autonoma de Sinaloa(UAS)

References

[1] V Ashkenazi and G W Roberts ldquoExperimental monitoringof the Humber bridge using GPSrdquo Proceedings of the In-stitution of Civil Engineers-Civil Engineering vol 120 no 4pp 177ndash182 1997

[2] W S Chan Y L Xu X L Ding andW J Dai ldquoAn integratedGPSndashaccelerometer data processing technique for structural

1

09

08

07

06

05

04

03

02

01

00 5 10

Frequency (Hz)15 20

SmartphoneAccelerometerGPS (vertical)

Figure 13 Dynamic frequencies in terms of periodogram (verticaldirection)

Advances in Civil Engineering 13

deformation monitoringrdquo Journal of Geodesy vol 80 no 12pp 705ndash719 2006

[3] S B Im S Hurlebaus and Y J Kang ldquoSummary review ofGPS technology for structural health monitoringrdquo Journal ofStructural Engineering vol 139 no 10 pp 1653ndash1664 2011

[4] M R Kaloop and H Li ldquoMonitoring of bridge deformationusing GPS techniquerdquo KSCE Journal of Civil Engineeringvol 13 no 6 pp 423ndash431 2009

[5] M Kaloop E Elbeltagi J Hu and A Elrefai ldquoRecent advancesof structures monitoring and evaluation using GPS-time seriesmonitoring systems a reviewrdquo ISPRS International Journal ofGeo-Information vol 6 no 12 p 382 2017

[6] X Li C Rizos L Ge Y Tamura and A Yoshida ldquo+ecomplementary characteristics of GPS and accelerometer inmonitoring structural deformationrdquo in Proceedings of the ION2005 National Technical Meeting Institute of Navigation(ION) San Diego CA USA January 2005

[7] P Psimoulis S Pytharouli D Karambalis and S StirosldquoPotential of Global Positioning System (GPS) to measurefrequencies of oscillations of engineering structuresrdquo Journalof Sound and Vibration vol 318 no 3 pp 606ndash623 2008

[8] C Watson T Watson and R Coleman ldquoStructural moni-toring of cable-stayed bridge analysis of GPS versus modeleddeflectionsrdquo Journal of Surveying Engineering vol 133 no 1pp 23ndash28 2007

[9] C Xiong and Y Niu ldquoInvestigation of the dynamic behavior ofa super high-rise structure using RTK-GNSS techniquerdquo KSCEJournal of Civil Engineering vol 23 no 2 pp 654ndash665 2018

[10] B Hofmann-Wellenhof H Lichtenegger and E WasleGNSSndashGlobal Navigation Satellite Systems GPS GLONASSGalileo and More Springer Science amp Business MediaVienna Austria 2007

[11] M Celebi ldquoGPS in dynamic monitoring of long-periodstructuresrdquo Soil Dynamics and Earthquake Engineeringvol 20 no 5ndash8 pp 477ndash483 2000

[12] D K Kim J I Kim and M Q Feng ldquoInstrumentation ofbridges for structural health monitoringrdquo KSCE Journal ofCivil Engineering vol 5 no 3 pp 231ndash242 2001

[13] G Q Wang D M Boore H Igel and X Y Zhou ldquoSomeobservations on colocated and closely spaced strong ground-motion records of the 1999 Chi-Chi Taiwan earthquakerdquoBulletin of the Seismological Society of America vol 93 no 2pp 674ndash693 2003

[14] J Hwang H Yun S K Park D Lee and S Hong ldquoOptimalmethods of RTK-GPSaccelerometer integration to monitorthe displacement of structuresrdquo Sensors vol 12 no 1pp 1014ndash1034 2012

[15] S Dashti J D Bray J Reilly S Glaser A Bayen and E MarildquoEvaluating the reliability of phones as seismic monitoringinstrumentsrdquo Earthquake Spectra vol 30 no 2 pp 721ndash7422014

[16] M Feng Y Fukuda M Mizuta and E Ozer ldquoCitizen sensorsfor SHM use of accelerometer data from smartphonesrdquoSensors vol 15 no 2 pp 2980ndash2998 2015

[17] X Zhao R Han Y Ding et al ldquoPortable and convenient cableforce measurement using smartphonerdquo Journal of CivilStructural Health Monitoring vol 5 no 4 pp 481ndash491 2015

[18] X Zhao K Ri R Han Y Yu M Li and J Ou ldquoExperimentalresearch on quick structural health monitoring technique forbridges using smartphonerdquoAdvances inMaterials Science andEngineering vol 2016 no 1 pp 1ndash14 2016

[19] Y Ding R Han H Liu S Li X Zhao and Y Yu ldquoBridgeinspection and management system based on smart phonerdquoin Proceedings of the ASME 2016 Conference on Smart

Materials Adaptive Structures and Intelligent SystemsAmerican Society of Mechanical Engineers Stowe VT USASeptember 2016

[20] X Zhao R Han K J Loh B Xie J Li and J Ou ldquoShakingtable tests for evaluating the damage features under earth-quake excitations using smartphonesrdquo in Proceedings of theHealth Monitoring of Structural and Biological Systems XIIInternational Society for Optics and Photonics Denver COUSA May 2018

[21] F Moschas and S Stiros ldquoMeasurement of the dynamicdisplacements and of the modal frequencies of a short-spanpedestrian bridge using GPS and an accelerometerrdquo Engi-neering Structures vol 33 no 1 pp 10ndash17 2011

[22] M R Kaloop ldquoBridge safety monitoring based-GPS tech-nique case study Zhujiang Huangpu bridgerdquo Smart Struc-tures and Systems vol 9 no 6 pp 473ndash487 2012

[23] G W Roberts X Meng and A H Dodson ldquoIntegrating aglobal positioning system and accelerometers to monitor thedeflection of bridgesrdquo Journal of Surveying Engineeringvol 130 no 2 pp 65ndash72 2004

[24] Accelerometer Analyzer App 2017 httpsplaygooglecomstoreappsdetailsidcomlulaccelerometeramphles

[25] X Meng ldquoReal-time deformationmonitoring of bridges usingGPSaccelerometersrdquo Doctoral dissertation University ofNottingham UK 2002

[26] G E Vazquez-Becerra J R Gaxiola-Camacho R BennettG M Guzman-Acevedo and I E Gaxiola-CamacholdquoStructural evaluation of dynamic and semi-static displace-ments of the Juarez bridge using GPS technologyrdquo Mea-surement vol 110 no 1 pp 146ndash153 2017

[27] P Psimoulis I Peppa L Bonenberg S Ince and X MengldquoCombination of GPS and RTS measurements for themonitoring of semi-static and dynamic motion of pedestrianbridgerdquo in Proceedings of the 3rd Joint International Sym-posium on Deformation Monitoring Vienna Austria April2016

[28] F Moschas and S Stiros ldquoNoise characteristics of high-frequencyshort-duration GPS records from analysis ofidentical collocated instrumentsrdquo Measurement vol 46no 4 pp 1488ndash1506 2013

[29] S W Smithgte Scientist amp Engineerrsquos Guide to Digital SignalProcessing California Technical Publishing San Diego CAUSA 1997

[30] J Wang X Meng C Qin and J Yi ldquoVibration frequenciesextraction of the forth road bridge using high sampling GPSdatardquo Shock and Vibration vol 2016 Article ID 980786118 pages 2016

[31] J Yu XMeng X Shao B Yan and L Yang ldquoIdentification ofdynamic displacements and modal frequencies of a medium-span suspension bridge using multimode GNSS processingrdquoEngineering Structures vol 81 no 1 pp 432ndash443 2014

[32] E Elbeltagi M Kaloop and M Elnabwy ldquoMonitoring andassessment of Mansoura railway steel bridge using RTK-GPStechniquerdquo in Proceedings of the 2nd Joint InternationalSymposium on Deformation Monitoring pp 9ndash11 Notting-ham England September 2013

[33] N Raziq ldquoGPS structural deformation monitoring the mid-height problemrdquo Doctoral dissertation University ofMelbourneAustralia 2008

[34] S S HaykinAdaptive Filtergteory Pearson Education India5th edition 2014

[35] L H Estey and C M Meertens ldquoTEQC the multi-purposetoolkit for GPSGLONASS datardquo GPS Solutions vol 3 no 1pp 42ndash49 1999

14 Advances in Civil Engineering

double integration which may involve some discrepancies inobtaining displacements In addition to numerical in-tegration some of the dierences observed in Figure 13 maybe associated with the applied 13ltering techniques to obtaindynamic displacements On the other hand GPS reportsdisplacements directly improving considerably data pro-cessing and interpretations However it is considered im-portant to analyze the three periodograms from Figure 13 tohave a better understanding about the coincidences of somefrequencies that appear in two or more periodograms and itmay be useful to have a better idea about de dynamic be-havior of a bridge

10 Conclusions

An alternative methodology to perform SHM on real-scalebridges is presented in this paper using GPS an acceler-ometer and a smartphone integrated in a fused smartsensor Such sensor is validated for the SHM of complexstructures as bridges under normal serviceability conditionsin real time A reinforced concrete bridge was studied byusing the proposed smart sensor because it presented ex-cessive vibration problems e accuracy of this smart sensorwas demonstrated comparing the gathered data from GPSaccelerometers and smartphones in both time and fre-quency domain Semistatic and dynamic displacements aswell as accelerations demonstrated that there are criticalperformance problems on the analyzed bridge particularlyin the vertical direction Since this critical behavior wasreported by other scholars that performed the SHM withdierent devices it can be stated that this smart sensor canbe implemented for SHM on bridges in real time In ad-dition it was demonstrated that the proposed smart sensor isadequate to perform SHM in real-scale bridges Conse-quently for the bridge under consideration it was calculated

that the fundamental frequency in the horizontal direction islocated at 9Hz Furthermore it was detected the 13rst fre-quency of interest around 35Hz In order to improvefrequency domain analysis the proposed smart sensormethodology performs a periodogram algorithm over thevertical direction data to obtain a more accurate SHM It wasobserved that there is a dierence between GPS displace-ment data and accelerometer-smartphone based data It canbe justi13ed due to the nature of each measurement meth-odology where GPS base data comes from atmosphere andaccelerometer-based data comes from a vibration sensorBecause of this the noise nature of the aforementionedmeasurement devices is not the same due to its dierentmeasurement principle On the other hand it was observed acoincidence of the three instruments approximately at937Hz for the vertical component us it can explain theusefulness of fusing GPS accelerometer and smartphoneinformation in order to discriminate undesired noise thatcan be generated by atmosphere disturbances of doubleintegration error from acceleration data when it appearsonly in one device data Finally it can be stated that thispaper contributed in understanding how GPS accelerom-eters and smartphones can be successfully fused in a smartsensor to perform SHM on real-scale bridges

Data Availability

e data used to support the 13ndings of this study areavailable from the corresponding author upon request

Disclosure

Any opinions 13ndings conclusions or recommendationsexpressed in this paper are those of the authors and do notnecessarily reect the views of the sponsors

Conflicts of Interest

e authors do not have any conicts of interest to declareabout this research

Acknowledgments

e authors would like to thank Septentrio Inc for pro-viding the AsteRx-U Receiver and the PolaNt-x MF antennaused during the GPS data collection is study is based onwork partly supported by the Consejo Nacional de Ciencia yTecnologıa (CONACYT-Mexico) as scholarships for the 13rstand seventh author respectively In addition the work isalso supported by the Universidad Autonoma de Sinaloa(UAS)

References

[1] V Ashkenazi and G W Roberts ldquoExperimental monitoringof the Humber bridge using GPSrdquo Proceedings of the In-stitution of Civil Engineers-Civil Engineering vol 120 no 4pp 177ndash182 1997

[2] W S Chan Y L Xu X L Ding andW J Dai ldquoAn integratedGPSndashaccelerometer data processing technique for structural

1

09

08

07

06

05

04

03

02

01

00 5 10

Frequency (Hz)15 20

SmartphoneAccelerometerGPS (vertical)

Figure 13 Dynamic frequencies in terms of periodogram (verticaldirection)

Advances in Civil Engineering 13

deformation monitoringrdquo Journal of Geodesy vol 80 no 12pp 705ndash719 2006

[3] S B Im S Hurlebaus and Y J Kang ldquoSummary review ofGPS technology for structural health monitoringrdquo Journal ofStructural Engineering vol 139 no 10 pp 1653ndash1664 2011

[4] M R Kaloop and H Li ldquoMonitoring of bridge deformationusing GPS techniquerdquo KSCE Journal of Civil Engineeringvol 13 no 6 pp 423ndash431 2009

[5] M Kaloop E Elbeltagi J Hu and A Elrefai ldquoRecent advancesof structures monitoring and evaluation using GPS-time seriesmonitoring systems a reviewrdquo ISPRS International Journal ofGeo-Information vol 6 no 12 p 382 2017

[6] X Li C Rizos L Ge Y Tamura and A Yoshida ldquo+ecomplementary characteristics of GPS and accelerometer inmonitoring structural deformationrdquo in Proceedings of the ION2005 National Technical Meeting Institute of Navigation(ION) San Diego CA USA January 2005

[7] P Psimoulis S Pytharouli D Karambalis and S StirosldquoPotential of Global Positioning System (GPS) to measurefrequencies of oscillations of engineering structuresrdquo Journalof Sound and Vibration vol 318 no 3 pp 606ndash623 2008

[8] C Watson T Watson and R Coleman ldquoStructural moni-toring of cable-stayed bridge analysis of GPS versus modeleddeflectionsrdquo Journal of Surveying Engineering vol 133 no 1pp 23ndash28 2007

[9] C Xiong and Y Niu ldquoInvestigation of the dynamic behavior ofa super high-rise structure using RTK-GNSS techniquerdquo KSCEJournal of Civil Engineering vol 23 no 2 pp 654ndash665 2018

[10] B Hofmann-Wellenhof H Lichtenegger and E WasleGNSSndashGlobal Navigation Satellite Systems GPS GLONASSGalileo and More Springer Science amp Business MediaVienna Austria 2007

[11] M Celebi ldquoGPS in dynamic monitoring of long-periodstructuresrdquo Soil Dynamics and Earthquake Engineeringvol 20 no 5ndash8 pp 477ndash483 2000

[12] D K Kim J I Kim and M Q Feng ldquoInstrumentation ofbridges for structural health monitoringrdquo KSCE Journal ofCivil Engineering vol 5 no 3 pp 231ndash242 2001

[13] G Q Wang D M Boore H Igel and X Y Zhou ldquoSomeobservations on colocated and closely spaced strong ground-motion records of the 1999 Chi-Chi Taiwan earthquakerdquoBulletin of the Seismological Society of America vol 93 no 2pp 674ndash693 2003

[14] J Hwang H Yun S K Park D Lee and S Hong ldquoOptimalmethods of RTK-GPSaccelerometer integration to monitorthe displacement of structuresrdquo Sensors vol 12 no 1pp 1014ndash1034 2012

[15] S Dashti J D Bray J Reilly S Glaser A Bayen and E MarildquoEvaluating the reliability of phones as seismic monitoringinstrumentsrdquo Earthquake Spectra vol 30 no 2 pp 721ndash7422014

[16] M Feng Y Fukuda M Mizuta and E Ozer ldquoCitizen sensorsfor SHM use of accelerometer data from smartphonesrdquoSensors vol 15 no 2 pp 2980ndash2998 2015

[17] X Zhao R Han Y Ding et al ldquoPortable and convenient cableforce measurement using smartphonerdquo Journal of CivilStructural Health Monitoring vol 5 no 4 pp 481ndash491 2015

[18] X Zhao K Ri R Han Y Yu M Li and J Ou ldquoExperimentalresearch on quick structural health monitoring technique forbridges using smartphonerdquoAdvances inMaterials Science andEngineering vol 2016 no 1 pp 1ndash14 2016

[19] Y Ding R Han H Liu S Li X Zhao and Y Yu ldquoBridgeinspection and management system based on smart phonerdquoin Proceedings of the ASME 2016 Conference on Smart

Materials Adaptive Structures and Intelligent SystemsAmerican Society of Mechanical Engineers Stowe VT USASeptember 2016

[20] X Zhao R Han K J Loh B Xie J Li and J Ou ldquoShakingtable tests for evaluating the damage features under earth-quake excitations using smartphonesrdquo in Proceedings of theHealth Monitoring of Structural and Biological Systems XIIInternational Society for Optics and Photonics Denver COUSA May 2018

[21] F Moschas and S Stiros ldquoMeasurement of the dynamicdisplacements and of the modal frequencies of a short-spanpedestrian bridge using GPS and an accelerometerrdquo Engi-neering Structures vol 33 no 1 pp 10ndash17 2011

[22] M R Kaloop ldquoBridge safety monitoring based-GPS tech-nique case study Zhujiang Huangpu bridgerdquo Smart Struc-tures and Systems vol 9 no 6 pp 473ndash487 2012

[23] G W Roberts X Meng and A H Dodson ldquoIntegrating aglobal positioning system and accelerometers to monitor thedeflection of bridgesrdquo Journal of Surveying Engineeringvol 130 no 2 pp 65ndash72 2004

[24] Accelerometer Analyzer App 2017 httpsplaygooglecomstoreappsdetailsidcomlulaccelerometeramphles

[25] X Meng ldquoReal-time deformationmonitoring of bridges usingGPSaccelerometersrdquo Doctoral dissertation University ofNottingham UK 2002

[26] G E Vazquez-Becerra J R Gaxiola-Camacho R BennettG M Guzman-Acevedo and I E Gaxiola-CamacholdquoStructural evaluation of dynamic and semi-static displace-ments of the Juarez bridge using GPS technologyrdquo Mea-surement vol 110 no 1 pp 146ndash153 2017

[27] P Psimoulis I Peppa L Bonenberg S Ince and X MengldquoCombination of GPS and RTS measurements for themonitoring of semi-static and dynamic motion of pedestrianbridgerdquo in Proceedings of the 3rd Joint International Sym-posium on Deformation Monitoring Vienna Austria April2016

[28] F Moschas and S Stiros ldquoNoise characteristics of high-frequencyshort-duration GPS records from analysis ofidentical collocated instrumentsrdquo Measurement vol 46no 4 pp 1488ndash1506 2013

[29] S W Smithgte Scientist amp Engineerrsquos Guide to Digital SignalProcessing California Technical Publishing San Diego CAUSA 1997

[30] J Wang X Meng C Qin and J Yi ldquoVibration frequenciesextraction of the forth road bridge using high sampling GPSdatardquo Shock and Vibration vol 2016 Article ID 980786118 pages 2016

[31] J Yu XMeng X Shao B Yan and L Yang ldquoIdentification ofdynamic displacements and modal frequencies of a medium-span suspension bridge using multimode GNSS processingrdquoEngineering Structures vol 81 no 1 pp 432ndash443 2014

[32] E Elbeltagi M Kaloop and M Elnabwy ldquoMonitoring andassessment of Mansoura railway steel bridge using RTK-GPStechniquerdquo in Proceedings of the 2nd Joint InternationalSymposium on Deformation Monitoring pp 9ndash11 Notting-ham England September 2013

[33] N Raziq ldquoGPS structural deformation monitoring the mid-height problemrdquo Doctoral dissertation University ofMelbourneAustralia 2008

[34] S S HaykinAdaptive Filtergteory Pearson Education India5th edition 2014

[35] L H Estey and C M Meertens ldquoTEQC the multi-purposetoolkit for GPSGLONASS datardquo GPS Solutions vol 3 no 1pp 42ndash49 1999

14 Advances in Civil Engineering

deformation monitoringrdquo Journal of Geodesy vol 80 no 12pp 705ndash719 2006

[3] S B Im S Hurlebaus and Y J Kang ldquoSummary review ofGPS technology for structural health monitoringrdquo Journal ofStructural Engineering vol 139 no 10 pp 1653ndash1664 2011

[4] M R Kaloop and H Li ldquoMonitoring of bridge deformationusing GPS techniquerdquo KSCE Journal of Civil Engineeringvol 13 no 6 pp 423ndash431 2009

[5] M Kaloop E Elbeltagi J Hu and A Elrefai ldquoRecent advancesof structures monitoring and evaluation using GPS-time seriesmonitoring systems a reviewrdquo ISPRS International Journal ofGeo-Information vol 6 no 12 p 382 2017

[6] X Li C Rizos L Ge Y Tamura and A Yoshida ldquo+ecomplementary characteristics of GPS and accelerometer inmonitoring structural deformationrdquo in Proceedings of the ION2005 National Technical Meeting Institute of Navigation(ION) San Diego CA USA January 2005

[7] P Psimoulis S Pytharouli D Karambalis and S StirosldquoPotential of Global Positioning System (GPS) to measurefrequencies of oscillations of engineering structuresrdquo Journalof Sound and Vibration vol 318 no 3 pp 606ndash623 2008

[8] C Watson T Watson and R Coleman ldquoStructural moni-toring of cable-stayed bridge analysis of GPS versus modeleddeflectionsrdquo Journal of Surveying Engineering vol 133 no 1pp 23ndash28 2007

[9] C Xiong and Y Niu ldquoInvestigation of the dynamic behavior ofa super high-rise structure using RTK-GNSS techniquerdquo KSCEJournal of Civil Engineering vol 23 no 2 pp 654ndash665 2018

[10] B Hofmann-Wellenhof H Lichtenegger and E WasleGNSSndashGlobal Navigation Satellite Systems GPS GLONASSGalileo and More Springer Science amp Business MediaVienna Austria 2007

[11] M Celebi ldquoGPS in dynamic monitoring of long-periodstructuresrdquo Soil Dynamics and Earthquake Engineeringvol 20 no 5ndash8 pp 477ndash483 2000

[12] D K Kim J I Kim and M Q Feng ldquoInstrumentation ofbridges for structural health monitoringrdquo KSCE Journal ofCivil Engineering vol 5 no 3 pp 231ndash242 2001

[13] G Q Wang D M Boore H Igel and X Y Zhou ldquoSomeobservations on colocated and closely spaced strong ground-motion records of the 1999 Chi-Chi Taiwan earthquakerdquoBulletin of the Seismological Society of America vol 93 no 2pp 674ndash693 2003

[14] J Hwang H Yun S K Park D Lee and S Hong ldquoOptimalmethods of RTK-GPSaccelerometer integration to monitorthe displacement of structuresrdquo Sensors vol 12 no 1pp 1014ndash1034 2012

[15] S Dashti J D Bray J Reilly S Glaser A Bayen and E MarildquoEvaluating the reliability of phones as seismic monitoringinstrumentsrdquo Earthquake Spectra vol 30 no 2 pp 721ndash7422014

[16] M Feng Y Fukuda M Mizuta and E Ozer ldquoCitizen sensorsfor SHM use of accelerometer data from smartphonesrdquoSensors vol 15 no 2 pp 2980ndash2998 2015

[17] X Zhao R Han Y Ding et al ldquoPortable and convenient cableforce measurement using smartphonerdquo Journal of CivilStructural Health Monitoring vol 5 no 4 pp 481ndash491 2015

[18] X Zhao K Ri R Han Y Yu M Li and J Ou ldquoExperimentalresearch on quick structural health monitoring technique forbridges using smartphonerdquoAdvances inMaterials Science andEngineering vol 2016 no 1 pp 1ndash14 2016

[19] Y Ding R Han H Liu S Li X Zhao and Y Yu ldquoBridgeinspection and management system based on smart phonerdquoin Proceedings of the ASME 2016 Conference on Smart

Materials Adaptive Structures and Intelligent SystemsAmerican Society of Mechanical Engineers Stowe VT USASeptember 2016

[20] X Zhao R Han K J Loh B Xie J Li and J Ou ldquoShakingtable tests for evaluating the damage features under earth-quake excitations using smartphonesrdquo in Proceedings of theHealth Monitoring of Structural and Biological Systems XIIInternational Society for Optics and Photonics Denver COUSA May 2018

[21] F Moschas and S Stiros ldquoMeasurement of the dynamicdisplacements and of the modal frequencies of a short-spanpedestrian bridge using GPS and an accelerometerrdquo Engi-neering Structures vol 33 no 1 pp 10ndash17 2011

[22] M R Kaloop ldquoBridge safety monitoring based-GPS tech-nique case study Zhujiang Huangpu bridgerdquo Smart Struc-tures and Systems vol 9 no 6 pp 473ndash487 2012

[23] G W Roberts X Meng and A H Dodson ldquoIntegrating aglobal positioning system and accelerometers to monitor thedeflection of bridgesrdquo Journal of Surveying Engineeringvol 130 no 2 pp 65ndash72 2004

[24] Accelerometer Analyzer App 2017 httpsplaygooglecomstoreappsdetailsidcomlulaccelerometeramphles

[25] X Meng ldquoReal-time deformationmonitoring of bridges usingGPSaccelerometersrdquo Doctoral dissertation University ofNottingham UK 2002

[26] G E Vazquez-Becerra J R Gaxiola-Camacho R BennettG M Guzman-Acevedo and I E Gaxiola-CamacholdquoStructural evaluation of dynamic and semi-static displace-ments of the Juarez bridge using GPS technologyrdquo Mea-surement vol 110 no 1 pp 146ndash153 2017

[27] P Psimoulis I Peppa L Bonenberg S Ince and X MengldquoCombination of GPS and RTS measurements for themonitoring of semi-static and dynamic motion of pedestrianbridgerdquo in Proceedings of the 3rd Joint International Sym-posium on Deformation Monitoring Vienna Austria April2016

[28] F Moschas and S Stiros ldquoNoise characteristics of high-frequencyshort-duration GPS records from analysis ofidentical collocated instrumentsrdquo Measurement vol 46no 4 pp 1488ndash1506 2013

[29] S W Smithgte Scientist amp Engineerrsquos Guide to Digital SignalProcessing California Technical Publishing San Diego CAUSA 1997

[30] J Wang X Meng C Qin and J Yi ldquoVibration frequenciesextraction of the forth road bridge using high sampling GPSdatardquo Shock and Vibration vol 2016 Article ID 980786118 pages 2016

[31] J Yu XMeng X Shao B Yan and L Yang ldquoIdentification ofdynamic displacements and modal frequencies of a medium-span suspension bridge using multimode GNSS processingrdquoEngineering Structures vol 81 no 1 pp 432ndash443 2014

[32] E Elbeltagi M Kaloop and M Elnabwy ldquoMonitoring andassessment of Mansoura railway steel bridge using RTK-GPStechniquerdquo in Proceedings of the 2nd Joint InternationalSymposium on Deformation Monitoring pp 9ndash11 Notting-ham England September 2013

[33] N Raziq ldquoGPS structural deformation monitoring the mid-height problemrdquo Doctoral dissertation University ofMelbourneAustralia 2008

[34] S S HaykinAdaptive Filtergteory Pearson Education India5th edition 2014

[35] L H Estey and C M Meertens ldquoTEQC the multi-purposetoolkit for GPSGLONASS datardquo GPS Solutions vol 3 no 1pp 42ndash49 1999

14 Advances in Civil Engineering

[36] G W Roberts C J Brown X Meng and P R B DallardldquoUsing GPS to measure the deflections and frequency re-sponses of the London millennium bridgerdquo in Bridge DesignConstruction and Maintenance Institution of Civil EngineersLondon UK 2015 httpswwwicevirtuallibrarycomdoi101680bdcam359350053

[37] X Meng A H Dodson and GW Roberts ldquoDetecting bridgedynamics with GPS and triaxial accelerometersrdquo EngineeringStructures vol 29 no 11 pp 3178ndash3184 2007

[38] D McCallen A Astaneh-Asl S Larsen and L HutchingsldquoDynamic response of the suspension spans of the sanFrancisco-Oakland bay bridgerdquo in Proceedings of the 8th USNational Conference on Earthquake Engineering San Fran-cisco CA USA April 2006

Advances in Civil Engineering 15

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom