Zoo Biology 28 : 259–270 (2009)

TECHNICAL REPORT

Development of a Noncontact andLong-Term Respiration MonitoringSystem Using Microwave Radar forHibernating Black BearSatoshi Suzuki,� Takemi Matsui, Hiroshi Kawahara, and Shinji Gotoh

Faculty of System Design, Tokyo Metropolitan University, Tokyo, Japan

The aim of this study is to develop a prototype system for noncontact,noninvasive and unconstrained vital sign monitoring using microwave radar andto use the system to measure the respiratory rate of a Japanese black bear (Ursusthibetanus japonicus) during hibernation for ensuring the bear’s safety. UenoZoological Gardens in Tokyo planned to help the Japanese black bear (female,approximately 2 years of age) going into hibernation. The prototype system has amicrowave Doppler radar antenna (10-GHz frequency, approximately 7mWoutput power) for measuring motion of the body surface caused by respiratoryactivity without making contact with the body. Monitoring using this system wasconducted from December 2006 to April 2007. As a result, from December 18,2006, to March 17, 2007, similar behaviors reported by earlier studies wereobserved, such as sleeping with curled up posture and not eating, urinating ordefecating. During this hibernation period and also around the time ofhibernation, the prototype system continuously measured cyclic oscillations.The presence of cyclic vibrations at 8-sec intervals (about 7 bpm) was confirmedby the system before she entered hibernation on December 3, 2006. Therespiratory rate gradually decreased, and during the hibernation period therespiratory rate was extremely low at approximately 2 bpm with almost nochange. The results show that motion on the body surface caused by respiratoryactivity can be measured without touching the animal’s body. Thus, themicrowave radar employed here can be utilized as an aid in observing vital signsof animals. Zoo Biol 28 : 259–270, 2009. r 2009 Wiley-Liss, Inc.

Published online 4 February 2009 in Wiley InterScience (www.interscience.wiley.com).

DOI 10.1002/zoo.20229

Received 19 March 2008; Revised 8 December 2008; Accepted 29 December 2008

�Correspondence to: Satoshi Suzuki, Faculty of System Design, Tokyo Metropolitan university, 6-6

Asahigaoka, Hino, Tokyo 191-0065, Japan. E-mail: ssuzuki@sd.tmu.ac.jp, ssuzuki@cc.tmit.ac.jp.

rr 2009 Wiley-Liss, Inc.

Keywords: noncontact; monitoring; microwave radar; respiratory rate; hibernation;Japanese black bear

INTRODUCTION

Seasonal hibernation, which is a characteristic behavior of some mammals, hasmany unclarified points despite considerable research. As for physiologicalapproaches, cardiac function was previously investigated to determine associationsbetween hibernation state, heart rate and systolic and diastolic myocardialperformance [Nelson et al., 2003]. In addition, the degree of disuse atrophy wasalso observed by monitoring hindleg flexion of bears in the field during late autumnafter denning and in early spring [Harlow et al., 2001], and the body surfacetemperature of hibernating bears was further investigated to monitor periodicmuscle activity [Harlow et al., 2004]. Moreover, a recent study reported a specificprotein as a candidate hormone responsible for inducing hibernation [Kondo et al.,2006]. Thus, clarification of the mechanism of hibernation is expected to contributeto the welfare and healthcare of human beings in the future [Hellgren, 1998].

On the other hand, in 2005, Ueno Zoological Gardens in Tokyo made plans toassist a Japanese black bear (Ursus thibetanus japonicus) to enter hibernation, becausethe bear showed extremely slow movements in winter. Moreover, the staff wantedvisitors to understand this instinctive behavior of bears as occurring in the wild. Thisattempt to assist the bear started from 2002 and the construction of new facilities forhibernation was completed in 2006. The bear’s condition during hibernation had tobe carefully observed to avoid the risk of long-term fasting; however, staff agreed thatit was advisable to avoid attaching any devices to the bear in the hibernating booth sothat visitors could appreciate the natural state of the animal. Therefore, to observethe physiological condition of the bear during hibernation, a night-vision camera wasset up in the hibernation booth. However, the monitoring of vital signs of thehibernating bear using a camera alone is not sufficient to prevent against long-termfasting; therefore, they looked for other methods.

In recent years, noninvasive sensing techniques have been developed formeasuring human vital signs. Several attempts at noninvasive pulse monitoring havebeen conducted using a strain gauge [Ciaccio et al., 2007] and pressure sensors[Jacobs et al., 2004]. Wang et al. [2006] reported a method for measuring heart beatand the respiratory rate from the subject’s back using a polyvinylidene fluoride filmpiezoelectric polymer sensor. We previously reported the use of this noncontactmethod for monitoring the heart rate and the respiratory rate of New Zealandrabbits [Matsui et al., 2004a,b]. Furthermore, we have developed a noncontactmethod using a ceiling-attached microwave radar to monitor the respiratory rate ofhuman subjects in a bed covered with transmitting bedding [Uenoyama et al., 2006].Additionally, for measuring the vital signs of casualties inside an isolation unit, wealso developed a noncontact monitoring system for the heart rate using a microwaveradar antenna with a frequency of 1.215MHz [Matsui et al., 2006]. The aim of thismethod was to measure extremely minute scales of motion appearing on the bodysurface caused by cardiac and respiratory motions. This method was originallydeveloped for searching for survivors under earthquake rubble [Chen et al., 1986,

260 Suzuki et al.

Zoo Biology

2000]. A microwave radar has the following characteristics: (1) it is possible to detectmovements of an object from some distance without actually touching it, and (2)microwaves can be transmitted through objects except metals or water. If we attemptto use this system in humans, it is possible to measure motion on the body surfacefrom some distance without removing the clothing. In this regard, we hypothesizedthat the respiratory rate of a hibernating bear can be monitored by applying thismethod, avoiding touching the animal.

The aim of this study was to develop a prototype noncontact, noninvasive andunconstrained vital sign monitoring system equipped with a microwave radar formonitoring the respiratory rate of a hibernating bear during hibernation period.Utilizing the acquired data, the system will be evaluated to determine whether it canmeasure physiological changes during hibernation without touching the animal’sbody.

MATERIALS AND METHODS

Preparation for Hibernation

The newly built facilities for hibernation consisted of three rooms: a holdingroom (12.93m2; 2.75m wide, 4.70m long and 2.82m high) for visitors, a preparationroom for hibernation (5.59m2; 2.33m wide, 2.40m long and 2.82m high) and ahibernation booth (2.19m2; 1.75m wide, 1.25m long and 1.68m high). Thepreparation room has air conditioners (LSVLP3A, Daikin Co., Tokyo Japan) thatcan control temperature from �5 to 201C. The preparation room had sufficientwater and dry hay for spreading in the hibernation booth before and duringhibernation. The small door from the preparation room leading to the hibernationbooth was always kept open during the hibernation period so that the bear coulddrink water and voluntarily discontinue its hibernation.

For the test animal, a female Japanese black bear, named as ‘‘Kuu,’’approximately 2 years of age, which was captured at Tohoku mountainarea in Japan in 2005, was selected for this project. From the beginning ofSeptember 2006, zoo staff started giving her large amounts of feed in order tostore body fat, with the largest amount of about 5 kg (potatoes (1.5 kg), carrots(2 sticks), pelleted feed (780 g), five apples, four mandarin oranges, acorns, oakleaves, etc.) given per day from October 22 to November 12. The staff graduallydecreased the amount of feed after November 13 because the bear left some feed andshowed slower movements on November 12. The amount of feed was initiallydecreased to 75% of that of the normal amount of feed from November 21, then to50% from November 27 and 25% from December 9. Feed was stopped onDecember 13.

The target weight gain was 20% higher than the original weight to compensatewhat some research refers to as weight loss caused by hibernation [Folk et al., 1972;Nelson et al., 1973; Folk, 1967]. The weight increased from 44.8 to 57.6 kg, 20%higher than that before going into hibernation.

The temperature and the lighting of the holding and the preparation rooms werecontrolled from December 3. The rooms for hibernation equipped with refrigerationand lighting cycle systems were built in the enclosure. The temperature of the holdingand preparation rooms was gradually lowered to �51C using an air conditioner, and

261Noncontact Respiration Monitoring

Zoo Biology

that of the hibernating booth was lowered to 31C from December 16. The amount oflight in the holding and preparation rooms was lowered to create winter conditionsunder where she was captured at Tohoku mountain area in Japan. The amount oflight in the hibernation booth was not controlled and was always dark (Fig. 1).

System Design of Prototype Noncontact Vital Sign Monitoring System

The prototype noncontact vital sign monitoring system consisted of threedevices: (1) a microwave Doppler radar antenna, (2) a device for controlling powersupply to the microwave Doppler radar antenna and (3) an analyzing system. Thefrequency of the radar for respiration monitoring was 10GHz, with a normal outputpower of 7mW (maximum 10mW), antenna gain of 10 dBi and diffusion angle ofapproximately 401. The system was developed based on the principle of measuringmotion appearing on the surface of the abdominal part of the body caused byrespiratory motions (Fig. 2). This 10-GHz frequency was within the confines offrequency limit in accordance with the domestic law on radio waves in Japan and theoutput power was much lower than the numeric value established in the guidelinesfor protection from radio waves in Japan.

Acquisition

The 10-GHz microwave Doppler radar antenna was attached to the ceiling 1.68mabove the floor surface of the hibernation booth. Data on the measured respiratory ratewas relayed through the local area network (LAN) and zoo staff confirmed the rateusing a PC web browser. A night-vision camera (TK-S549, Victor Co. Tokyo, Japan)was also set up by attaching to the ceiling in the hibernation booth; the condition of theblack bear in the booth could be checked through this night-vision camera on a TVmonitor (Fig. 3). The monitoring continued from December 1, 2006, to April 4, 2007,about 3 months for hibernation and 2 weeks before and after the hibernation to observethe process of entering and recovering from hibernation (Fig. 1).

Settings and Analysis

Our prototype system is a Doppler-type radar system that can measure changes ofvelocity of a targeted moving object. The difference between targeted motion induced byrespiration and another movement involves the periodicity and the volume. There is a

12/3

Temperature / Light control

VTR / Respiration monitoringusing the prototype system

HibernationReactivationperiod

2006/ 2007/

Preparationperiod

1 2 3 4

Fig. 1. Time schedule of hibernation and items for control and monitoring. Temperature andlighting of the three rooms for hibernation were controlled from December 3. Video tape recordingand respiration monitoring using the prototype system were performed from December 3.

262 Suzuki et al.

Zoo Biology

1.68m40°

0.35m

Ceiling

Floor

JapaneseBlackBear

Swollen abdomen caused by respiration

Dry hay

Microwave Dopplerradar antenna

Fig. 2. Setting position of the 10-GHz microwave Doppler radar antenna in the hibernationbooth and principle of measurement. The microwave Doppler radar antenna was constructedbased on the principle of measuring motion appearing on the surface of the abdominal part ofthe body caused by respiratory motions.

HDD2

HDD1

A/DBoard

Controller

Backup

Networkrouter

LAN

Microwave Doppler radar antennain the ceiling attached

RespiratorySignal

Night-vision camera

1.68

m

TV monitor

PC/Analyzing System

Fig. 3. Schematic diagram of devices for the prototype noncontact vital sign monitoringsystem. The 10-GHz microwave Doppler radar antenna and a night-vision camera wereattached to the ceiling (1.68m high) of the hibernation booth. The output signal from themicrowave radar antenna was sent to a PC for analyzing the signal. The measured respiratoryrate was transmitted through LAN for staff to observe the bear’s condition via a PC. Visitorscould also observe the bear’s condition in the booth on a TV monitor through a night-visioncamera. LAN, local network area.

263Noncontact Respiration Monitoring

Zoo Biology

technical problem with this type of radar system separating each component related toindividual motion. However, it is possible to determine the features periodically usinginformation from the time component in the output signal by time–frequency analysise.g. using fast Fourier transform (FFT). We aimed only to capture the featureperiodically induced by respiration. Data were acquired at a sampling frequency of100Hz. A band-pass filter was used for transferring data from the 10-GHz microwaveDoppler radar antenna in order to reduce noise and select data related to respiratorypeaks. To date, few surveys have been conducted regarding the respiration of animalsduring hibernation, although it has been shown that the heart rate of black bearsdecreases from 40bpm in summer down to 10 or 8bpm during hibernation [Folk et al.,1972; Nelson et al., 1973]. Here, we simply assumed that the respiratory rate alsodecreased at the same rate as the change in the heart rate. Therefore, the frequency ofthe analyzing system was set between the lower cut-off frequency of 0.03Hz and theupper cut-off frequency of 0.25Hz, and this setting sufficiently covers the range from 1.8to 15bpm for measuring respiration.

After band-pass filtering, FFT was conducted with a normal Hanning window toobtain the respiratory rate. It was thought that the maximum value of the FFTspectrum indicated the cyclic oscillation caused by the respiratory movement of thebody surface similar to the case of humans. The respiratory rate was counted backwardfrom the frequency corresponding to this maximum value of the FFT spectrum withinthe range from 0.03 to 0.25Hz.

All acquisitions and analytical steps proceeded using real-time sampling and adisplay program employing a self-produced system by LabVIEW (NationalInstruments, Co., TX), as shown in Fig. 4.

RESULTS

Observation of the Bear’s Behavior

Feeding was stopped on December 13 and the temperature of the preparationroom was set to �51C using an air conditioner. From December 17, the bear rakedup the dry hay into the hibernation booth and slept curled up in a ball with her legsplaced under the body in the center of the booth. Figure 5 shows a video image

Fig. 4. Screen image of the analyzing system of the prototype noncontact vital signmonitoring system.

264 Suzuki et al.

Zoo Biology

obtained using the night-vision camera at 18:00 on December 26, 2006. She mostlyremained with curled up posture and moved rarely until mid-March. This statecontinued for 89 days and the bear slept with curled up posture and neither urinatednor defecated during this period. On March 17, 2007, excrement was found and thebear was observed to leave the hibernation booth for a short period in the daytime.

Respiratory Rate

Figure 6a shows sample data acquired by the prototype system at 23:50 onDecember 3, 2006, when the bear was not yet in hibernation. The presence of cyclicvibrations occurring at 7-sec intervals, 8 bpm, was recognized and these consecutivepeaks were observed when the bear slept in the booth before, during and also afterhibernation; therefore, it was thought that the reflected microwave amplitude containedcyclic oscillations caused by the bear’s respiratory motion. Figure 6b shows sample dataduring the hibernation period, at 23:10 on February 8, 2007. Three cyclic oscillationswith 15-sec intervals in 1min, 3bpm, were recognized in the signal. It means that therespiratory rate during the hibernation period was lower and the intervals between cyclicoscillations caused by respiratory motion during the hibernation were longer than thoseat normal condition. This indicates that the prototype system using a microwaveDoppler radar was able to measure the changes in the respiratory rate of the twoperiods. On the other hand, our prototype system monitors the motion speed. If the bearmoves, the system monitor cannot track the respiratory rate. The targeted bear movedin the daytime prior to hibernation and slept stably and long at night. She was alreadyactive in the daytime of December 3, though her movements were slow. She stepped intothe booth for hibernation only during the night; therefore, we could monitor therespiratory rate only at night. We thought that the data should be compared under thesame conditions and we presented the respiratory rates taken at night in Figure 6. Inaddition, this prototype system is a Doppler-type radar system that can measurechanges of velocity of a targeted moving object. As the output power is quite small andthe frequency quite high, the signal transmitted by the antenna is clearly reflected on the

Fig. 5. Video image of a female Japanese black bear ‘‘Kuu’’ in the hibernation boothcaptured using a night-vision camera (at 18:00 on December 26, 2006).

265Noncontact Respiration Monitoring

Zoo Biology

body surface of the targeted test animal. If the targeted test animal does not changeposition or move any limbs, the system captures only the motion speed on the bodysurface induced by physiological activity. Therefore, if the value of this signal wereintegrated, we could determine theoretically the change of distance from the radarantenna to the body surface, making it possible to determine any of the bear’srespiratory movements by comparing waveforms. However, it is very difficult tomeasure the distance with our current technology. In fact, we have attempted tocalculate the distance, but the estimated values did not match exactly with the truevalues. In addition, we could not find a consistent and regular pattern because it variedwith time. As a consequence, we decided to discuss only the changes of respiratory rateas a temporal feature rather than waveforms as a feature of the amplitude.

Figure 7a shows the change of respiratory rate at 10-min intervals for 3 consecutivedays from February 8 to 10. This result confirms the trend that the respiratory rate islower at night than in the daytime, and the trend fluctuated slightly between the days.Subsequently, in order to determine the trend for diurnal changes, the respiratory ratewas calculated at 10-min intervals for 7 days from February 8 to 14 by averaging thevalues obtained. The result in Figure 7b shows a gentle slope upward from left to rightfrom about 16:00 to 18:00. The trend is similar to a circadian rhythm.

Figure 8 shows the changes in the average daily respiratory rate during about 3months. This figure demonstrates that the respiratory rate decreased while the bear wasentering hibernation, and became extremely low at approximately 2bpm and showedalmost no change. OnMarch 17, 2007, the respiratory rate increased coinciding with the

0

(a)

(b)

1

2

3

4

5

6

0

Am

plitu

de (

V)

Am

plitu

de (

V)

0

1

2

3

4

5

6

Time (sec)

10 20 30 40 50 60

0

Time (sec)

10 20 30 40 50 60

1 2 5 6 7 83

1 2 3

4

Fig. 6. Sample data of an output signal over a 1-min period derived from the microwaveDoppler radar antenna before hibernation and during the hibernation period. (a) Eight cyclicoscillations at 7-sec intervals were observed (at 23:50 on December 3, 2006). (b) Three cyclicoscillations with 15-sec intervals were confirmed in the signal (at 23:10 on February 8, 2007).

266 Suzuki et al.

Zoo Biology

final temperature and lighting control. Two more weeks were needed for the respiratoryrate to return to the same level as that on December 3, 2006.

DISCUSSION

Based on our observations, it was thought that the hibernating period of theblack bear lasted for 89 days from December 18, 2006, to March 17, 2007, startingfrom the time when the animal spent most of the day sleeping and from the time

1

(a)

(b)

1.5

2

2.5

3

Res

pira

tory

Rat

e (b

pm)

2/8 2/9 2/10

1.5

1.7

1.9

2.1

2.3

2.5

0:00

2:00

4:00

6:00

8:00

10:00

12:00

14:00

16:00

18:00

20:00

22:00 0:0

0

Time

0:00

2:00

4:00

6:00

8:00

10:00

12:00

14:00

16:00

18:00

20:00

22:00 0:0

0

Time

Res

pira

tory

Rat

e (b

pm)

Fig. 7. Sample data of diurnal variation during the hibernation period. (a) Change ofrespiratory rate at 10-min intervals for 3 consecutive days (February 8–10, 2007). (b) Averagerespiratory rate at 10-min intervals for 7 days (February 8–14, 2007).

0

2

4

6

8

10

12

12/1

Date

Ave

rage

of

resp

irat

ory

rate

in a

day

(bp

m)

12/13

Dropping start

12/18 3/17Hibernating period

1/1 2/1 3/4 4/4

Fasting period 3/17

Fig. 8. Average of the estimated respiratory rate. On March 17, 2007, during the maskingperiod, excrement was found and the bear was observed to leave the hibernation booth for ashort period in the daytime. This state continued for 89 days up to March 17, where the bearslept with curled up posture and neither urinated nor defecated during this period.

267Noncontact Respiration Monitoring

Zoo Biology

excrement was found. During this period, similar behaviors reported by earlierstudies were also observed, such as sleeping with curled up posture [Nelson et al.,1973; Folk, 1967] and not eating [Folk et al., 1972; Nelson et al., 1973; Watts andCuyler, 1988], urinating or defecating [Nelson et al., 1973; Folk, 1967].

In addition, we tried to monitor the respiratory activity observed on the bodysurface without directly touching the bear’s body by applying the same devices usedin humans. This investigation showed that the respiratory rate decreased by about2 bpm during the hibernation period. The respiratory rate before the hibernatingperiod starting from December was approximately 10 bpm and decreased by about80% in the hibernation period. According to early research studies, the heart rate inthe hibernation period is reduced to about one-fourth or one-fifth [Folk, 1967] of thenormal level, and these results and our findings are almost identical. After March 17,the respiratory rate increased gradually and returned to the same level as at thebeginning of December, suggesting that the changes in reduction of metabolism inthe cardiovascular and respiratory systems were the same.

Though there is a tendency that the rate of body weight loss during hibernationis higher in the wild than in captivity, a 15–20% body weight loss during hibernationwas observed in the case of an American black bear [Folk et al., 1972]. In anotherreport, a 7–28% weight loss was observed in a simulated den [Watts and Cuyler,1988]. The Japanese black bear in our research lost 10.9 kg, about 18.9% of bodyweight, during hibernation. This percentage weight loss is in the range reportedpreviously.

On the other hand, Nelson et al. [1973] reported that the respiratory quotienttaken prior to hibernation decreased from 0.78 to values approaching 0.60 duringhibernation, before returning to 0.80 after the bears became awake and active. Fat isconsidered to be the main source of metabolic energy during hibernation. If the totalreduction in body weight in our investigation is assumed to result from fatmetabolism, where the calorific value of fat is approximately 9.1 kcal/g, it isestimated that an average of 1,114.5 kcal/day was consumed during the hibernatingperiod taken as 89 days. This earlier research indicated the relationships betweenchange of body weight and metabolism, and the calories consumed per day, whichwe estimated here in a black bear with a body weight of around 50 kg, which is abody weight in the range of earlier research [e.g. Watts and Cuyler, 1988]. Averagecalorie intake from spring to autumn was about 4,500 kcal; the estimated valueduring hibernation was reduced to one-fourth. The amount of decrease intemperature was 5–71C below active body temperature of approximately 381C[Harlow et al., 2004; Hellgren, 1998; Hissa et al., 2004], which is about a 10–15%reduction. Therefore, it is likely that these calories are used mainly for basalmetabolism such as regulation of constant body temperature.

In addition, the cyclic oscillation induced by respiratory activity acquired fromour prototype system was 2–3 bpm on average. The oscillation prior to hibernationwas about 10 bpm, a reduction to about one-fourth. The reduction rate of calorieswas roughly in accordance with the rate of decrease in the respiration rate. Thereduction rate estimated from our investigation is acceptable and comparable withthat indicated in the earlier study [Folk, 1967]. It demonstrates a proportionalrelationship between the change of calories and the change of respiratory rate duringhibernation. Additionally, the rate of weight loss was almost equal to the rate ofreduction in respiration. Therefore, this prototype system using a microwave

268 Suzuki et al.

Zoo Biology

Doppler radar was able to accurately measure the motion on the body surfacecaused by the respiratory activity, and the respiratory rate of the bear duringhibernation was obtained without touching the animal’s body. In this study, weshowed that it is possible to monitor changes in the animal’s physiological conditionwhile keeping it safe during hibernation without directly applying any special devicesto its body, as carried out in an artificial raising environment.

CONCLUSION AND FUTURE WORK

We constructed a prototype system for monitoring the physiological changes inthe bear without touching the animal’s body directly in order to keep the bear safeduring hibernation. The results show that the respiratory rate can be measured over along period without touching the animal’s body using our microwave Doppler radar.Additionally, we found using the microwave radar system that there is a diurnaltrend during hibernation, similar to a circadian rhythm. We cannot confirm whetherit is a circadian rhythm at this stage; however, the possibility of the existence of acircadian rhythm cannot be denied, as a diurnal cyclic change was confirmed in ourdata.

This attempt to assist the bear to enter hibernation and to show how theanimal hibernates proved to be popular among visitors; thus, we plan to furtherinvestigate other changes in the physiological condition of the animal duringhibernation in the future. We previously reported the monitoring of the heart rate inhumans using the same type of microwave Doppler radar (frequency of 24GHz andoutput power of 7mW). Although we also set this 24-GHz microwave Doppler radarantenna 10 cm under the floor of the booth during this attempt and tried to analyzecardiac motion from the output of this antenna during the hibernation period, wecould not clearly measure cardiac motion, probably because the microwaves had nosufficient gain and power for them to be transmitted through the dry hay. However,there were some periods when the heart rate was observed to be similar to thatreported by a previous study [Folk, 1967]. In addition, the changes in the heart ratein a day in these periods were similar to the changes in the circadian rhythm [Folket al., 1972]. Therefore, further investigation is necessary but our system appearspromising for noncontact, noninvasive and unconstrained monitoring of not onlythe respiratory rate but also the heart rate of hibernating bears.

ACKNOWLEDGMENTS

The authors make a most cordial acknowledgment to Teruyuki Komiya, theDirector of Ueno Zoological Gardens in Tokyo, Japan. We thank Kazuyoshi Ito,General Curator, Mitsuhiro Terada, Curator, and keepers Keiko Ide, YoshinoriKojima, Mariko Saito, Yu Mizumoto and Ko Fujioka in Ueno Zoological Gardens,Tokyo, for their support and cooperation during this research.

REFERENCES

Chen KM, Misra D, Wang H, Chuang HR,Postow E. 1986. An X-band microwave life-detection system. IEEE Trans Biomed Eng33:697–702.

Chen KM, Huang Y, Zhang J . 2000. Microwavelife-detection systems for searching human sub-jects under earthquake rubble or behind barrier.IEEE Trans Biomed Eng 27:105–113.

269Noncontact Respiration Monitoring

Zoo Biology

Ciaccio EJ, Hiatt M, Hegyi T, Drzewiecki GM.2007. Measurement and monitoring of electro-cardiogram belt tension in premature infants forassessment of respiratory function. Biomed EngOnline 6:1–11.

Folk Jr GE. 1967. Physiological observations ofsubarctic bears under winter den conditions. In:Fisher K, editor. Mammalian hibernation. NewYork: Elsevier. p 75–85.

Folk Jr GE, Folk MA, Minor JJ. 1972. Physiolo-gical condition of three species of bears in winterdens. Int Conf Bear Res Manage 2:107–124.

Harlow HJ, Lohuis T, Beck TD, Iaizzo PA. 2001.Muscle strength in overwintering bears. Nature409:997.

Harlow HJ, Lohuis T, Anderson-Sprecher RC,Beck TDI. 2004. Body surface temperature ofhibernating black bears may be related toperiodic muscle activity. J Mammal 85:414.

Hellgren EC. 1998. Physiology of hibernation inbears. Ursus 10:467–477.

Hissa R, Siekkinen J, Hohtola E, Saarela S,Hakala S, Pudas J. 2004. Seasonal patterns inthe physiology of the European brown bear(Ursus arctos arctos) in Finland. Comp BiochemPhysiol A Physiol 109:781–791.

Jacobs J, Embree P, Glei M, Christensen S,Sullivan P. 2004. Characterization of a novelheart and respiratory rate sensor. Conf ProcIEEE Eng Med Biol Soc 3:2223–2226.

Kondo N, Sekijima T, Kondo J, Takamatsu N,Tohya T, Ohtsu T. 2006. Circannual control ofhibernation by HP complex in the brain. Cell125:161–172.

Matsui T, Ishizuka T, Takase B, Ishihara M,Kikuchi M. 2004a. Non-contact determination

of vital sign alterations in hypovolaemic statesinduced by massive haemorrhage: an experimen-tal attempt to monitor the condition of injuredpersons behind barriers or under disaster rubble.Med Biol Eng Comput 42:807–811.

Matsui T, Hagisawa K, Ishizuka T, Takase B,Ishihara M, Kikuchi M. 2004b. A novel methodto prevent secondary exposure of medical andrescue personnel to toxic materials under bio-chemical hazard conditions using microwaveradar and infrared thermography. IEEE TransBiomed Eng 51:2184–2188.

Matsui T, Gotoh S, Arai I, Hattori H, Fujita M,Obara K, Masuoka K, Nakamura S, Takase B,Ishihara M, Kikuchi M. 2006. Noncontact vitalsign monitoring system for isolation unit(casualty care system). Mil Med 171:639–643.

Nelson RA, Wahner HW, Jones JD, Ellefson RD,Zollman PE. 1973. Metabolism of bears before,during, and after winter sleep. Am J Physiol224:491–496.

Nelson OL, McEwen MM, Robbins CT, FelicettiL, Christensen WF. 2003. Evaluation of cardiacfunction in active and hibernating grizzly bears. JAm Vet Med Assoc 223:1170–1175.

Uenoyama M, Matsui T, Yamada K, Suzuki S,et al. 2006. Non-contact respiratory monitoringsystem using a ceiling-attached microwave an-tenna. Med Biol Eng Comput 44:835–840.

Wang F, Tanaka M, Chonan S. 2006. Develop-ment of a wearable mental stress evaluationsystem using PVDF film sensor. J Adv Sci18:170–173.

Watts P, Cuyler C. 1988. Metabolism of the blackbear under simulated denning conditions. ActaPhysiol Scand 134:149–152.

270 Suzuki et al.

Zoo Biology