SUSTAINABLE ROAD SAFETY: A NEW NEIGHBOURHOOD

ROAD PATTERNDr. Gordon R. Lovegrove, P.Eng, PhDgord.lovegrove@ubc.caAssistant Professor, School of EngineeringUniversity of British Columbia, Kelowna, Canada

27 May,2010

University of British Columbia, Kelowna, Canada

• Research Assistants:

– Mr. James Sun, MASc, UBCO

– Ahsan Alam, UBCO

Acknowledgements

• Research Sponsors:

– Canada Mortgage & Housing Corporation

– Cities of Ottawa, Vancouver, and Victoria

– Natural Sciences & Engineering Research Council

2

Outline

• Motivation• Methodology• Methodology• Results• Conclusions• Questions?

3

� Each year more than 1.2 million people die

worldwide. (WHO, 2009)

Motivation

� One of the top ten leading “Causes of Death”.

o 9th in 2004 (WHO, 2004)

o 4th in 2020 (Gasper, 2004)

4

• Economic cost of road injuries varies from 1% to 2% ofGross National Product (GNP).

• Economic cost : US$ 518 billion annually worldwide.

Road Safety Burden

• Economic cost : US$ 518 billion annually worldwide.

(WHO 2004)

• The economic cost to Canadians is estimated to exceed$25 billion annually. (Transport Canada 2004)

• Traffic collisions account for around 50% of all accidentaldeaths. (Transport Canada,2004)

5

Canadian Road Safety

3.000

3.500

Observed

collisions

An

nu

al t

raff

ic fa

talit

ies

6

1.500

2.000

2.500

1990 1995 2000 2005 2010

collisions

Fatalities

trendline

Vision:2010

An

nu

al t

raff

ic fa

talit

ies

Year

Road Safety Improvement Programs (RSIPs)

7

•Transportation engineering programs focus on

road environment factors

• Traditional RSIP approach:

o Reactive

o But.... (disadvantages)

Road Safety Improvement Programs (RSIPs)

• Proactive road safety

o Focuses on predicting and improving the

safety of planned facilities

o Collision Prediction Models

8

• Relate Road collisions to land use/road patterns

• Traditional neighbourhood road patterns:

o Grid pattern

Loops and Culs-de-sac

Methodology

o Loops and Culs-de-sac

• Emerging patterns:

– Sustainable Road Safety (Dutch researchers)

– 3-way Offset (UBC researchers)

– Fused grid (CMHC researchers)

9

10

Fused Grid

11

o Sources:

o Vancouver , Victoria, Ottawa, Census

o Auto insurance claims, police records

Data

o Auto insurance claims, police records

o Types:

o Collisions(severe, total), GIS, demographics,

traffic volumes, mode splits, road network, signals

12

• Stratification

4 main themes related to neighbourhood traits

• Exposure: VKT, TLKM, VC

• Socio-Demographic: POPD, WKGD, FS, etc.

• TDM: TCM, DRIVE, SCC, etc.

Variables

• TDM: TCM, DRIVE, SCC, etc.

• Network: INTD, SIGD, I3WP, etc.

VKT = Vehicle Km Travelled

VC = average congestion level

TLKM = Total Lane Km

POPD = Population Density = population/ha

SCC = ShortCut Capacity

INTD = INTersection Density = Intersections/ha

I3WP = 3-way Intersection Percentage = I3W / INT 13

Models

14

• Regression Technique

o Generalized Linear Modelling (GLM) regression method

o Assumes a Negative Binominal error structure

• Model Form

Models

• Model Form

oo E(Λ) E(Λ) = predicted collision frequency= predicted collision frequency

oo aa00, a, a11, , bbjj= model parameters= model parameters

oo ZZ= leading exposure variables (e.g. VKT, TLKM)= leading exposure variables (e.g. VKT, TLKM)

oo XXjj= explanatory variables = explanatory variables (e.g. VC, POPD, INTD)(e.g. VC, POPD, INTD)

15

∑=Λ ii Xba eZaE 1)(o

• 32 models to analyze Victoria neighborhoods

• 32 models to analyze Ottawa

• 32 models to analyze Vancouver

Results

• 32 models to analyze Vancouver

• In each case, 16 Urban & 16 Rural models, at

least one in each model group

16

Models• Exposure:

• Socio-Demographic :

evc

VKTyrsCollisionsTotal3364.25685.03355.03/ =

17

• Socio-Demographic :

• TDM:

• Network:

ecorescvc

VKTyrsCollisionsTotal)0106.00134.0(915.01028.03/

−⋅=

ellkpwpidsigd

VKTyrsCollisionsTotal)0086.0300405.0int829.04246.0(8509.01757.03/

−−+⋅=

Results

18

Results256 hectare module

2.55

1.46

2.39

2

2.5

3

19

1.46

0.88 1

0

0.5

1

1.5

2

GridNetwork

Cul-de-sac

DutchSRS

3-wayOffset

FusedGrid

t-Statistics 2.57 1.88 2.41 1.15 --

(t8, 90% = 1.86)

• Land use patterns effects road collisions

– If we drive, we crash

• In traditional neighbourhoods

– short-cutting is the dominant cause of road collisions

Conclusions

– short-cutting is the dominant cause of road collisions

– Traffic calming is needed, or, retro-fit with new patterns

• New neighbourhood road patterns safer by 50%:

– 3-way Offset: Accessibility, mobility, & safety

– Fused Grid: Accessibility via off-road paths needs

– 30%-60% fewer road collisions – stay tuned!20

21

Fused Grid

22

SUSTAINABLE ROAD SAFETY: A NEW NEIGHBOURHOOD

ROAD PATTERNDr. Gordon R. Lovegrove, P.Eng, PhDgord.lovegrove@ubc.caAssistant Professor, School of EngineeringUniversity of British Columbia, Kelowna, Canada

27 May,2010

University of British Columbia, Kelowna, Canada