Exploring the Determinants of Vulnerable Road Users’ Crash Severity in State Roads

DETERMINANTS OF VULNERABLE ROAD USERS’ CRASH SEVERITY IN STATE ROADS

Álvaro CaviedesMiguel Figliozzi

January 19th

TABLEOF CONTENTS 1. BACKGROUND

2. PROBLEM STATEMENT3. DATA AND METHODS4. RESULTS5. CONCLUSIONS

BACKGROUND

1. Department of Transportation’s (DOT) interest in increasing non-motorized transportation

BACKGROUND

1. Focus on driver safety2. Crash frequency vs crash severity

BACKGROUND

1. Urban environments2. Risk factors:• Location,• Environmental, • Crash, • Road,

• Demographic, and • Traffic characteristics.

PROBLEM STATEMENTExplore risk factors of crash severity for pedestrians and bicyclists in the Oregon Highway Network System

DATA AND METHODS

1. Oregon statewide crash database (2007-2014)2. ODOT’s TransGIS database3. Neighborhood concepts (Currans et al. 2015)

INJURY SEVERITYLEVELS A (INCAPACITATED)- ALMOST KILLED

K (FATAL) - KILLED

B (VISIBLE INJURY) – BRUISE

C – COMPLAIN OF PAIN

0 – ONLY PROPERTY DAMAGE

CRASHES IN THE OREGON HIGHWAY STATE NETWORK

VARIABLE

PEDESTRIANCRASHESIN OREGON

ONLY STATE HIGHWAYS

BICYCLISTCRASHESIN OREGON

ONLY STATE HIGHWAYS

TOTAL CRASHES

6,162 1,840(30%)

7,147 1,584(22%)

CRASHES AT INTERSECTIONS

3,629 1,088 4,702 1,045

CRASHES AT SEGMENTS

1,822 561 864 169

OTHERS 711 191 1481 370

DESCRIPTIVE ANALYSIS

1. Location characteristics2. Environmental conditions3. Crash characteristics4. Traffic characteristics5. Road characteristics

LOCA

TION

CHAR

ACTE

RISTIC

SDE

SCRI

PTIV

E AN

ALYS

ISLOCATION LAND USE

PEDESTRIANBICYCLIST

0% 5%

10% 15% 20% 25%

Intersection Segment

K A

0% 5%

10% 15% 20%

Intersection Segment

K A

0% 5%

10% 15% 20% 25%

Urban Suburban Rural

K A

0% 5%

10% 15% 20%

Urban Suburban Rural

K A

(n=496) (n=1039)

(n=133) (n=867)

(n=602) (n=724) (n=209)

(n=372) (n=487) (n=141)

ENVIR

ONME

NTAL

COND

ITION

SDE

SCRI

PTIV

E AN

ALYS

IS LIGHT CONDITIONS WEATHER

PEDESTRIANBICYCLIST

0%

10%

20%

30%

Daylight Dark + streetlight

Dark Twilight

K A

0%

10%

20%

30%

Daylight Dark + streetlight

Dark Twilight

K A

0%

10%

20%

30%

Clear day Bad conditions

K A

0%

10%

20%

30%

Clear day Bad conditions

K A

(n=775) (n=410) (n=228) (n=122)

(n=786) (n=113) (n=42) (n=59)

(n=856) (n=679)

(n=749) (n=251)

CRAS

H CH

ARAC

TERIS

TICS

DESC

RIPT

IVE

ANAL

YSIS ALCOHOL INTOXICATION USER LOCATION

PEDESTRIANBICYCLIST

0%

10%

20%

30%

No alcohol Yes alcohol

K A

0%

10%

20%

30%

No alcohol Yes alcohol

K A

0%

10%

20%

30%

Crosswalk Roadway Midblock

K A

0%

10%

20%

30%

Crosswalk Roadway Bike lane

K A

(n=1292) (n=243)

(n=961) (n=39)

(n=1042) (n=452) (n=41)

(n=423) (n=541) (n=36)

CRAS

H CH

ARAC

TERIS

TICS

DESC

RIPT

IVE

ANAL

YSIS VEHICLE TYPE VEHICLE MOVEMENT

PEDESTRIANBICYCLIST

0%

10%

20%

30%

40%

Passenger car Heavy vehicle

K A

0% 10% 20% 30% 40%

Passenger car Heavy vehicle

K A

0% 10% 20% 30% 40%

Straight Turning

K A

0% 10% 20% 30% 40%

Straight Turning

K A

(n=1496) (n=39)

(n=989) (n=11)

(n=893) (n=642)

(n=393) (n=607)

TRAF

FIC CO

NDITI

ONS

DESC

RIPT

IVE

ANAL

YSIS

AADT (ONLY TRUCKS)

PEDESTRIAN BICYCLIST

0% 5%

10% 15% 20% 25% 30%

<1.5k 1.5k-5k 5k-7.5k >7.5k

K A

0% 5%

10% 15% 20% 25% 30%

<1.5k 1.5k-5k 5k-7.5k >7.5k

K A(n=1107) (n=418) (n=9) (n=1) (n=716) (n=271) (n=8) (n=5)

ROAD

CHAR

ACTE

RISTIC

SDE

SCRI

PTIV

E AN

ALYS

ISROAD SURFACE POSTED SPEED LIMIT

PEDESTRIANBICYCLIST

0% 20% 40% 60% 80%

100%

<20 20-35 35-50 50-65 >65

K A

0% 20% 40% 60% 80%

100%

<20 20-35 35-50 50-65 >65

K A

0% 5%

10% 15% 20%

Dry Wet

K A

0% 5%

10% 15% 20%

Dry Wet

K A

(n=55) (n=1032) (n=316) (n=131) (n=1)

(n=26) (n=616) (n=239) (n=114) (n=5)

(n=1075) (n=460)

(n=862) (n=138)

ROAD CHARACTERISTICS –EXPOSURE ANALYSISDESCRIPTIVE ANALYSIS

- Crash risk under various road characteristics- Likelihood of crash involvement - Depends on exposure- Risk ratio ~ concentration of crashes (number of lanes, road width, and road

classification)- Exposure controlled by estimating the proportion of VMT

Example: 𝑅𝑖𝑠𝑘𝑟𝑎𝑡𝑖𝑜 = +,-.,/012,34.-156,57/-31859706:;1/,+,-.,/012,34<=>7/-31859706:;1/,

EXPO

SURE

ANAL

YSIS

-ROA

D CH

ARAC

.DE

SCRI

PTIV

E AN

ALYS

IS

0

1

2

3

4

5

6

7

8

9

10

1 2 3 4

0

1

2

3

4

5

6

7

8

9

10

1 2 3 4

0

1

2

3

4

5

6

7

8

9

10

Local Collector Arterial

0

1

2

3

4

5

6

7

8

9

10

Local Collector Arterial

0

1

2

3

4

5

6

7

8

9

10

10 - 20 20 - 30 30 - 40 40 - 50

0

1

2

3

4

5

6

7

8

9

10

10 - 20 20 - 30 30 - 40 40 - 50

NUMBER OF LANES ROAD WIDTH ROAD CLASSIFICATIONRIS

K RAT

IORIS

K RAT

IO

RISK R

ATIO

RISK R

ATIO

RISK R

ATIO

RISK R

ATIO

K A B C O

1 2 3 4

1 2 3 4

10-20 20-30 30-40 40-50

10-20 20-30 30-40 40-50

Local Collector Arterial

Local Collector Arterial

PEDESTRIANBICYCLIST

10 10 10

101010

00 0

000

RESULTS

1. METHODOLOGY2. INDIVIDUAL MODEL3. POOLED MODELS4. SENSITIVITY ANALYSIS

POOLEDMODELS

POOLEDMODELS

POOLEDMODELS

POOLEDMODELS

Other significant variables:- Land use- Segment vs Inters.- Time of the day

POOLEDMODELSOnly road and traffic characteristics

POOLEDMODELSOnly road and traffic characteristics

POOLEDMODELSOnly road and traffic characteristics

SENSITIVITY ANALYSISPEDESTRIAN MODEL

CRASH SEVERITY ~ Light conditions +Road classification +Posted speed limit +Pedestrian location+AADT(Only truck)

SENSITIVITY ANALYSISPEDESTRIAN MODEL

CRASH SEVERITY ~ BASELINE SCENARIO

Prob. (K+A)=7.4%

Light conditions (DAYLIGHT) +Road classification (LOCAL STREET) +Posted speed limit (<50 MPH) +Pedestrian location (CROSSWALK) +AADT(Only truck) (<700)

SENSITIVITY ANALYSISPEDESTRIAN MODEL

CRASH SEVERITY ~ Light conditions (DARKNESS) +Road classification (LOCAL STREET) +Posted speed limit (<50 MPH) +Pedestrian location (CROSSWALK) +AADT(Only truck) (<700)

Prob. (K+A)=7.4%Prob. (K+A)=16.1% Change = +8.6%

BAD LIGHTING SCENARIO VS BASELINE SCENARIO

% CH

ANGE

PEDESTRIAN MODELSENSITIVITY ANALYSIS

59%

9% 7% 7% 4% 3% 0%

20% 40% 60% 80%

100%

K+A

SENSITIVITY ANALYSISBICYCLIST MODEL

CRASH SEVERITY ~ Bicyclist location+AADT

% CH

ANGE

13%

8%

3%

0%

5%

10%

15%

20%

Worst case scenario Location: Segment AADT: >10.000

K+A

CONCLUSIONS

1. TAKEAWAYS- Age- Alcohol intoxication- Vehicle size- Vehicle movement

- Posted speed limit- Light conditions- Road classification- Road surface- Location of the user- Land use and AADT

CONCLUSIONS

2. MITIGATION- Educational campaigns- Training courses- More strict enforcement- Protection vulnerable users

CONCLUSIONS

3. LIMITATIONS and FUTURE RESEARCH- Underreporting- Speed- Bike facilities- Special pedestrian signals

QUESTIONS?ACKNOWLEDMENTS

DISSERTATION COMMITTEE:- Miguel Figliozzi- Chris Monsere- Avinash Unnikrishnan

TTP LABTransportation engineering and planningMaster students