Run-Off-Road (ROR) Crashes: A Description of Conventional and Contemporary Countermeasures

Page | 1

Sohrab Siddiqui

Date: 04/18/2015

Run-Off-Road (ROR) Crashes: A Description of Conventional and Contemporary

Countermeasures

Page | 1

Contents Introduction ................................................................................................................................................. 1

Rural and Urban ROR Crashes ................................................................................................................. 6

Critical Reasons for ROR Crashes .......................................................................................................... 10

Critical Reasons Attributed to Drivers: ................................................................................................... 11

Driver Alcohol Presence ..................................................................................................................... 13

Critical Reasons Attributed to Vehicles .................................................................................................. 15

Critical Reasons Attributed to Environment ........................................................................................... 15

Critical Reasons for ROR Crashes involving Large Trucks ................................................................... 15

Crash-Associated Factors in Single-Vehicle ROR Crashes .................................................................... 17

Conventional Countermeasures for reducing ROR Crashes ................................................................ 21

Strategies to Keep Vehicles from Encroaching on the Roadside ............................................................ 24

Shoulder Rumble Strips: ..................................................................................................................... 24

Edgeline Rumble Strips for Roads with Narrow or Unpaved Shoulders: ........................................... 25

Midlane Rumble Strips: ...................................................................................................................... 25

Enhanced Delineation of Sharp Curves: ............................................................................................. 25

Improved Highway Geometry for Horizontal Curves: ....................................................................... 25

Enhanced Pavement Markings at Appropriate Locations ................................................................... 25

Skid-Resistant Pavements: .................................................................................................................. 26

Shoulder Treatments: .......................................................................................................................... 26

Minimize the Likelihood of Crashing into an Object or Overturning if the Vehicle Travels Off the

Shoulder .................................................................................................................................................. 28

Design Safer Slopes and Ditches to Prevent Rollovers: ..................................................................... 28

Remove/Relocate Objects in Hazardous Locations: ........................................................................... 28

Delineation of Roadside Objects: ....................................................................................................... 29

Reduce the Severity of the Crash ............................................................................................................ 29

Improve Design of Roadside Hardware or Application of Barrier and Attenuation Systems: ........... 29

Contemporary Countermeasures for reducing ROR Crashes ............................................................. 30

Effects of ABS and ESC on ROR Crashes ............................................................................................. 30

Effects of LDWS on ROR Crashes ......................................................................................................... 32

Conclusion / Discussion ............................................................................................................................ 33

Areas of Future Research ........................................................................................................................ 34

References .................................................................................................................................................. 34

Page | 1

Executive Summary A vehicle sometimes leaves the travel lane and encroaches onto the shoulder, median, roadside, parking

lane, gore, or a separator and hits one or more natural or artificial objects. This type of crash is called a run-

off-road (ROR) or Roadway Departure (RwD) crash. If this crash involves a single vehicle hitting objects

on the roadside or medians, it is called a single vehicle ROR (SVROR) crash. SVROR crashes result in

large proportion of fatalities and serious injuries in the United States. Around 70% of all fatal single vehicle

crashes are ROR crashes. Furthermore single vehicle ROR crashes accounts for more than 40% of total

fatalities in the United States.

Each year roadside crashes in the United States cost society an estimated $110 billion, killing approximately

15,000 people and injuring another 1,000,000. Besides the United States, ROR crashes are gaining attention

of other countries. According to World Health Organization (WHO, 2013), 1.24 million people die annually

due to road accident and this number is expected to rise by 67% in the year 2020. Of this 1.2 million

fatalities, single-vehicle ROR crashes constitute more than 35 percent.

This report has thoroughly investigated the problems associated with ROR crashes, their relative

importance, the critical reasons associated with pre-crash events, and contributing factors to the crashes.

Furthermore, a thorough review of the conventional and contemporary countermeasures are presented. A

comprehensive review of literature on run-off-road crashes was done in order to identify the reasons behind

ROR crashes and ways to mitigate the risks associated with ROR crashes. References are provided in the

last part of the report for the reader who is interested to know more about this issue.

The following are some highlights from this study:

Single vehicle ROR crashes constituted 40.6% of the fatal crashes, 20% of the injury crashes and

11.2% of the Property Damage Only (PDO) in the United States in the year 2012.

Of the fatal SVROR crashes, 37% occurred off roadway, 2% on shoulder and 4% on median.

The first harmful event for 25% of fatal SVROR crashes were trees and shrubberies followed by

rollover (19%) and embankment (10%). The most harmful event for more than 40% of fatal

SVROR crashes were rollovers followed by shrubberies/ trees (29%) and utility poles (8%).

Around 80% of fatal SVROR crashes occur on rural roadways of which 90% occur on two-lane

highways. In Montana which is mostly a rural states, 60% percent of all fatal crashes are SVROR

crashes. In rural counties of Montana, more than 65% percent of all fatal crashes are SVROR

crashes.

Of the critical reasons for pre-crash events in fatal and non-fatal ROR crashes, 95.1% is attributed

to drivers, 1% to vehicles and 1.1% to roadway and environment. This shows the importance of

finding countermeasures to minimize driver errors.

The most prominent driver errors include, internal distraction (15%), overcompensation (14.3%),

poor directional control (12.6%), too fast of the curve (11%) and driver sleeping (10.3%). Alcohol

use increases the overcompensation problem to 23.4% and directional control problem to 21.7%.

Brake failure (32.7%) and tire failure (25.6%) are the two common vehicle –attributed errors. Slick

road conditions contributed to more than 60% of environment-attributed errors for SVROR crashes.

Sleeping and heart-attack problems were the critical reason for almost 50% of truck-related

SVROR crashes.

Among the crash-associated factors (see the difference between critical reasons and factors in the

report) for SVROR crashes, driver inattention, driver fatigue and driver in a hurry carried an odds-

ratio of more than 3 (meaning that the odds of being involved in a crash is almost 3 times higher if

a factor is present than if the factor is not present).

Page | 1

Of the tried and proven conventional countermeasures for reducing ROR crashes, the most famous

are: shoulder rumble strips, improved horizontal curves, skid-resistant pavements, shoulder

treatments, safer sideslopes and ditches, improved roadside hardware and wider clearzones. Some

experimental countermeasures like edgeline rumble strips, midlane rumble strips, enhanced

pavement markings and enhanced delineation of sharp curves and roadside objects are also

described in this report.

Of the contemporary countermeasures, anti-lock braking system (ABS), electronic stability control

(ESC) and lane departure warning system (LDWS) together with their benefits are presented in this

report.

In the last part of the report, a conclusion / discussion of the topics described within the report and areas of

future research that needs to be expanded further in order to better understand ROR crashes and their

countermeasures are provided. The reader is persuaded to refer to the references in the final part of the

report if he/she wants to know more about ROR crashes.

Page | 1

List of Tables Table 1. Crashes by crash type, relation to roadway, and crash severity…………………………………….2

Table 2. Crashes by First Harmful Event, Manner of Collision, and Crash Severity………………………4

Table 3. Single Vehicle, Run-Off-the-Road Fatal and Total Crashes in Montana…………………………7

Table 4. Single Vehicle, Rural Run-Off-the-Road Fatal and Total Rural Crashes in Montana……………8

Table 5. Critical Reasons Coded for the Single-Vehicle ROR and “Other” Crash Events…………………11

Table 6. Critical Reasons for the Single-Vehicle ROR and “Other” Crash Events Attributed to Drivers…12

Table 7. Critical Reasons for the Single-Vehicle ROR Crash Events Attributed to Drivers with Versus

without the Presence of Alcohol in the Driver……………………………………………………………..14

Table 8. Critical Reasons for the Single-Vehicle ROR and “Other” Crash Events Attributed to Vehicles…15

Table 9. Critical Reasons for the Single-Vehicle ROR and “Other” Crash Events Attributed to

Environment……………………………………………………………………………………………….15

Table 10. Critical Reasons for the Large-Truck Single-Vehicle ROR and “Other” Crash Events…………16

Table 11. Crash-Associated Factors in Single-Vehicle ROR Crashes……………………………………18

Table 12. Logistic Regression Coefficients and Odds Ratios……………………………………………...21

Table 13. Benefits of using ABS and ESC in reducing ROR crashes………………………………………31

Page | 1

List of Figures Figure 1. Single-Vehicle ROR Crashes as a Percentage of All Fatal Crashes……………………………...1

Figure 2. Distribution of Fatal ROR Crashes…………………………………………………………….....2

Figure 3. Fatal Single Vehicle Crashes……………………………………………………………………..3

Figure 4. Fatal Multiple Vehicles Crashes………………………………………………………………….3

Figure 5. Distribution of Single-Vehicle ROR Fatalities for Two-Lane, Undivided, Noninterchange,

Nonjunction Roads by First Harmful Event………………………………………………………………5

Figure 6. Distribution of Single-Vehicle ROR Fatalities for Two-Lane, Undivided, Noninterchange,

Nonjunction Roads by Most Harmful Event………………………………………………………………..5

Figure 7. Distribution of Single-Vehicle ROR Fatalities on Two-Lane, Undivided, Noninterchange,

Nonjunction Roads by Highway Type……………………………………………………………………...6

Figure 8. Fatalities for Rural and Urban SVROR Crashes…………………………………………………7

Figure 9. Single-Vehicle ROR Crashes in Montana………………………………………………………..8

Figure 10. Single-Vehicle ROR Rural Crashes in Montana………………………………………………..8

Figure 11. Comparison of Statewide and Rural Fatal SVROR in Montana………………………………..9

Figure 12. SVROR Crash Percentages for States…………………………………………………………..9

Figure 13. ROR Crash Percentages for Montana Counties……………………………………………….11

Figure 14. Critical Reasons for SVROR Crashes and "Other" Crashes Comparison……………………..11

Figure 15. Driver-Related Error Categories for ROR and "Other" Crashes………………………………12

Figure 16. Top 5 Critical Reasons attributed to Drivers for SVROR Crashes……………………………13

Figure 17. Comparison of Driver-Related Error Categories for drivers with and without the presence of

Alcohol…………………………………………………………………………………………………….14

Figure 18. Comparison of Major Critical Reasons for the Passenger Vehicle and Large-Truck Single-

Vehicle ROR Crash Events………………………………………………………………………………..17

Figure 19. FHWA RwD Team strategies to mitigate most common roadway departure fatal and serious

injury crashes……………………………………………………………………………………………...23

Figure 20. Before and After Data for Selected Single-Vehicle ROR Crashes on the New York Thruway

(Source: New York State Police…………………………………………………………………………..24

Figure 21. Accident Modification Factor for Paved Shoulder Width (Relative to 6-Foot Paved Shoulder)

on Two-Lane Rural Highways…………………………………………………………………………….27

Figure 22. Accident Modification Factor for Shoulder Type on Two-Lane Rural Highways…………….27

Figure 23. Percent Reduction in Related Accident Types (i.e., ROR+ head-on+ sideswipe)…………….29

Figure 24. Lane Drift Warning System Functional Blocks…………………………………………….…32

Figure 25. Citroen Designed LDWS………………………………………………………………………32

Page | 1

Introduction American Association of State Highway and Transportation Officials (AASHTO) has embraced a vision of

“Towards Zero Deaths” and a goal of cutting fatalities in half by 2030. According to Federal Highway

Administration (FHWA, 2013), in order to accomplish this goal we need to reduce approximately 1,000

fatalities per year. Although many types of crashes contribute to the current fatality rates in the US, Run-

off-Road (ROR) crashes continue to account for more than 40% of the fatalities and fatal crashes.

According to FHWA, a roadway departure (RwD) crash or ROR crash is defined as a non-intersection crash

which occurs after a vehicle leaves the traveled way, crossing the center line of undivided highways, or

crossing an edge line (longitudinal pavement marking located at the edge of the traveled lane and the

shoulder) of the roadway.

Single vehicle ROR crashes which occurs as a result of a vehicle travelling off the roadway and hitting a

natural or artificial object has more fatalities involved than multiple vehicle ROR crashes. Multiple vehicle

ROR crash occurs when a vehicle encroaching on the median where the highway is separated or on the

opposite side when the vehicle crosses the opposing lanes of a nondivided highway hits a parked vehicle or

a vehicle on the opposing direction.

An estimated societal cost of $110 billion has been imposed each year due to roadside crashes (McGinnis

et al. 2001). These losses are equivalent to $1,600 per year for an average family of 4 persons. The

importance of the roadside safety problem has been recognized by different organizations, and efforts have

been made to reduce the types of errors most likely to cause roadside crashes.

NHTSA traffic safety facts reports provide the estimates for different types of crashes and fatalities resulted

from these crashes in the United States. The 1999 statistics from the Fatality Analysis Reporting System

(FARS) show that nearly 39 percent of the 37,043 fatal crashes were single-vehicle ROR crashes on various

road types. (See Figure 1)

Comparison of the statistics from 1999 and 2012 showed a slight increase in the number of single vehicle

ROR crashes. Based upon a compilation of motor vehicle crash data from the FARS and the General

Estimates System (GES), 30,800 fatal crashes, 1,634,000 injury crashes, and 3,950,000 property-damage-

only (PDO) crashes occurred on the U.S. highway system in 2012, totaling 5,614,800 crashes (see Table

1). It shows that of the 30,800 fatal crashes, 10,972 (35.6 percent) were single vehicle crashes that occurred

Sideswipe2%

Angle20%

SV ROR39%

Rear End5%

Head On14%

Other20%

Figure 1. Single-Vehicle ROR Crashes as a Percentage of All Fatal Crashes

Page | 2

off the roadway. An additional 474 fatal crashes (1.5 percent) occurred on the shoulder, and 1,071 (3.5

percent) occurred on the median. Thus, of the 30,800 fatal crashes, 12,517 crashes (40.6 percent) occurred

off the roadway, on the shoulder, or within the median. Of the injury crashes, 20 percent (327,000) were

single-vehicle crashes that occurred off the roadway, on the shoulder, or within the median. Of the property-

damage-only crashes, 11.2 percent (629,000) were single-vehicle crashes that occurred off the roadway, on

the shoulder, or within the median.

Figure 2 shows the distribution of fatal ROR crashes. As can be seen, ROR crashes contributes to 43 percent

of the total fatalities. Of this 43 percent, 41% is for the single vehicle ROR crashes (SVROR) and 2% for

multiple vehicle ROR crashes.

Table 1. Crashes by crash type, relation to roadway, and crash severity

Crash Types

Relation to Roadway

On

Roadway

Off

Roadway Shoulder Median Other/Unknown Total

Fatal Crashes

Single Vehicle 5890 10972 474 1071 298 18705

Multiple Vehicles 11515 284 115 156 25 12095

Total 17405 11256 589 1227 323 30800

Injury Crashes

Single Vehicle 162,000 278,000 7,000 42,000 37,000 525,000

Multiple Vehicles 1,097,000 5,000 1,000 4,000 2,000 1,109,000

Total 1,259,000 283,000 8,000 46,000 39,000 1,634,000

Property Damage Only Crashes

Single Vehicle 323,000 541,000 8,000 80,000 249,000 1,202,000


Total 3,052,000 550,000 10,000 85,000 253,000 3,950,000

All Crashes

Single Vehicle 490,890 829,972 15,474 123,071 286,298 1,745,705


Total 4,328,405 844,256 18,589 132,227 292,323 5,614,800

(Source: NHTSA Traffic Safety Facts 2012)

On Roadway56%

Off Roadway37%

Shoulder2%

Median4%

Other/Unknown1%

Figure 2. Distribution of Fatal ROR Crashes

Page | 3

Figure 3 and Figure 4 show the percentage of ROR and Non-ROR crashes involving single and multiple

vehicles. As can be seen, only 5 percent of crashes involve multiple vehicles while this percentage is 68%

for single vehicles.

These numbers show that SVROR crashes that occur off the roadway, on the shoulder or within the median

account for a significant portion of all accidents (17.2 percent). It is also evident from the higher percentage

of fatal and injury crashes that these crashes typically cause severe injuries or fatalities while not causing

many PDO’s.

FARS also provides the first harmful event that resulted in a fatal crash. According to this data for the year

2012, of the 30,800 fatal crashes, 10,347 (33.5 percent) were single-vehicle collisions with fixed objects;

while another 6,051 (19.6 percent) were single-vehicle collisions with objects that were not fixed. Overturns

caused 3,009 fatal crashes which is almost 10% of total fatal crashes. (See Table 2)

These tables do not indicate where these crashes occurred relative to junctions (i.e., whether the crashes

should be attributed to an intersection or to a roadway segment). Usually SVROR crashes occur on roadway

and they rarely occur on intersection. Table 2 shows a large percentage of crashes attributed to pedestrian

and pedalcyclists. This should always be the case for locations near or at intersections.

32%

68%

Figure 3. Fatal Single Vehicle Crashes

Non-ROR Crashes ROR Crashes

95%

5%

Figure 4. Fatal Multiple Vehicles Crashes

Non-ROR Crashes ROR Crashes

Page | 4

(Neuman et al., 2003) analyzed the extent of the SVROR problem specifically related to two-lane,

undivided, non-interchange, and nonjunction roadways using 1999 FARS data showing the first harmful

and the most harmful events in crash fatalities. Figure 5 shows that overturning which mostly results in

rollover crashes, trees, and embankments are the first harmful events for 54% of fatalities from SVROR

Table 2. Crashes by First Harmful Event, Manner of Collision, and Crash Severity

First Harmful Event

Crash Severity

Total Fatal Injury

Property Damage

Only

Number Percent Number Percent Number Percent Number Percent

Collision With Motor

Vehicle In Transport:

Angle 5,359 17.4 416,000 25.5 774,000 19.6 1,195,000 21.3

Rear End 1,827 5.9 518,000 31.7 1,327,000 33.6 1,847,000 32.9

Sideswipe 797 2.6 90,000 5.5 482,000 12.2 573,000 10.2

Head On 2,895 9.4 61,000 3.7 60,000 1.5 123,000 2.2

Other/Unknown 128 0.4 8,000 0.5 71,000 1.8 80,000 1.4

Subtotal 11,006 35.7 1,093,000 66.9 2,713,000 68.7 3,817,000 68

Collision With

Fixed Object:

Pole/Post 1,471 4.8 49,000 3 127,000 3.2 178,000 3.2

Culvert/Curb/Ditch 2,392 7.8 62,000 3.8 113,000 2.9 178,000 3.2

Shrubbery/Tree 2,561 8.3 46,000 2.8 61,000 1.6 110,000 2

Guard Rail 872 2.8 28,000 1.7 72,000 1.8 101,000 1.8

Embankment 1,105 3.6 20,000 1.2 25,000 0.6 47,000 0.8

Bridge 204 0.7 4,000 0.3 11,000 0.3 16,000 0.3

Other/Unknown 1,742 5.7 68,000 4.2 168,000 4.2 238,000 4.2

Subtotal 10,347 33.6 279,000 17.1 578,000 14.6 867,000 15.4

Collision With

Object Not Fixed:

Parked Motor Vehicle 313 1 40,000 2.4 278,000 7 318,000 5.7

Animal 171 0.6 13,000 0.8 258,000 6.5 271,000 4.8

Pedestrian 4,383 14.2 69,000 4.2 2,000 * 76,000 1.3

Pedalcyclist 719 2.3 49,000 3 6,000 0.2 56,000 1

Train 111 0.4 * * 1,000 * 1,000 *

Other/Unknown 354 1.1 12,000 0.7 60,000 1.5 72,000 1.3

Subtotal 6,051 19.6 183,000 11.2 604,000 15.3 794,000 14.1

Noncollision:

Rollover 3,009 9.8 73,000 4.4 37,000 0.9 112,000 2

Other/Unknown 363 1.2 6,000 0.4 18,000 0.5 24,000 0.4

Subtotal 3,372 10.9 78,000 4.8 55,000 1.4 137,000 2.4

Total 30,800 100 1,634,000 100 3,950,000 100 5,615,000 100

(Source: NHTSA Traffic Safety Facts 2012)

Page | 5

crashes. Figure 6 shows that the most harmful events for 70% of the fatalities from SVROR crashes are

trees and overturns.

0% 5% 10% 15% 20% 25% 30%

Shrubbery/Tree

Overturn

Embankment

Ditch

Utility Pole

Culvert

Other Fixed Object

Fence

Pole, post, supports

Guardrail

Other

Curb

Pedestrian

Other Object (Not Fixed)

Parked Motor Vehicle

Bridge Pier or Abutment

Boulder

Bridge Rail

Figure 5. Distribution of Single-Vehicle ROR Fatalities for Two-Lane, Undivided, Noninterchange, Nonjunction Roads by First Harmful Event

0% 5% 10% 15% 20% 25% 30% 35% 40% 45%

Shrubbery/Tree

Overturn

Embankment

Ditch

Utility Pole

Culvert

Other Fixed Object

Fence

Pole, post, supports

Guardrail

Other

Fire/Explosion

Pedestrian

Other Object (Not Fixed)

Parked Motor Vehicle

Bridge Pier or Abutment

Immersion

Building

Figure 6. Distribution of Single-Vehicle ROR Fatalities for Two-Lane, Undivided, Noninterchange, Nonjunction Roads by Most Harmful

Event

Page | 6

Rural and Urban ROR Crashes According to (FHWA, 2006), 80 percent of ROR fatalities occurred on rural roadways, with about 90

percent of those rural crashes occurring on two-lane highways alone. Based on FARS data, (Neuman et al.,

2003) analyzed how single vehicle ROR crashes are distributed by roadway functional classification. (See

Figure 7). There were more than four times as many ROR fatal crashes on rural roads than on urban roads

(Figure 8), partly due to the higher speeds on rural roads and to the greater mileage.

3%3%5%

7%

2%

8%

9%

13%

24%

26%

Figure 7. Distribution of Single-Vehicle ROR Fatalities on Two-Lane, Undivided, Noninterchange, Nonjunction Roads

by Highway Type

Urban Principal Arterial

Urban Collector

Urban Minor Arterial

Urban Local Road or Street

Unknown Rural

Rural Principal Arterial

Rural Minor Collector

Rural Minor Arterial

Rural Local Road or Street

Rural Major Collector

18%

82%

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Urban Roads Rural Roads

Figure 8. Fatalities for Rural and Urban SVROR Crashes

Page | 7

For the purpose of illustration, and in order to compare rural to total SVROR crashes, rates of fatal SVROR

crashes for rural Montana is compared to statewide fatal SVROR crash rates. According to Montana

Department of Transportation (MDT) office of safety management systems, in Montana almost 60 percent

of all fatal crashes are ROR crashes. Table 3 indicates these percentages from 2000 to 2009.

Table 3. Single Vehicle, Run-Off-the-Road Fatal and Total Crashes in Montana

Year All Crashes Fatal Crashes

Single Vehicle

Run-off-road

Crashes

All

Crashes

Percent of

All Crashes

Single Vehicle

Run-off-road

Fatal Crashes

All

Fatal

Crashes

Percent of

All Fatal

Crashes

2000 6882 22254 30.9 107 203 52.7

2001 6265 21846 28.7 122 201 60.7

2002 7211 23527 30.6 139 232 59.9

2003 7216 23160 31.2 144 239 60.3

2004 6395 21783 29.4 131 209 62.7

2005 6808 22376 30.4 139 224 62.1

2006 6727 22186 30.3 138 226 61.1

2007 6406 21829 29.3 154 249 61.8

2008 6740 21971 30.7 117 208 56.3

2009 6054 20967 28.9 117 198 59.1

(Source: MDT- Office of Safety Management Systems)

Figure 9 depicts the data in the table in a graphical format. As we can see, in 2009, although ROR crashes

constituted 28.9% of all crashes, 59.1% of them were involved in fatal crashes.

The same information is given for rural crashes in Montana in Table 4 and Figure 10. As it can be seen

from Figure 11, single vehicle ROR rural crashes account for an even higher percentage of all fatal crashes

in rural Montana.

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

Figure 9. Single-Vehicle ROR Crashes in Montana

Percent of All Crashes Percent of Fatal Crashes

Page | 8

Table 4. Single Vehicle, Rural Run-Off-the-Road Fatal and Total Rural Crashes in Montana

Year All Rural Crashes Fatal Rural Crashes

Single Vehicle

Run-Off-Road

Crashes

All Crashes Percent Of

All Crashes

Single Vehicle

Run-Off-Road

Fatal Crashes

All

Fatal

Crashes

Percent Of All

Fatal Crashes

2000 5889 11637 50.6 102 185 55.1

2001 5246 10452 50.2 118 187 63.1

2002 6033 11489 52.5 132 209 63.2

2003 6106 11746 52.0 139 214 65.0

2004 5409 10576 51.1 126 184 68.5

2005 5694 10934 52.1 130 194 67.0

2006 5648 10939 51.6 133 209 63.6

2007 5236 10573 49.5 144 230 62.6

2008 5384 10135 53.1 106 175 60.6

2009 4721 9097 51.9 110 180 61.1

(Source: MDT- Office of Safety Management Systems)

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

Figure 10. Single-Vehicle ROR Rural Crashes in Montana

Percent of All Crashes Percent of Fatal Crashes

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

Figure 11. Comparison of Statewide and Rural Fatal SVROR in Montana

Fatal Statewide ROR Crashes Fatal Rural SVROR Crashes

Page | 9

Montana usually ranks high in the nation as far as percentage of fatal crashes that are run-off-road.

According to MDT this is partly due to Montana not having a high amount of two or more vehicle fatal

crashes which mostly occur in more congested areas with many access points. Other possible factor

according to MDT can be long distances between communities which causes tiredness and will result in

driver making dangerous maneuvers. Speeding vehicles, driving under the influence, and adverse weather

conditions can also contribute to fatal ROR crashes.

Figure 12 shows the percentage of fatal SVROR crashes for entire United States. As can be seen, the

percentages are the highest for the rural states like Montana, Wyoming, South Dakota, West Virginia, and

Mississippi.

Figure 13 compares the percentages of fatal SVROR crashes for the state of Montana. As can be seen,

mostly rural counties have the highest percentage of fatal SVROR crashes.

Figure 12. SVROR Crash Percentages for States

Figure 13. ROR Crash Percentages for Montana Counties

Page | 10

Critical Reasons for ROR Crashes Many studies have identified the array of reasons for fatal and injurious ROR crashes. Many of these studies

are published in transportation safety journals like Accident Analysis and Prevention, Human Factors,

Transportation Research Part F, technical reports by FHWA and NHTSA, and technical reports by private

agencies like American Automobile Association, Foundation for Traffic Safety (AAA-FTS) and Crash

Avoidance Metrics Partnership (CAMP).

Data from National Motor Vehicle Crash Causation Survey (NMVCCS) is used in this study. While most

researchers have concentrated mainly on the factors which lead up to a crash, data from (NMVCCS) helps

to understand the entire crash envelope which comprises of a series of events which lead to crash

(NMVCCS, 2015). These events were captured by NHTSA’s National Center for Statistics and Analysis

(NCSA) from 2005 to 2007. The purpose of this survey was to collect on-scene information about the events

and associated factors leading up to crashes involving light vehicles. This survey contains four data

elements: (i) movement prior to critical pre-crash event (i.e., the movement of the vehicle immediately

before the occurrence of the critical event); (ii) critical pre-crash event (i.e., the circumstances that caused

the crash inevitable); (iii) critical reason for the critical pre-crash event (i.e., the immediate reason for the

critical event, which is often the last failure in the causal chain of events leading up to the crash); and (iv)

the crash-associated factors (i.e., the factors that are likely to add to the probability of crash occurrence).

NMVCCS data was used by (Liu and Ye, 2011) to identify the critical pre-crash event, the critical reasons

underlying the critical pre-crash event and associated factors present in the pre-crash phase of the single-

vehicle ROR crash. In order to better understand these chain of events and NMVCCS data, the following

example is provided to illustrate a SVROR crash and the manner it was coded in NMVCCS.

Case description: A crash involving a 2004 Subaru Forester (a compact SUV) occurred on the late weekday

afternoon on a dry roadway with a posted speed limit of 40 mph (64 km/h). The Subaru driver was an 82-

year-old man. He had some pre-existing physical or mental health condition and reported taking

drugs/medications in the past 24 hours. The driver tried to avoid a non-contact truck approached from the

opposite direction by steering right and the vehicle ran off the edge of the road on the right side. The vehicle

was equipped with ABS but not with ESC.

NMVCCS coding

Critical pre-crash event – “this vehicle ran off the edge of the road on the right side”

Critical reason for the critical event – “poor directional control (e.g., failing to control vehicle with

skill ordinarily expected)”

Crash-associated factors: pre-existing physical or mental health conditions; taking

drugs/medications in the 24 hours; attempted an avoidance maneuver by steering right; age; gender

…

This study by (Liu and Ye, 2011) is not limited to fatal crashes but consists of both fatal and non-fatal

crashes from NMVCCS database and thoroughly investigates driver-, vehicle-, and environment-related

factors, with a focus on driver-related factor due to physical and mental conditions, as well as driver’s

activities prior to the crash.

The term “other” in this report is generally referred to on-road crashes in which after the crash the vehicles

remained on the road. The critical pre-crash event for “other” crashes can include “vehicle loss of control

due to blow out/flat tire, poor road condition and other cause,” “vehicle turning at or passing through

Page | 11

intersection,” “pedestrian, pedal-cyclist, or other non-motorist in or approaching roadway,” and “animal in

or approaching roadway.”

According to statistics for critical reasons for SVROR crashes and “other” crashes (Table 5 and Figure 14)

64.4 percent (434,412) of the estimated 674,002 single-vehicle crashes were ROR crashes. From the critical

reasons for SVROR crashes, those attributed to drivers were dominating with 95.1%.

Critical Reasons Attributed to Drivers: Of the 95.1% of critical reasons that are attributed to drivers, the most frequently occurring errors are driver

performance errors (27.7%), followed by driver decision errors (25.4%), critical non-performance errors

(22.5%), and recognition errors (19.8%). In contrast for “Other” crashes, driver decision errors (59.7%) and

performance errors (26.3%) are the most frequently occurring errors. (See Figure 15). Table 6 and shows

the critical reasons involved in each of the four categories with the weighted frequency and percentages.

Table 5. Critical Reasons Coded for the Single-Vehicle ROR and “Other” Crash Events

Critical Reasons

Attributed to

ROR “Other”

Weighted

Frequency

Weighted

Percent

Weighted

Frequency

Weighted

Percent

Driver 413,070 95.10% 201,408 84.10%

Vehicle 4,456 1.00% 20,631 8.60%

Environment (roadway

and weather conditions) 4,950 1.10% 16,385 6.80%

Unknown reason for the

critical event 11,937 2.70% 1,087 0.50%

Critical reason not

coded to the vehicle 0 0.00% 80 0.00%

Total 434,412 Col 100%

239,590 Col 100%

Row 64.4% Row 35.6%

(Source: Liu and Ye, 2011)

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Driver Vehicle Environment

Figure 14. Critical Reasons for SVROR Crashes and "Other" Crashes Comparison

SVROR Crashes "Other" Crashes

Page | 12

Table 6. Critical Reasons for the Single-Vehicle ROR and “Other” Crash Events Attributed to Drivers

Critical Reasons ROR “Other”

Weighted

Frequency

Weighted

Percent

Weighted

Frequency

Weighted

Percent

Critical Non-

Performance

Errors

Sleeping/actually asleep 42,586 10.30% 886 0.40%

Heart attack/other physical impairment 29,226 7.10% 1,646 0.80%

Other/unknown critical nonperformance 20,961 5.10% 311 0.20%

Subtotal 92,773 22.50% 2,843 1.40%

Recognition

Errors

Internal distraction 62,048 15.00% 10,561 5.20%

External distraction 11,324 2.70% 591 0.30%

Inattention 5,644 1.40% 2,262 1.10%

Inadequate surveillance 1,651 0.40% 8,303 4.10%

Other/unknown recognition error 1,313 0.30% 545 0.30%

Subtotal 81,980 19.80% 22,262 11.00%

Decision

Errors

Too fast for curve 45,429 11.00% 39,813 19.80%

Too fast for conditions 27,983 6.80% 55,092 27.40%

Incorrect evasion 13,529 3.30% 8,626 4.30%

Aggressive driving 6,894 1.70% 4,813 2.40%

Too fast to be able to respond 5,314 1.30% 5,819 2.90%

Inadequate evasion 2,173 0.50% 2,450 1.20%

Other/unknown decision error 1,432 0.40% 1,644 0.80%

Illegal maneuver 203 0.10% 1,074 0.50%

Misjudgment of gap 915 0.20% 42 0.00%

Following too closely 334 0.10% 700 0.40%

Subtotal 104,206 25.40% 120,073 59.70%

Performance

Errors

Overcompensation 59,155 14.30% 31,410 15.60%

Poor directional control 51,991 12.60% 19,004 9.40%

Other/unknown performance error 2,088 0.50% 1,193 0.60%

Panic/freezing 1,149 0.30% 1,346 0.70%

Subtotal 114,383 27.70% 52,953 26.30%

Other/Unknown Driver Errors 19,726 4.80% 3,276 1.60%

Total 413,070 100% 201,408 100%


22.5%

19.8%

25.4%

27.7%

1.4%

11.0%

59.7%

26.3%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90%

Critical Non-Performance Errors

Recognition Errors

Decision Errors

Performance Errors

Figure 15. Driver-Related Error Categories for ROR and "Other" Crashes

ROR "Other"

Page | 13

As can be seen from the table above, the most frequently occurring critical reasons for non-performance

error category are attributed to driver being asleep during driving (10.3%) and heart attacks (7.1%). For

the recognition error category which means that the driver failed to correctly recognize the pre-crash

situation, internal distraction (15%) and external distraction (2.7%) are the most frequently occurring

critical reasons.

Among performance errors, “overcompensation” (14.3%) and “poor directional control” (12.6%) were the

top two critical reasons for SVROR crashes. Too fast for curves (11%), “too fast for conditions” (6.8%),

and “incorrect evasion” (3.3%) were the top three critical reasons for SVROR crashes among driver

decision errors.

Figure 16 shows the top 5 critical reasons attributed to drivers for SVROR Crashes. These results are

consistent with 2009 NHTSA report on fatal ROR crashes (Liu and Subramanian, 2009).

Driver Alcohol Presence

According to (Spainhour and Mishra, 2007) and (Liu and Subramanian, 2009), drivers with high levels of

blood-alcohol-content are more prone to ROR crashes than sober drivers. Driver alcohol presence is

described as an associated factor in SVROR crashes later in the paper. This section shows the result of

using alcohol on critical reasons attributed to drivers for SVROR crashes.

Table 7 compares the critical reasons for SVROR crashes attributed to drivers with and without the presence

of alcohol. As can be seen, with the presence of alcohol, the overcompensation problem has increased from

12.6% to 23.4%. Poor directional control has also increased from 11.2% to 21.7% with the presence of

alcohol.

Figure 17 compares four error categories for SVROR crashes attributed to drivers. As can be seen

performance errors have increased considerably from 24.4% to 46.9% with the presence of alcohol.

0% 2% 4% 6% 8% 10% 12% 14% 16%

Sleeping/actually asleep

Too fast for curve

Poor directional control

Overcompensation

Internal distraction

Figure 16. Top 5 Critical Reasons attributed to Drivers for SVROR Crashes

Page | 14

Table 7. Critical Reasons for the Single-Vehicle ROR Crash Events Attributed to Drivers With Versus

Without the Presence of Alcohol in the Driver

Critical Reasons

Alcohol Present Alcohol Not Present

Weighted

Frequency

Weighted

Percent

Weighted

Frequency

Weighted

Percent

Critical Non-

Performance

Errors

Sleeping/actually asleep 3,220 4.20% 37,795 12.20%

Heart attack/other physical impairment 794 1.00% 27,650 9.00%

Other/unknown critical nonperformance 8,377 11.00% 11,890 3.90%

Subtotal 12,391 16.20% 77,335 25.10%

Recognition

Errors

Internal distraction 10,578 13.90% 44,979 14.60%

External distraction 961 1.30% 10,070 3.30%

Inattention 0 0% 5,082 1.70%

Inadequate surveillance 0 0% 1,651 0.50%

Other/unknown recognition error 1,038 1.40% 275 0.10%

Subtotal 12,577 16.60% 62,057 20.20%

Decision

Errors

Too fast for curve 6,447 8.50% 36,331 11.80%

Too fast for conditions 5,436 7.10% 22,314 7.20%

Incorrect evasion 0 0% 13,529 4.40%

Aggressive driving 767 1.00% 6,127 2.00%

Too fast to be able to respond 718 0.90% 4,268 1.40%

Inadequate evasion 0 0% 2,009 0.70%

Other/unknown decision error 0 0% 1,432 0.50%

Illegal maneuver 0 0% 203 0.10%

Misjudgment of gap 0 0% 915 0.30%

Following too closely 0 0% 334 0.10%

Subtotal 13,368 17.50% 87,462 28.50%

Performance

Errors

Overcompensation 17,864 23.40% 39,057 12.60%

Poor directional control 16,564 21.70% 34,657 11.20%

Other/unknown performance error 1,391 1.80% 698 0.20%

Panic/freezing 0 0% 1,149 0.40%

Subtotal 35,819 46.90% 75,561 24.40%

Other/Unknown Driver Errors 2,129 2.80% 6,518 2.10%

Total 76,283 100% 308,932 100%


0%

5%

10%15%

20%

25%

30%

35%

40%

45%50%

Critical Non-Performance

Errors

Recognition Errors Decision Errors PerformanceErrors

Figure 17. Comparison of Driver-Related Error Categories for drivers with and without the presence of Alcohol

Alcohol Present Alcohol Not Present

Page | 15

Critical Reasons Attributed to Vehicles Table 8 shows the critical reasons for SVROR crashes attributed to vehicles. In ROR crashes, the most

frequently occurring critical reason attributed to vehicles was “brakes failed/degraded” (32.7%), followed

by “tires failed or degradation/wheel failed” (25.6%), “steering/suspension/transmission/engine failed”

(19.1%), and “other vehicle failure/deficiency” (18.8%). For “Other” single vehicle crashes “tires failed or

degradation/wheel failed” was the most frequently occurring vehicle-related critical reason with a

percentage of 71.7%.


Vehicles

Critical Reasons

ROR “Other”

Weighted

Frequency

Weighted

Percent

Weighted

Frequency

Weighted

Percent

Tires failed or degradation/wheel failed 1,142 25.60% 14,790 71.70%

Steering/suspension/transmission/engine failed 850 19.10% 3,272 15.80%

Brakes failed/degraded 1,457 32.70% 2,155 10.50%

Other vehicle failure/deficiency 839 18.80% 413 2.00%

Unknown vehicle failure 167 3.80% 0 0.00%

Total 4,456 100% 20,631 100%


Critical Reasons Attributed to Environment Table 9 shows the critical reasons for SVROR crashes attributed to environment (roadway and weather

conditions). Among the ROR crashes in which the critical reason was attributed to environment, about 96.5

percent were related to roadway conditions (slick roads, 64.3%; other highway-related conditions, 32.2%)

while only 3.5 percent to the weather (predominantly rain or snow.)

For “other” single vehicle crashes, 84.2% percent were related to roadway conditions and 15.8 percent to

weather conditions.


Environment

Critical Reasons

ROR “Other”

Weighted

Frequency

Weighted

Percent

Weighted

Frequency

Weighted

Percent

Roadway

Slick roads (ice, loose debris, etc.) 3,183 64.30% 11,942 72.90%

Other highway-related

(sign/signal/road design/view

obstructions, etc.) conditions

1,592 32.20% 1,843 11.30%

Subtotal 4,775 96.50% 13,785 84.20%

Weather

Rain/snow 174 18.80% 413 2.00%

Other weather-related

(fog/glare/wind, etc.) condition 0 0.00% 1,565 9.50%

Subtotal 174 3.50% 2,600 15.80%

Total 4,950 100% 16,385 100%


Critical Reasons for ROR Crashes involving Large Trucks Based on Large-Truck Crash Causation Study (LTCCS), a large truck is defined as a truck with a gross

weight rating of over 10,000 pounds. For the purpose of investigating the critical reasons for large truck

SVROR crashes, the data from LTCCS is used. Likewise NMVCCS, LTCCS collected the driver-, vehicle-

Page | 16

, and environment-related on-scene information from April 2001 to December 2003. The LTCCS collected

data on approximately 1,000 variables for each crash (FMCSA, 2006), (Starnes, 2006). Table 10 shows the

statistics of the critical reasons for large truck SVROR crashes and “other” crashes. As shown, only 37.2

percent of all single-vehicle ROR crashes involves large trucks.

Table 10. Critical Reasons for the Large-Truck Single-Vehicle ROR and “Other” Crash Events

Critical Reasons

ROR “Other”

Weighted

Frequency

Weighted

Percent

Weighted

Frequency

Weighted

Percent

Driver-Related

Too Fast for curve/turn 782 5.50% 7,719 32.30%

Sleeping /actually asleep 4,696 33.10% 180 0.80%

Too fast for conditions to be able to respond... 964 6.80% 1,494 6.20%

Inattention (i.e., daydreaming) 963 6.80% 1,290 5.40%

Heart attack or other physical impairment of the

ability... 2,114 14.90% 127 0.50%

Overcompensation 479 3.40% 1,117 4.70%

Poor directional control e.g., failing to control vehicle... 570 4.00% 943 3.90%

Inadequate surveillance (e.g., failed to look, looked

but...) 0 0.00% 1,376 5.80%

Internal distraction 125 0.90% 1,160 4.80%

Aggressive driving behavior 0 0.00% 814 3.40%

Other decision error 295 2.10% 190 0.80%

External distraction 272 1.90% 69 0.30%

Illegal maneuver 0 0.00% 157 0.70%

Following too closely to respond to unexpected actions 144 1.00% 0 0.00%

Other critical non-performance 138 1.00% 0 0.00%

Misjudgment of gap or other's speed 0 0.00% 62 0.30%

Type of driver error unknown 1,248 8.80% 69 0.30%

Unknown recognition error 396 2.80% 659 2.70%

Unknown critical non-performance 489 3.40% 0 0.00%

Vehicle-Related

Cargo shifted 280 2.00% 2,223 9.30%

Tires/wheels failed 0 0.00% 387 1.60%

Brakes failed 94 0.70% 230 1.00%

Steering failed 0 0.00% 69 0.30%

Degraded braking capability 0 0.00% 758 3.20%

Suspension failed 0 0.00% 791 3.30%

Environment (roadway and weather conditions)

Road design – other 0 0.00% 356 1.50%

Slick roads (low friction road surface due to ice...) 125 0.90% 115 0.50%

Wind gust 0 0.00% 127 0.50%

Road design - roadway geometry (e.g., ramp curvature) 2 0.00% 0 0.00%

Unknown reason for critical event 23 0.20% 0 0.00%

Critical event not coded to this vehicle 0 0.00% 1,447 6.10%

Total 14,198 Col. 100%

23,928 Col. 100%

Row 37.20% Row 62.80%


Page | 17

Figure 18 shows the comparison between the major critical reasons for passenger vehicle and large truck

SVROR crashes. As depicted, sleeping and physical impairment cause the majority of truck SVROR

crashes. These high rates are probably due to the fact that drivers of large trucks usually drive long-distances

on highways.

Crash-Associated Factors in Single-Vehicle ROR Crashes The NMVCCS data provides information about driver-, vehicle-, and environment-related factor present in

a pre-crash phase. In this section, the factors of interest include, driver inattention, driver alcohol presence,

driver fatigue status, driver’s gender, driver’s work-related stress or pressure, driver’s pre-existing physical

and/or mental health conditions, driver’s familiarity with the roadway, and whether the driver was in a

hurry.

To study the effect of each of crash-associated factors, (Liu and Ye, 2011) categorized single vehicle crashes

into two groups: the crashes in which the factor was present and those in which it was not present. Wald

chi square was used to study if the difference between them is statistically significant. Most authors refer

to statistically significant as P < 0.05 and statistically highly significant as P < 0.001 (less than one in a

thousand chance of being wrong).

Table 11 describes each of the above mentioned factors in details, studying their significance in the

occurrence of SVROR crashes.

0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50%

Sleeping /actually asleep

Heart attack or other physical impairment of the ability...

Internal distraction

Overcompensation

Poor directional control e.g., failing to control vehicle...

Too Fast for curve/turn

Inattention (i.e., daydreaming)

External distraction

Figure 18. Comparison of Major Critical Reasons for the Passenger Vehicle and Large-Truck Single-Vehicle ROR Crash Events

Large Trucks Passenger Vehicles

Page | 18

Table 11. Crash-Associated Factors in Single-Vehicle ROR Crashes

The NMVCCS data assesses a driver to be inattentive if he

or she was pre-occupied with concerns or the nature of

these concerns such as personal problems, family problems,

financial problems, preceding arguments, and future events

(e.g., vacation and wedding).

As can be seen, among the SV crashes that inattention was

the factor, 85.4% were SVROR crashes, while this

percentage when SV crash did not involve any inattention

was 57.1%.

Difference between these two percentages was statistically

significant at the 90 percent confidence level (𝜒2=4.23, p-

value=0.0622). So we can say that inattentive driving is

significantly associated with SVROR crashes.

NMVCCS data includes the information about “police

reported alcohol presence” that records the presence of

alcohol for the driver as reported by police in the police

accident report (PAR).

As can be seen, among the SV crashes that alcohol-

presence was the factor, 83.6% were SVROR crashes,

while this percentage when SV crash did not involve any

alcohol-presence was 60.9%.


highly significant at the 90 percent confidence level (𝜒2=

19.26, p-value= 0.0009). So we can say that driver alcohol-

presence is significantly associated with SVROR crashes.

Based on NMVCCS data fatigue is based on driver's current

and preceding sleep schedules, current and preceding work

schedules, and a variety of other fatigue-related factors

including recreational and non-work activities.

As can be seen, among the SV crashes that fatigue was the

factor, 83.9% were SVROR crashes, while this percentage

when SV crash did not involve any alcohol-presence was

55.5%.


significant at the 90 percent confidence level (𝜒2= 9.39, p-

value= 0.0098). So we can say that driver fatigue is


Dri

ver

Inat

tenti

on

85.457.1

14.642.9

0

20

40

60

80

100

Inattention (Total: 68,725) No Inattention Factors(Total: 242,111)

Single-Vehicle Crashes by Driver Inattention

ROR Other

83.660.9

16.439.1

0

20

40

60

80

100

Driver Alcohol Present(Total: 94,202)

No Driver Alcohol Present(Total: 529,985)

Single-Vehicle Crashes by Driver Alcohol Presence

ROR Other

83.9

55.5

16.1

44.5

0

20

40

60

80

100

Driver Fatigued (Total:115,196)

Driver Not Fatigued(Total: 372,208)

Single-Vehicle Crashes by Driver Fatigue Status

ROR Other

Dri

ver

Alc

ohol

Pre

sence

F

atig

ue

Page | 19

As can be seen, among the SV crashes that pre-existing

physical and mental health problem was the factor, 75.6 %

were SVROR crashes, while this percentage when SV crash

did not involve any pre-existing health problem was 58.7%.



25.3, p-value= 0.0003). So we can say that pre-existing

physical and mental health problem is significantly

associated with SVROR crashes.

As depicted, among the SV crashes with male drivers,

68.1% were SVROR crashes, while this percentage when

SV crash involving female was 61.1%.



5.17, p-value= 0.0421). So we can say that male drivers are

more prone to SVROR crashes than female drivers.

Driver’s self-reported driving frequency shows the

familiarity with the roadway in NMVCCS data. If driver

has used the highway on a daily or weekly basis or have

used it several times a month, it is defined as driver is

familiar with the roadway. If the driver has rarely used or

used the roadway for the first time, then it is reported as

driver’s unfamiliarity with the roadway.

As can be seen, among the SV crashes that driver’s

unfamiliarity with the roadway was the factor, 63.9 % were

SVROR crashes, while this percentage when SV crash did

not involve any unfamiliarity problem was 54.1%.



value= 0.002). So we can say that driver’s unfamiliarity

with the roadway is significantly associated with SVROR

crashes.

75.658.7

24.441.3

0

20

40

60

80

100

Driver Had Pre-existingPhysical or Mental Health

Conditions (Total:152,729)

Driver Did NOT Have Pre-existing Health Conditions

(Total: 360,240)

Single-Vehicle Crashes by Driver’s Pre-Existing Physical or Mental Health

Conditions

ROR Other

68.1 61.1

31.9 38.9

0

20

40

60

80

100

Male Drivers (Total:372,464)

Female Drivers (Total:289,872)

Single-Vehicle Crashes by Driver’s Gender

ROR Other

63.9 54.1

36.1 45.9

0

20

40

60

80

100

Driver Was Familiar Withthe Roadway (Total:

404,627)

Driver Was Not FamiliarWith the Roadway (Total:

108,220)

Single-Vehicle Crashes by Driver’s Familiarity with the Roadway

ROR Other

Pre

-Exis

tin

g P

hy

sica

l or

Men

tal

Hea

lth

Co

nd

itio

ns

Gen

der

F

amil

iari

ty w

ith

Ro

adw

ay

Page | 20

Work-related stress or pressure was reported in NMVCCS

if the driver had been in this state in the days leading up to

the crash.

As can be seen, among the SV crashes that work-related

stress or pressure was the factor, 86.4 % were SVROR

crashes, while this percentage when SV crash did not

involve any unfamiliarity problem was 59.5%.



value= 0.0973). So we can say that driver’s work-related

stress or pressure is significantly associated with SVROR

crashes.

As depicted, among the SV crashes in which the road

surface was dry, 70.6% were SVROR crashes, while this

percentage when SV crash occurred on a wet roadway

surface with water or ice or snow was 47.3%.



58.19, p-value < 0.0001. So we can say SVROR crashes

are more likely to occur on dry roadway surfaces than wet

roadway surfaces.

One possible reason for this can be that while driving on

wet roadways drivers usually exercise more cautions.

NMVCCS data has recorded the information about being in

a hurry as, late for start of work shift, late for start of school

classes, late for business appointment, work related

delivery schedule, late for social appointment,

pursuing/fleeing, and normal driving pattern.

As can be seen, among the SV crashes that being in a hurry

was the factor, 82.9 % were SVROR crashes, while this

percentage when being in a hurry was not the problem was

59.9%.


significant at the 90 percent confidence level (𝜒2= 16.69,

p-value= 0.0015). So we can say hurrying while driving is


86.459.5

13.640.5

0

20

40

60

80

100

Driver Was Feeling SomeWork-Related Stress orPressure (Total: 36,887)

Driver Was Not FeelingAny Work-Related Stress

or Pressure (Total:470,902)

Single-Vehicle Crashes by Driver’s Work-Related Stress or Pressure

ROR Other

70.647.3

29.452.7

0

20

40

60

80

100

The Roadway SurfaceWas Dry (Total: 493,547)

The Roadway SurfaceWas Wet With Water or

Ice or Snow (Total:178,365)

Single-Vehicle Crashes by Roadway Surface Conditions

ROR Other

82.959.9

17.140.1

0

20

40

60

80

100

Driver Was In a Hurry(Total: 39,693)

Driver Was Not In a Hurry(Total: 477,581)

Single-Vehicle Crashes Based on Whether the Driver Was in a Hurry

ROR Other

Wo

rk-R

elat

ed S

tres

s or

Pre

ssu

re

Road

way

Surf

ace

Condit

ions

In a

Hurr

y

Page | 21

To assess the relative influence of each of the above contributors (Liu and Ye, 2011) performed a logistic

regression which predicts the probability of occurrence of an event as a consequence of certain factors. The

end results were the odds-ratio which can tell us about the risk a certain factor carry in contributing to the

occurrence of ROR crashes and a coefficient corresponding to that odds-ratio which tell us about the

importance of each factor in the occurrence of ROR crashes (See Table 12).

Table 12. Logistic Regression Coefficients and Odds Ratios

Variable Coefficient Odds

Ratio p-value

Driver Inattention 1.2967 3.66 <.0001

Driver Was Fatigued 1.2463 3.48 <.0001

Driver Was In a Hurry 1.1630 3.2 <.0001

The Roadway Surface Was Dry 0.9928 2.7 <.0001

Driver Alcohol Present 0.9215 2.51 0.0218

Driver Was Familiar with the Roadway 0.7265 2.07 0.0032

Driver Had Pre-Existing Physical/Mental Health Conditions 0.5924 1.81 0.0031

Driver Was Male 0.2787 1.32 0.0217

Driver Was Feeling Work-Related Stress or Pressure 0.2252 1.25 0.5457


As shown in the table, the three most influential factors to the occurrence of SVROR crashes are “driver

inattention”, “fatigue” and “being in a hurry”. The odds of being involved in a SVROR crash for an

inattentive driver is 3.66 times higher than an attentive driver and it is statistically highly significant. For

driver feeling work-related stress, although the odds of being involved in a SVROR crash is 1.25 but it is

not statistically significant in 90 percent confidence interval.

Other studies by (Lord et al., 2011), (Liu and Subramanian, 2009), (Neuman et al., 2003), (Davis et al.,

2006), (Najm et al., 2002), (Liu and Jianqiang, 2011), (Wood et al., 2006), (Shanmugaratnam, 2008), (Ball

et al., 2009), (Romoser and Fisher, 2009) and (Hallmark et al., 2009) provide more specific details on the

roadway/environment, driver and other factors related to ROR crashes.

Conventional Countermeasures for reducing ROR Crashes In order to significantly reduce the number of crashes, American Association of State Highway and

Transportation Officials Strategic Highway Safety Plan (AASHTO-SHSP) has identified 22 goals. Goal 15

is Keeping Vehicles on the Roadway, and Goal 16 is Minimizing the Consequences of Leaving the Road.

The common solution to both of these goals is keeping the vehicle in the proper lane. Although this would

not eliminate the collisions between vehicles and pedestrian, bicyclists and trains, it would reduce the

fatalities and injuries caused by ROR crashes. Solutions proposed by AASHTO is to move away from

independent activities toward more coordinated efforts which can reduce the fatality and injuries rates in a

national scale.

FHWA Roadway Departure Team which is also responsible to provide important information for

transportation practitioners, decision makers, and others to assist them in preventing and reducing the

severity of roadway departure (RwD) crashes or ROR crashes have published Roadway Departure Strategic

Plan that is an approach toward “Zero ROR fatalities” and “serious injuries”.

Page | 22

In this study, a more comprehensive review of the methods of both agencies together with independent

studies by some researchers to reduce fatalities and serious injuries occurred as a result of ROR crashes are

presented.

FHWA RwD Team have published some recommended strategies and actions to reduce the number of ROR

crashes. To start with some, FHWA RwD Team strategies to mitigate most common roadway departure

fatal and serious injury crashes are shown in Figure 19.

As a response to AASHTO-SHSP, National Cooperative Highway Research Program (NCHRP) has also

published a report named Report 500, Volume 6: “Guide for Addressing Run-Off-Road Collisions”. This

report contains a comprehensive approach to minimize ROR crashes and its fatalities and injuries. The main

objective of this report is firstly to develop strategies to keep the vehicle in the travel lane and prevent them

from encroaching to the outside edges of the lanes. Factors that can cause driver errors include avoiding a

vehicle, object, or animal in the travel lane; inattentive driving due to distraction, fatigue, sleep, or drugs;

the effects of weather on pavement conditions; and traveling too fast through a curve or down a grade;

travel lanes that are too narrow, substandard curves, and unforgiving shoulders and roadsides. Strategies

are developed to deal with these problems.

The second objective is that if for any reason the vehicle travels on to the roadside, strategies should be

developed to minimize the likelihood of crashing or overturning. The probability of a crash occurring

depends on many factors like presence and location of fixed objects, shoulder edge drop-off, sideslopes,

ditches, trees and soil properties. Strategies directed at reducing the number and density of possibly

hazardous roadside features that would contribute to the likelihood of an ROR crash are provided.

The third objective is that given that the crash occurs, strategies needs to be developed to reduce the severity

of these crashes. These strategies may include making roadside hardware more forgiving or modifying

sideslopes to prevent rollovers) and by changes in the vehicle (e.g., better restraint systems or improved

side protection) or by increased occupant use of available restraints. So the three objectives of this NCHRP

Report are as follows:

• Keep vehicles from encroaching on the roadside

• Minimize the likelihood of crashing or overturning if the vehicle travels off the shoulder

• Reduce the severity of the crash

According to NCHRP 500 volume 6, strategies developed have been classified into three types which

include:

• Tried (T): These are strategies that have been implemented in a number of locations and may even

be accepted as standards, but for which there have not been found valid evaluations. While applying

these strategies, caution needs to be taken.

• Experimental (E): Strategies that have been suggested and may have been tried by some agencies

in one or two locations. These strategies need to be applied after careful testing and evaluations.

• Proven (P): These are strategies that properly designed evaluations have been conducted on them

and it is now proven that they are effective. They can be applied with great confidence.

In this paper the strategies are also divided into three parts based on the objectives of the NCHRP Report

and various other sources are used to include strategies that are not mentioned in this NCHRP Report.

Page | 23

83%

68% 68%

32%

UndividedRoads

Rural Areas Posted Speed ≥ 50 mph

Curves

Risk Factors for Opposite Direction Crashes

68%

52%48% 46%

Rural Areas Posted Speed ≥ 50 mph

Posted Speed ≤ 45 mph

Curves

Risk Factors for Tree and Shrub Crashes

76%72%

43%

Rural Areas Posted Speed > 50mph

Curves

Risk Factors for Overturn Crashes

Opposite Direction Crash Strategies:

1. Centerline rumble strips

2. Friction treatment in curves

3. Increased separation between opposing

lanes, particularly in curves.

4. Median Barriers

Roadside Trees and Shrub Crash Strategies:

1. Edge line and shoulder rumble strips

2. Curve delineation

3. Friction treatments in curves

4. Clear zone improvements, particularly

on the outside of the curves

5. Barriers to shield trees of curves

Overturn Crash Strategies:

1. Curve delineation

2. Friction treatments in curves

3. Edge line and shoulder rumble strips

4. Safety Edge

5. Clear zones

6. Traversable roadside slopes

7. Barriers to shield fixed objects and

slopes

Only 4% of head-on crashes involve a

passing vehicle 48% of tree-related fatalities occur where the

posted speed limit is 45 mph or less

43 % of overturn fatalities occur in curves

Figure 19. FHWA RwD Team strategies to

mitigate most common roadway departure

fatal and serious injury crashes

Page | 24

Strategies to Keep Vehicles from Encroaching on the Roadside

Shoulder Rumble Strips:

Shoulder Rumble Strips are generally about 0.5 inches deep, spaced about 7 inches apart, and cut in groups

of four or five. They can provide a sudden rumbling sound and can cause the vehicle to vibrate, and can

alert an inattentive, drowsy, or sleeping driver of encroachment on the shoulder and possibly onto the

roadside. They are generally less expensive and can be applied anytime during the construction or

maintenance of the project. Details regarding current practice with rumble strips can be found on the Federal

Highway Administration’s “Rumble Strip Community of Practice” Website provided in the Reference

section.

On freeways shoulder rumble strips are proven to be very effective to warn drivers of encroaching to the

roadside. According to FHWA, several studies have estimated that rumble strips can reduce the rate of ROR

crashes by 20 to 50 percent. Although they have been applied to many nonfreeways their effectiveness is

not documented and further evaluation is needed.

New York State Thruway Authority (NYSTA) installed continuous milled-in shoulder rumble strips on all

four shoulders of 485 roadway miles of thruway between 1992 and 1993. Before and after studies were

performed to measure the safety benefits of shoulder rumble strips for this particular corridor. (See Figure

20) Single-vehicle ROR crashes with certain “causes” were selected for the study.

As can be seen, applying shoulder rumble strips have decreased the number of fatalities, injuries and PDO’s

resulted from SVROR crashes to a considerable amount. Other studies by (Patel et al., 2007) and (Lord et

al., 2011) can be used for further information.

557

358

17

182

7454

119

0

100

200

300

400

500

Selected SVRORCrashes

Total Injuries Total Fatalities Property Damage Only

Figure 20. Before and After Data for Selected Single-Vehicle ROR Crashes on the New York Thruway (Source: New York State Police)

1991 Crashes (before installation; 6674 million vehicle miles)

1997 Crashes (Installation Complete; 8692 million vehicle miles)

Page | 25

Edgeline Rumble Strips for Roads with Narrow or Unpaved Shoulders:

Edgeline rumble strips are used when shoulders are narrow and unpaved. Although used by some agencies

in some locations, this strategy is still in its experimental phase requiring many tests and evaluations before

using.

Midlane Rumble Strips:

This treatment is also in its experimental phase and should be pilot tested and evaluated before widespread

use. They are similar to shoulder rumble strips 0.5 inch deep, spaced about 4 inches apart, and cut in groups

of four or five. There is fear among some designers and safety engineers that the strip in the center of the

lane may cause distraction to the driver. It is better if they are considered at locations with both an ROR

and a head-on crash problem.

Enhanced Delineation of Sharp Curves:

Sharper curves result in more shoulder encroachments and crashes and flattening of the curve may be too

costly. One way to consider both safety and budgetary constraints is to delineate sharp curves with enhanced

markings. This can be done by improved shoulder delineation (adding more chevrons or high-intensity

chevrons, large arrow signs, or delineators on guardrails); improved curve warning signs (warning signs

with flashing beacons); or innovative on pavement markings (warning arrows on the pavement prior to the

curve).

According to many studies, well-placed shoulder delineators are proven to reduce ROR crashes. On the

other hand on-pavement treatments have been evaluated in terms of speed reduction but not crash reduction.

This is a categorized as a tried strategy. In a very well-designed early study of post-mounted delineators on

rural two-lane curves (Foody and Taylor, 1966) found that they reduce ROR crashes by 15 percent. Many

other studies also have documented the benefits of enhanced delineation of sharp curves.

Improved Highway Geometry for Horizontal Curves:

According to (Glennon et al., 1985), both ROR and head-on crashes are 1.5 to 4 times more likely to occur

on curves than on tangents. In a study done by (Zegeer et al. 1992) it was found that flattening curves on

two-lane rural roadways from 30 degrees to 5 degree would result in total curve crash reductions of up to

80 percent. So this strategy is proven to reduce ROR crashes. This strategy is among the higher-cost

alternatives of those considered.

Enhanced Pavement Markings at Appropriate Locations

Pavement marking should be enhanced at locations where they drivers might leave the roadway. This may

be done through higher contrast or wider markings or raised pavement markers (RPMs). The main purpose

behind this strategy is to make the drivers use the information to stay in their lanes and not merely to

maintain or increase their speed. They are considered to be in the “tried” category of strategies. This is

because some studies have raised questions about the effectiveness of this measure.

In a study by (Pendleton, 1996) and a research conducted by Bellomo-McGee, Inc., for NCHRP indicate a

lack of significant effect or even a possible increase in crashes on some locations. This may be because

drivers tend to drive faster when presented with a clearer delineation.

But in another study of six rural two-lane roads (over 126 miles) with RPMs in northern New Jersey, it was

found that there was a statistically significant reduction in various nighttime crashes including total, injury,

property damage, overturn, head-on, fixed object, and between intersection crashes. (State of New Jersey,

1986).

Page | 26

Skid-Resistant Pavements:

A vehicle will skid during braking and maneuvering when frictional demand exceeds the friction force that

can be developed at the tire-road interface. The probability of this occurring on a wet pavement is much

higher than a dry pavement. According to a study by (Hall et al., 2009), even as little as 0.002 inches of

water on the pavement can reduce the coefficient of friction by 20 to 30 percent.

There has been a large amount of research funded by the FHWA, AASHTO, and pavement associations

concerning designing better pavements—durable and cost-effective pavements. An important parameter in

all this work is to design skid-resistant pavement with good drainage capability.

Although there is only limited research on such site-specific programs, the results of research on the general

effectiveness of decreasing skidding has placed this strategy in proven category.

New York State Department of Transportation (NYDOT), implemented a statewide program that identified

sites with low pavement skid resistance and treated them with overlays or microsurfacing. Between 1995

and 1997, 36 sites were treated and resulting in a reduction of more than 800 wet crashes per year. This was

a reduction of 50 percent in wet-road accidents and 20 percent for total accidents. Reduction of ROR crashes

are also perceived to have decreased by the same amount proportion as total crashes.

Shoulder Treatments:

If a shoulder area can allow the vehicle to safely recover from a dangerous maneuver that caused the vehicle

to leave the lane and enter the shoulder, it can prevent a ROR crash. It can further increase the safety if the

shoulder area is paved and wide enough to fit a vehicle. Shoulder treatments that promote safe recovery

include shoulder widening, shoulder paving, and the reduction of pavement edgedrops.

Although widening or paving a shoulder is a self-evident alternative, treating an edgedrop in locations

where it is very obvious can increase safety. Edgedrops may result from repaving, where material is added

to the lane but not to the adjacent shoulder, or from weather or vehicle-caused “erosion” of unpaved

shoulders. According to (Humphreys and Parham, 1994), the best treatment to an edgedrop is to make it

level with the pavement surface. Since this may be costly and sometimes difficult, they came up with an

excellent alternative for both paved and unpaved shoulders. They proposed adding a 45-degree fillet at the

lane/pavement edge that would allow the vehicle to safely return to the roadway. Based on the current

researches, shoulder paving and widening are considered proven strategies, while edgedrop treatments are

considered experimental.

In FHWA’s effort related to determining accident modification factors for use with the Interactive Highway

Safety Design Model (IHSDM), (Harwood et al., 2000), studied the effects of shoulder widening on

reducing the number of ROR crashes. As can be seen from Figure 21 for an ADT of less than 400 veh/day,

the effect of shoulder widening is minimal. But as the ADT increases, the benefits of widening the shoulder

also increases and will stay constant after an ADT of more than 2000 veh/day.

Same source has documented the benefits of using paved or unpaved shoulder. Figure 22 shows the benefits

of using paved shoulder. As can be seen, AMF is lower in paved shoulder than gravel, composite or turf

shoulders. Much less is known about the effectiveness of edgedrop treatments, as it is difficult to identify

the crashes that were caused as a result of overcorrection. (Humphreys and Parham, 1994) concluded that

a 45-degree-angle asphalt fillet at the lane edge would eliminate this type of crash, even in cases where the

shoulder is unpaved and suffers from erosion damage.

Page | 27

0.98

1

1.02

1.04

1.06

1.08

1.1

1.12

1.14

1.16

0 2 4 6 8 10

AM

F

Shoulder Width (ft)

Figure 22. Accident Modification Factor for Shoulder Type on Two-Lane Rural Highways (Source: Harwood et al., 2000)

Paved Gravel Composite Turf

Figure 21. Accident Modification Factor for Paved Shoulder Width (Relative to 6-Foot

Paved Shoulder) on Two-Lane Rural Highways (Source: Harwood et al., 2000)

Page | 28

Minimize the Likelihood of Crashing into an Object or Overturning if the Vehicle

Travels Off the Shoulder

Design Safer Slopes and Ditches to Prevent Rollovers:

As shown in the first part of the report, the most harmful event in a SVROR is overturn or rollover.

Decreasing the number of rollover crashes can contribute significantly to the safety of a roadway. The

roadside design features most likely to affect rollover include the sideslopes (particularly fill slopes), ditch

design, the nature of the soil on the slope, and the design of roadside hardware that might lead to rollovers

(e.g., poorly designed guardrail ends).

Flattening the sideslopes is a proven strategy to decrease the number of ROR crashes. According to

FHWA’s Model Inventory of Roadway Elements (MIRE, 2010) roadside rating the best roadside with a

rating of 1 has sideslopes flatter than 1:4.

Additional rollover (and other ROR) crash reduction could come from improved designs of roadside

ditches. Although AASHTO Roadside Design Guide includes preferred foreslopes and backslopes for basic

ditch configurations, they are only applied to interstates or major highways and are rarely applied to other

low profile roadways.

(Zegeer et al., 1987) examined the effects of sideslope on both rollover crashes and total single-vehicle

crashes. They used field-measured crash, sideslope, cross section, and traffic data from approximately 1,800

miles on rural two-lane roads in three states. They found that rollover rates were significantly higher on

slopes of 1:4 or steeper as compared with slopes of 1:5 or flatter.

So based on the current research evidence, the set of strategies related to rollover (and ROR crash) reduction

through changes in the “vertical” component of the sideslopes and changing the foreslopes and backslopes

of ditches would be considered “proven” strategies.

Remove/Relocate Objects in Hazardous Locations:

Although a considerable amount of rollover crashes could be reduced by improving sideslopes and ditches

which are part of the clear zone- the recovery area provided to vehicles that leave the roadway, additional

fatalities and injuries can be reduced by modifying the width of the clear zone. The wider the clear zone,

the more likely that a vehicle can recover from a dangerous maneuver. Usually the width of a clear zone is

established by natural objects (trees, rock outcrops, embankments, etc.), or by roadside hardware (guardrails

etc.).

According to the Roadside Design Guide, a “safe clear zone width” on higher-speed roads is approximately

30 feet but there is no single width that defines maximum safety. Width of the clear zone usually depends

on design speed of the roadway, design ADT, the prevailing sideslope, and curvature.

A study for the Texas DOT by (Fambro et al., 1995) used a benefit-cost analysis to establish guidelines for

clear zone on suburban high-speed roadways with curb and gutter. The baseline minimum clear zone width

used in the calculation was approximately 10 feet. It was found that it is not cost-beneficial to purchase 5

feet or less of additional right-of-way since the cost of relocating the objects would be higher than cost of

purchasing the land. It was also found that if a roadway has less roadside hazards, it is not cost beneficial

to purchase additional clear zone.

(Zegeer et al., 1987) also estimated the effects of clear zone widening on two-lane rural roads. If the existing

recovery area measured from the edgeline is less than 10 to 15 feet, Figure 23 shows the percent reduction

in ROR, head-on and sideswipe crashes due to widening the clear zone.

Page | 29

As indicated, and increase of 20 feet has three times more benefits in reducing related crashes than an

increase of 5 feet.

So based on the current research concerning crash-related effectiveness, strategies widening the clear zone

are considered to have proven effectiveness.

Delineation of Roadside Objects:

The main purpose of delineating roadside objects is that they become more visible to drivers at night.

Pennsylvania Department of Transportation (PennDOT) has an experimental program targeting delineation

to potentially hazardous objects on road segments with high run-off-road utility pole- and tree-related crash

frequencies. This program is proving effective when it is not feasible to remove or relocate an object

because of budget constraints or the object is on private property. PennDOT marks the tree or utility pole

with a round of reflective tape. One round is used on each tree and each utility pole. Two rounds are used

for poles at intersections.

Since no evaluation has yet been documented on the effectiveness of this practice, this strategy is considered

as experimental.

Reduce the Severity of the Crash

Improve Design of Roadside Hardware or Application of Barrier and Attenuation Systems:

Although the concept of forgiving roadside requires us to place all the roadside fixed objects outside the

clear zone, so objects need to be placed near the traveled way. These objects include hardware or objects

related to traffic guidance or control (signs, some lighting supports, etc.); protection of more hazardous

objects or situations (guardrails or median barriers); roadway design requirements (culverts); and traditional

right-of-way uses (utility poles, mail boxes).

0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

50%

5 10 15 20Pe

rce

nt

Re

du

ctio

n in

Re

late

d A

ccid

en

t Ty

pe

s (i

.e.,

R

OR

+he

ad-o

n+s

ide

swip

e)

Amount of Increased Roadside Recovery Distance (feet)

Figure 23. Percent Reduction in Related Accident Types (i.e., ROR+head-on+sideswipe)

Page | 30

If relocating the objects farther from traffic flow or hazardous locations is impossible, shielding or replacing

“harder” objects with less hazardous breakaway devices (use of breakaway luminaire supports, or use of

crash cushions in front of hazardous immovable objects) can be a good option.

Based upon current crash-related research, the relocation of “hard” objects farther from the roadway or their

replacement with more forgiving designs (e.g., breakaway designs) are considered “proven” strategies to

reduce roadside harm.

Contemporary Countermeasures for reducing ROR Crashes During the recent two decades, several intelligent vehicle technologies have emerged that can provide

instantaneous on the road information to the motorist and can result in timely decisions by the driver.

NHTSA has recognized that technologies such as electronic stability control (ESC) and Anti-lock Braking

System (ABS) can potentially reduce a great number of these fatal crashes. Emerging technologies like lane

departure warning systems (LDWS) can also have great impacts in crash avoidance.

A detailed description of the functions and guidelines for using these technologies is beyond the scope of

this paper. Instead, this paper looks into the major benefits of using these systems in terms of preventing

fatal and major injury crashes.

In the first part of this section, an introduction and the potential benefits of using ESC and ABS in reducing

ROR crashes is discussed. The second section includes information about how LDWS systems function

and what are their benefits in reducing ROR crashes. The reader is persuaded to refer to the references

provided in the report for more information.

Effects of ABS and ESC on ROR Crashes ABS is a four-wheel system that modulates the brake pressure during an emergency stop, hence preventing

wheel lock-up. Preventing the wheel from an entire lock-up, ABS helps the driver to maintain steering

control and to stop in the shortest possible distance. On the other hand ESC which is an evolution of ABS

system, uses a technology that maintains vehicle control that is lost due to oversteering or skidding.

According to many studies by (Hertz and Johnson, 1998) and (Kahane and Dang, 2009) ABS-supplied

vehicles have zero or even negative impacts in reducing the number of fatal ROR crashes. But according

to (National Automotive Sampling System-General Estimates System (NASS-GES), 1995-2007) ABS has

been quite effective in reducing non-fatal crashes. It has reduced the overall crash involvement of passenger

cars by 6% and light trucks by 8%.

In order to show the effects of ESC and ABS system, the NMVCCS data which also contains information

about the availability and use of equipment on-board the vehicles, both original equipment (OEM) and

after-market. Table 13 shows the benefits of using ABS and ESC in reducing ROR crashes. Three cases are

considered.

ABS Versus “Neither ABS nor ESC”

“ABS Only” Versus “Both ABS and ESC”

“Both ABS and ESC” Versus “Neither ABS Nor ESC”

Page | 31

Table 13. Benefits of using ABS and ESC in reducing ROR crashes

As can be seen, of the vehicle equipped with ABS

only, 13.2% ran off the road and of the vehicle

equipped neither with ABS nor with ESC, did 14.6%

run off the road.

The odds of being involved in a ROR crash if the car

is equipped with neither ABS nor ESC is 1.1 times

the odds for the vehicle equipped with ABS only.

According to (Liu and Ye, 2011) the positive effect

of ABS on reducing the ROR crashes is not

statistically significant at the 90 percent confidence

level (p-value=0.500).

As can be seen, of the vehicle equipped with ABS

only, 13.2% ran off the road and of the vehicle

equipped with both ABS nor with ESC, did 7.5% run

off the road.


is equipped with ABS only is 1.9 times the odds for

the vehicle equipped with both ABS and ESC.

According to (Liu and Ye, 2011) the positive effect

of both ABS and ESC on reducing the ROR crashes

is not statistically significant at the 90 percent

confidence level (p-value=0.154).

As can be seen, of the vehicle equipped with neither

ABS nor ESC, 14.6% ran off the road and of the

vehicle equipped with both ABS nor with ESC, did

7.5% run off the road.


is equipped with neither ABS nor ESC is 2.1 times

the odds for the vehicle equipped with both ABS and

ESC. According to (Liu and Ye, 2011) the positive

effect of both ABS and ESC on reducing the ROR

crashes is statistically significant at the 90 percent

confidence level (p-value= 0.065).

In another study by (Kahane and Dang, 2009) it was

confirmed that a combination of ABS and ESC

would reduce fatal ROR crashes by an estimated 30

percent in passenger cars and by 68 percent in light

trucks.

13.214.6

0

5

10

15

20

ABS only (Total:2,153,596)

Neither ABS Nor ESC(Total: 1,090,053)

Pe

rce

nt

Percent of ROR Vehicles

7.5

13.2

0

5

10

15

20

ABS & ESC (Total:155,077)

ABS Only (Total:2,153,596)

Pe

rce

nt


7.5

14.6

0

5

10

15

20

ABS & ESC (Total:155,077)

Neither ABS Nor ESC(Total: 1,090,053)

Pe

rce

nt


AB

S V

ersu

s “N

eit

her

AB

S n

or

ES

C”

“AB

S O

nly

” v

ersu

s “B

oth

AB

S a

nd

ES

C”

“Bo

th A

BS

an

d E

SC

“v

ersu

s “N

eith

er A

BS

no

r E

SC

”

Page | 32

Effects of LDWS on ROR Crashes One emerging technology that NHTSA believes can have a great impact on decreasing fatalities and serious

injuries caused by ROR crashes is lane departure warning systems (LDWS). This technology was a key

technology introduced by NHTSA at the start of Intelligent Vehicle Highway System (IVHS) in 1992

(Barickman et al., 2007).

The LDWS uses sensors to determine the vehicle’s position and velocity relative to the road. A collision

warning algorithm interprets this state to determine if the vehicle is in danger of unintentionally drifting out

of the travel lane. If so, the system provides a warning to the driver. The block diagram in Figure 24 shows

the functional blocks within a LDWS system. Sensors senses the relative position and velocity of the vehicle

relative to the roadway and if the vehicle is trying to move out of the roadway, a collision warning algorithm

sends an sound alert to the driver interface.

LDWS can be effective technology in preventing ROR crashes as it can prevent the vehicle from departing

the lane by either warning the driver or actively controlling the vehicle. This ESC system can prevent the

driver from oversteering thus preventing it from spinning out of control or reduce understeering, thereby

preventing a vehicle from running off the

road in a sharp curve.

Different projects are currently on going in

order to improve this time of process and to

deal with more and more case. Citroen

which is a major French automobile

manufacturer, has an interesting system.

This system activates when driving speed is

above 80 km/h (50 mph) and lanes are

changed without using the turn signals. A

vibrator, mounted under the driver seat,

alerts the driver of the possible crash. This

is done with six infrared sensors, three on

each side, that are mounted on the front

bumper that monitors when a line

unintentional crossed, and then sending a

signal to one of the two vibrators. (See

Figure 25 for more details).

LDWS was introduced as an original

equipment (OE) with the 2007 models of

vehicles made in Japan, Europe, and North

America. A detailed description of the

functions and guidelines for using LDWS is beyond the

Figure 24. Lane Drift Warning System Functional Blocks

Driver

Warning

Display /

Interface

Collision

Warning

Algorithm

Road and

Host Vehicle

State Sensing

Figure 25. Citroen Designed LDWS http://www.citroen.co.uk/about-

citroen/technology/lane-departure-

warning-system

Page | 33

scope of this paper. Instead, this paper looks into the major benefits of using LDWS in terms of preventing

fatal and major injury crashes.

In a report by (Pomerleau et al., 1999) the effects of LDWS on ROR crashes were studied. The results of

the study indicated that LDWS have the potential to reduce road departure crashes in passenger vehicles by

approximately 10 percent, and reduce road departure crashes in heavy trucks by approximately 30 percent.

These reductions would result in approximately 160,000 fewer crashes and 1,500 fewer fatalities in

passenger vehicles each year. In heavy trucks, a 30 percent reduction in run-off-road crashes would result

in approximately 9,300 fewer crashes and 96 fewer fatalities each year.

Conclusion / Discussion The main purpose of this paper was to discuss the following two issues:

Critical reasons and contributing factors for ROR crashes

Countermeasures to reduce the number of ROR crashes

Run-off-road crashes account for a significant percentage (around 70%) of all fatal single-vehicle crashes.

FARS data which includes detailed information about ROR crashes provides sufficient evidence to

conclude that certain driver-, environment-, and vehicle-related factors closely associated with the

occurrence of these crashes. Finding countermeasures based on the critical reasons and associated factors

can help reduce the number of ROR crashes considerably.

Among all critical reasons for passenger vehicle SVROR crashes, more than 95 percent were driver-related.

The dominant critical reasons were “internal distraction,” “overcompensation,” “poor directional control,”

“too fast for curve,” and “sleeping/actually asleep.” For trucks, sleeping and hearth attack problems were

more prominent. Driver-related errors are more prevalent in both trucks and passenger vehicles. Finding

countermeasures to deal with driver problems can help increase safety considerably. Although allocating

the budget on building new highways can help to some extents, driver-related countermeasures can help a

lot.

The paper also described the most important contributing factors to ROR crashes. The dominant factors

were “driver Inattention,” “driver fatigue,” and “driver in a hurry”. Focusing the attention to deal with these

problems (e.g. shoulder rumble strips, LDWS etc.) can help reduce roadway departure crashes to a good

extent.

Conventional Countermeasures described in this report like shoulder rumble strips, edgeline delineating,

widening clearzones, and treating shoulders have shown themselves in reducing the fatalities caused by

ROR crashes. Although some of these treatments are in their experimental stage, applying it to some pilot

locations and testing the results can be a good starting step in accepting or denying some of these

countermeasures.

This report also summarized some of the state-of-the-art technologies currently used to intelligently solve

ROR crash problems. In the evaluation of ABS and ESC in reducing ROR crashes, it was found that the

combined effect of ABS and ESC systems on reducing the ROR crashes is significant. But if was use only

one of them, it cannot improve safety significantly. LDWS is also studied in this report. It is documented

that LDWS reduce road departure crashes in passenger vehicles by approximately 10 percent, and reduce

road departure crashes in heavy trucks by approximately 30 percent. If this technology is accepted in the

car manufacturing industry, it can help reduce ROR crashes to a considerable amount.

Page | 34

Areas of Future Research As discussed in this document, many factors associated with roadway departure crashes are human-related

(e.g., Liu and Ye, 2011). In many cases, a driver commits an error, and the end result that governs the type

and severity of the crash is related to where the error is committed, whether it is on a curve or a tangent

section, or at a location with a high pavement edge drop-off. With this in mind, future research can examine

countermeasures that would help reduce the likelihood for these errors to occur and those that would

minimize the severity when the driver leaves the traveled-way. This may include conventional or new

advanced technologies.

Fatigue is also one of the largest sources to ROR crashes. Further development of drowsiness detection

systems can have significant impact in reducing fatalities and serious injuries occurred by ROR crashes due

to fatigue. The potential benefit of applying these systems in trucks and buses is unimaginable.

References AASHTO (2002). Roadside Design Guide. American Association of State Highway and Transportation

Officials, Washington, DC.

AASHTO-SHSP, http://safety.transportation.org/

Ball, K., Owsley, C., Sloane, E., Roenker, D., and Bruni, J. “Visual attention problems as a predictor of

vehicle crashes in older drivers,” Investigative Ophthalmology & Visual Science, 34, 3110-3123.

Citroen Designed LDWS. http://www.citroen.co.uk/about-citroen/technology/lane-departure-warning-

system

Critical Reasons for Crashes Investigated in the National Motor Vehicle Crash Causation Survey 2015.

Report No. DOT HS 812 115. National Highway Transportation Safety Agency, Washington, D.C.

Davis, G.A., S. Davuluri, and J. Pei (2006). Speed as a Risk Factor in Serious Run-off-Road Crashes:

Bayesian Case-Control Analysis with Case Speed Uncertainty. Journal of Transportation and Statistics,

Vol. 9, No. 1. Research and Innovative Technology Administration, U.S. Department of Transportation,

Washington, DC. P. 17-28.

F Barickman, R Jones, L Smith, Lane departure warning system research and test development, in

Proceedings of the 20th International Conference on the Enhanced Safety of Vehicles, (Lyon, 18–21 Jun

2007), pp. 1–8

Fambro, D. B., R. L. Nowlin, S. P. Warren, K. A. Lienau, J. M. Mounce, R. P. Bligh, K. K. Mak, and H.

E. Ross. Geometric Design Guidelines for Suburban High-Speed Curb and Gutter Roadways. Texas

Transportation Institute, Texas A&M University, College Station, Texas. 1995. -008. 1987.

FHWA (2006). Safety Compass: Highway Safety Solutions for Saving Lives. A Publication of the

Federal Highway Administration Safety Program, December 2006 - Volume 1, Issue 1.Accessed from:

http://safety.fhwa.dot.gov/newsletter/safetycompass/2006/dec06_v1/ .Accessed on Aug 23, 2011.

FHWA Roadway Departure (RwD) Strategic Plan, March 2013

FHWA, Office of Safety, NY State Thruway Selected Crash Experience,

http://safety.fhwa.dot.gov/roadway_dept/pavement/rumble_strips/effectiveness/state_ny/ny_crash.cfm

http://safety.transportation.org/

http://www.citroen.co.uk/about-citroen/technology/lane-departure-warning-system

http://www.citroen.co.uk/about-citroen/technology/lane-departure-warning-system

http://safety.fhwa.dot.gov/newsletter/safetycompass/2006/dec06_v1/

http://safety.fhwa.dot.gov/roadway_dept/pavement/rumble_strips/effectiveness/state_ny/ny_crash.cfm

Page | 35

FMCSA. (2006, March). Report to Congress on the Large Truck Crash Causation Study. Washington,

DC: Federal Motor Carrier Safety Administration. March 2006.

Foody, T. J., and W. C. Taylor. Curve Delineation and Accidents. Ohio Department of Highways, Bureau

of Traffic. Columbus, Ohio. 1966.

Glennon, J. C., T. R. Newman, and J. E. Leisch. Safety and Operational Considerations for Design of

Rural Highway Curves, FHWA/RD-86/035. FHWA, U.S. Department of Transportation. 1985.

Hall, J.W., K.L. Smith, L. Titus-Glover, J.C. Wambold, T.J. Yager, and Z. Rado, Guide for Pavement

Friction, NCHRP Web-Only Document 108, Transportation Research Board of the National Academies,

Washington D.C., 2009.

Hallmark, S.L., Hsu, Y.Y., Maze, T., McDonald, T., and Fitzsimmons, E. (2009). Investigating factors

contributing to large truck lane departure crashes using the federal motor carrier safety administration’s

large truck crash causation study (LTCCS) database. Center for Transportation Research and Education,

Iowa State University, Ames, IA.

Harwood, D. W., F. M. Council, E. Hauer, W. E. Hughes, and A. Vogt. Prediction of the Expected Safety

Performance of Rural Two-Lane Highways, FHWA-RD-99-207, Federal Highway Administration. 2000.

Hertz, E., Hilton, J. and Johnson, D.M. (1998, August). Analysis of the Crash Experiences of Vehicles

Equipped with Antilock Braking Systems: An Update. DOT HS 808 758. Washington, DC: National

Highway Traffic Safety Administration.

Humphreys, J. B., and J. A. Parham. The Elimination or Mitigation of Hazards Associated with Pavement

Edge Drop-Offs During Roadway Resurfacing. AAA Foundation for Traffic Safety, Washington, D.C.

1994.

Kahane, C. J., & Dang, J. N. (2009, August). The Long-Term Effect of ABS in Passenger Cars and LTVs.

DOT HS 811 182. Washington, DC: National Highway Traffic Safety Administration.

Liu C. and T. G. Ye (2011). Run-Off-Road Crashes: An On-Scene Perspective. Report No. DOT HS 811

500. National Highway Transportation Safety Agency, Washington, D.C.

Liu, C., & Subramanian, R. (2009, November). Factors Related to Fatal Single-Vehicle Run-Off-Road

Crashes. DOT HS 811 232. Washington, DC: National Highway Traffic Safety Administration.

Lord, Dominique, Brewer, M.A., Fitzpatrick, Kay, Geedipally, S.R., Peng, Yichuan. Analysis of roadway

departure crashes on two-lane rural roads in Texas, Report, September 2011

McGinnis, R. G., M. J. Davis, and E. A. Hathaway, 2001. Longitudinal Analysis of Fatal Run-off-road

Crashes, 1975 to 1997, in Transportation Research Record of the National Academies: Journal of the

Transportation Research Board, vol. 1746, no. 1, pp.47-58.

Model Inventory of Roadway Elements (MIRE), Federal Highway Administration, U.S. Department of

Transportation. 2010

Najm, W., J. Koopmann, L. Boyle, and D. Smith. 2002. Development of Test Scenarios for Off-Roadway

Crash Countermeasures Based on Crash Statistics. DOT-HS-809-505. Washington, DC: National

Highway Traffic Safety Center.

Page | 36

National Highway Traffic Safety Administration (NHTSA), Traffic Safety Facts 2012: A Compilation of

Motor Vehicle Crash Data from the Fatality Analysis Reporting System and the General Estimates

System, Report HS 812 032. U.S. Department of Transportation, 2013.

National Highway Traffic Safety Administration (NHTSA), Traffic Safety Facts 1999: A Compilation of

Motor Vehicle Crash Data from the Fatality Analysis Reporting System and the General Estimates

System, Report HS 809 100. U.S. Department of Transportation, 2000.

Neuman, T. R., R. Pfefer, K. L. Slack, K. K. Hardy, F. Council, H. McGee, L. Prothe, and K. Eccles,

NCHRP Report 500: Guidance for Implementation of the AASHTO Strategic Highway Safety Plan.

Volume 6: A Guide for Addressing Run-Off-Road Collisions, Transportation Research Board,

Washington D.C., 2003.

Patel, R.B., F.M. Council, and M.S. Griffith (2007). Estimating Safety Benefits of Shoulder Rumble

Strips on Two-Lane Rural Highways in Minnesota: Empirical Bayes Observational Before-and-After

Study. TRB 86th Annual Meeting Compendium of Papers, Washington, DC.

Pendleton, O. Evaluation of Accident Analysis Methodology. Federal Highway Administration,

Washington, D.C. FHWA RD-96-039. 1996.

Pomerleau, Dean, et al. (1999 December). Run-Off-Road Collision Avoidance Using IVHS

Countermeasures. DOT HS 809 170. Washington, DC: National Highway Traffic Safety Administration.

Romoser, M. and Fisher, D. (2009). “Effects of cognitive and physical decline on older drivers’ side-to-

side scanning for hazards while executing turns,” Proceedings of the fifth international driving

symposium on human factors in driving assessment, training, and vehicle design, Big Sky, MT, 61-68.

Rumble Strip Community of Practice http://safety.fhwa.dot.gov/programs/rumble.htm.

Shanmugaratnam, S. (2008). Predicting simulated driving performance under high and low workload

environments using cognitive, visual, and psychomotor tasks. Unpublished master’s thesis, University of

West Florida, Pensacola, FL.

Spainhour, L. K., & Mishra, A. (2007). Analysis of Fatal Run-off-the-Road Crashes Involving

Overcorrection. TRB 87th Annual Meeting, Paper No. 08-0661. Washington, DC: Transportation

Research Board.

Starnes, M. (2006, August). Large-Truck Crash Causation Study – An Initial Overview. DOT HS 810

646. Washington, DC: National Highway Traffic Safety Administration.

State Highway Traffic Safety Office, Montana Traffic Safety Problem Identification, 2009 Data, Montana

Department of Transportation (MDT), 2011.

State of New Jersey. Annual Safety Report. Fiscal Year 1986.

WHO.2013. Facts Sheet 358. http://www.who.int/mediacentre/factsheets/fs358/en/

Wood, J., Chaparro, A., Hickson, L., Thyer, N., Carter, P., Hancock, J., and Hoe, A. (2006). The effect of

auditory and visual distracters on the useful field of view: Implications on the useful field of view:

Implications for the driving task. Investigative Ophthalmology & Visual Science, 47, 4646-4650.

Zegeer, C. V., J. Hummer, D. Reinfurt, L. Herf, W. Hunter. Safety Effects of Cross-Section Design for

Two-Lane Roads—Volumes I and II. Federal Highway Administration, Washington, D.C. FHWA-RD-87

http://safety.fhwa.dot.gov/programs/rumble.htm

http://www.who.int/mediacentre/factsheets/fs358/en/

Page | 37

Zegeer, C. V., J. R. Stewart, F. M. Council, D. W. Reinfurt, and E. Hamilton. “Safety Effects of

Geometric Improvements on Horizontal Curves,” Transportation Research Record 1356. Transportation

Research Board of the National Academies. 1992.

Run-Off-Road (ROR) Crashes: A Description of Conventional and Contemporary Countermeasures

Documents