Final Report: Advanced Technologies for Improving Large-Truck Safety on Two-Lane Secondary Roads

Final Report: Advanced Technologies for Improving Large-truck Safety on Two-Lane Secondary Roads

Click HERE for graphic.

FINAL REPORT

ADVANCED TECHNOLOGIES FOR IMPROVING LARGE-TRUCK
SAFETY ON TWO-LANE SECONDARY ROADS

Nicholas J. Garber, Ph.D.
Faculty Research Engineer
and
Professor of Civil Engineering

Kirsten A. Black, E.I.T.
Graduate Research Assistant

(The opinions, findings, and conclusions expressed in this
report are those of the authors and not necessarily
those of the sponsoring agencies.)

Virginia Transportation Research Council
(A Cooperative Organization Sponsored Jointly by the
Virginia Department of Transportation and
the University of Virginia)

In Cooperation with the U.S. Department of Transportation
Federal Highway Administration

Charlottesville, Virginia

March 1995
VTRC 95-R17

SAFETY RESEARCH ADVISORY COMMITTEE

D. F. MICHAEL, Chairman, Commissioner for Field Operations,
Department of Motor Vehicles
J. D. JERNIGAN, Executive Secretary, Senior Research Scientist,
VTRC
J. D. AUSTIN, Transportation Engineering Program Supervisor,
Department of Rail & Public Transportation
J. L. BLAND, Manager, Plans, Programs & Services, Department of
Aviation
R. J. BREITENBACH, Director, Transportation Safety Training Center,
Virginia Commonwealth University
J. L. Butler, Traffic Engineering Division Administrator, VDOT
Maj. J. K. COOKE, Assistant Chief of Law Enforcement,
Department of Game & Inland Fisheries
V. L. CROZIER, Associate Specialist, Driver Education, Department
of Education
W. S. FELTON, JR., Administrative Coordinator, Commonwealth's
Attorneys' Services & Training Council
P. D. FERRARA, Director, Division of Forensic Sciences, Department
of General Services
D. R. GEHR, Commissioner, VDOT
J. T. HANNA, Assistant Professor, Transportation Safety Training
Center, Virginia Commonwealth University
T. A. JENNINGS, Safety/Technology Transfer Coordinator, Federal
Highway Administration
Lt. Colonel W. G. MASSENGILL, Director, Bureau of Field
Operations, Virginia Department of State Police
W. T. McCOLLUM, Executive Director, Commission on VASAP
S. D. McHENRY, Director, Division of Emergency Medical Services,
Department of Health
Lt. S. E. NEWTON, Commander, Patrol Division, Albemarle County
Police Department
J. T. PHIPPS, Director, Roanoke Valley ASAP
J. A. SPENCER, Assistant Attorney General, Office of the Attorney
General
E. W. TIMMONS, Director of Public Affairs, Tidewater AAA of
Virginia
A. R. WOODROOE Esq., Assistant Attorney General (Retired)

ABSTRACT

The use of large trucks has steadily increased since the passage of
the Surface Transportation Assistance Act to the point where they now
account for over 50% of vehicle traffic on some highways in Virginia.
Projections now forecast that large-truck travel will grow at twice
the rate of personal vehicle travel in the near future.

Although several studies have been conducted to determine the
effects of large trucks on safety on multilane primary and interstate
highway systems, the effects on two-lane secondary roads have been
largely ignored. This study identified the causal factors and
predominant types of large-truck crashes on two-lane secondary roads
in Virginia and compared the large-truck crash rates for two-lane
secondary roads and two-lane primary roads. The study also identified
advanced technologies associated with intelligent transportation
systems (ITS) that can be used to minimize the causal factors of
large-truck crashes on these roads.

The results showed that large-truck crash rates are significantly
higher on two-lane secondary roads than on two-lane primary roads,
with the predominant types of crashes being angle, rear end, sideswipe
same direction, and sideswipe opposite direction. The study
identified several ITS technologies that can be used to mitigate the
predominant causal factors and recommends a pilot study to test the
effectiveness of one such system.

iii

FINAL REPORT

ADVANCED TECHNOLOGIES FOR IMPROVING LARGE-TRUCK
SAFETY ON TWO-LANE SECONDARY ROADS

Nicholas J. Garber, Ph.D.
Faculty Research Engineer
and
Professor of Civil Engineering

Kirsten A. Black, E.I.T
Graduate Research Assistant

INTRODUCTION

Supporters of the Surface Transportation Assistance Act (STAA) of
1982 and the Tandem Truck Safety Act (TTSA) of 1984 claimed that
passage of these two acts would reduce the overall vehicle miles of
travel (VMT) of large trucks (trucks having six or more wheels in
contact with the road and having a gross weight greater than 4535.9 kg
[10,000 lb]) since fewer of the longer and wider trucks would be
needed for the transportation of goods in the United States.
Supporters also believed that the increased use of twin-trailer trucks
(truck-tractors pulling two trailing units) would have little overall
effect on highway safety because the reduction in truck VMT would
approximately offset the small possible increase in crash involvement
per mile traveled.1

The predicted reduction in large-truck VMT, however, has not
occurred. Apparently, increasing the size of the trucks has simply
reduced the expense of distribution. Consequently, more businesses
have begun using this mode of transporting goods, thus increasing the
number of large trucks using the highway system and the annual VMT.
In fact, their use has steadily increased over the years to the point
where large trucks now account for over 50% of vehicle traffic on some
highways in Virginia. Projections now forecast that large-truck
travel will grow at twice the rate of personal vehicle travel in the
near future.2 This will probably result in increasing numbers of
large-truck crashes, particularly on two-lane secondary roads, as
motorists move from the congested interstate and primary roads onto
the secondary road system.

Several studies have been conducted to determine the effect of
large-truck operation on multilane primary and interstate highway
systems. However, two-lane secondary roadways have been largely
disregarded under the assumption that crash characteristics and crash
rates for large trucks are the same for these roads as for multilane
and interstate highways. In fact, geometric characteristics of two-
lane secondary routes often limit maneuverability and visibility for
large vehicles, thereby increasing the risk of a crash involving a
large truck. One study determined that collisions between passenger
cars and large trucks on undivided rural roads are more severe than on
divided roads under all conditions. This disparity is even more
predominant during

nighttime hours.3 These results pertain particularly to secondary
roads, which tend to be undivided and, more often than not, two lane.
Sharp curves, steep grades, and limited sight distance are common on
secondary roads throughout the United States and are particularly
abundant in certain areas of Virginia where mountains and other
natural land features have dictated many of the road locations and
designs. These limitations can lead to more large-truck crashes and
traffic delays on these two-lane secondary roads than on multilane
highways.

Other pertinent factors with regard to secondary roads are the wide
variations in their physical characteristics and average annual daily
traffic (AADT). For example, lane width varies from about 2.44 m (8.0
ft) to 3.66 m (12.0 ft), and AADT can be as low as 30 and as high as
30,000.

Since statistics suggest that large-truck traffic will continue to
increase on two-lane secondary roads, and their inherent geometric
characteristics will not change, suitable countermeasures must be
identified that will reduce the risk of large trucks being involved in
crashes on these roads. Most often, a large-truck crash is caused by
driver error or inattention rather than a problem with the highway
environment. Traditional improvements to the roadway such as
geometric changes, increased/improved signing, and altered/improved
pavement markings cannot counter driver error significantly, or have
not appeared to be effective to this point. It is, however, feasible
that advanced technologies, particularly those associated with
intelligent transportation systems (ITS), which consist of a number of
technologies including information processing, communications,
control, and electronics, could be used to minimize the effect of the
lower geometric standards and other causal factors of large-truck
crashes on two-lane secondary roads.

Toward this end, the Virginia Transportation Research Council
(VTRC) conducted this study to determine the crash characteristics and
identify the causal factors of large-truck crashes on secondary roads
and identify ITS and other advanced technologies that would eliminate
or minimize the effect of the identified factors. In addition, the
study contributed to achieving the Virginia Department of
Transportation's (VDOT's) fundamental objective to use ITS technology
to enhance safety.

PURPOSE AND SCOPE

The purpose of this project was to identify the predominant
collision types and the principal causal factors of large-truck
crashes on two-lane secondary roads in Virginia through fault tree
analysis and then identify existing or future ITS and other advanced
technologies that could be used to develop appropriate countermeasures
to eliminate or reduce the detrimental effects of the identified
factors. The specific objectives were:

1. Determine the characteristics of crashes involving large trucks
on two-lane secondary roads.

2. Determine whether large-truck crash rates for two-lane secondary
roads are significantly different from those for two-lane
primary roads.

3. Determine whether large trucks are over-represented in crashes
on two-lane secondary roads.

4. Identify predominant causal factors for large-truck crashes on
two-lane secondary roads.

5. Identify existing or proposed ITS and other advanced
technologies that could be used to develop countermeasures that
would minimize or eliminate large-truck crashes on two-lane
secondary roads.

LITERATURE REVIEW

Sources for identifying information relevant to this study included
the Transportation Research Information Service (TRIS), the VTRC
Library, and the University of Virginia libraries. Completed studies
over the past 15 years relating to this project were identified and
their reports reviewed. In addition, reports on ITS and other
advanced technologies were continuously reviewed in order to ensure
inclusion of the most up-to-date material on this ever changing field.
The materials reviewed were classified under the following
subheadings:

- large-truck crash characteristics on two-lane primary roads

- large-truck crash characteristics on secondary roads

- large-truck access laws

- large-truck safety

- ITS and other advanced technologies.

Large-Truck Crash Characteristics on Two-Lane Primary Roads

Several reports have been published by the Transportation Research
Board relating to the impact of the STAA of 1982 and the TTSA of 1984
on traffic safety on two-lane highways. For example, Hedlund4
reported that large-truck crashes are more likely to result in a
fatality on two-lane primary roads than on four-lane primary roads.
Hedlund attributed this finding to the increased likelihood of high-
speed, head-on collisions on two-lane primary roads. On the other
hand, low-speed crashes, with less severe effects, predominated in
residential and business areas.

Preliminary analysis of large-truck crashes on two-lane primary
highways in Virginia indicated that large trucks have higher rates of
injury crashes, property damage crashes, and overall crashes than
passenger cars.5 Between 1988 and 1990, large-truck fatality rates
increased significantly for two-lane primary roads whereas overall
vehicle fatality rates declined.5 The greatest percentage of crashes
were rear-end collisions and fixed-object-off-the-road crashes.5

Cleveland et al.6 examined the influence of geometric
characteristics and traffic variables on crash rates on rural two-lane
highways by grouping them into compatible classes based on their
geometric and traffic characteristics. A total of 21 models were
tested with different combinations of the variables within each class.
These models were able to explain between 30% and 75% of the variation
in crashes, and the most significant variable was found to be the
average daily traffic (ADT) of the roadway. They concluded that the
geometric characteristics of the roadway were relatively
insignificant. However, their analysis was based on all crashes, not
just on those involving large trucks.

Large-Truck Crash Characteristics on Secondary Roads

Roads classified as secondary roads in the United States can be
grouped into two categories: those that were formally under the
federal-aid system, consisting of 398,000 miles of rural major
collector roads linking towns and smaller communities with the primary
system,7 and those that are off the federal-aid system, consisting of
about 2.6 million miles of two-lane rural highways. These secondary
roads account for 80% of all U.S. road miles.8 In addition, roads in
mountainous and rolling terrain account for more than two-thirds of
the secondary two-lane mileage.8 The roads that traverse these types
of terrain are characterized by steep grades and sharp curves. Within
the secondary system, geometric design standards vary considerably,
the use of traffic control devices is limited, and estimates indicate
that 68% of rural travel and 30% of all travel occur on the rural two-
lane system.8 In addition, 80% to 90% of two-lane crashes occur in
this rural environment and certain crash categories including passing
maneuver, run-off-the-road, and railroad crossing crashes are
predominant among them.8 In Virginia, the total mileage of secondary
roads is about 45,710.

The literature review revealed no specific study on large-truck
crashes on two-lane secondary roads.

Large-Truck Access Laws

The access of large trucks onto secondary roads is governed by both
federal and state laws. Although a state's regulations cannot reduce
the restrictions called for by federal regulations, they can further
restrict large-truck access over and above that of the federal
regulations. This in some cases results in litigation. For example,
prior to 1988, Virginia had state regulations in effect that prevented
single pup trailer units 2.6 m (102 in) wide from reaching local
customer

Table 1. Federal and Virginia Codes Regulating Large-Truck Access
Statute

Federal Regulations Pertinent Language

49 U.S.C.S. Appx. 2311(a) (1992) No state shall impose regulation on
tractor-trailer combinations of
less than 48 ft [14.6 m] or semi-
trailers of less than 28 ft [8.5 m]

49 U.S.C.S. Appx. 2311(b) Federal regulation is limited to
only trailers

49 U.S.C.S. Appx. 2311(i) The Governor can petition the
Department of Transportation for
any road to be exempted from
federal length regulation for
purely safety reasons

49 U.S.C.S. Appx. 2312(b) States can impose reasonable
regulations on trailers less than
28.5 ft [8.7 m] for purely safety
reasons

49 U.S.C.S. Appx. 2316(a) Prohibits state regulation of
trailers less than 102 in [2.6 m]
wide (does not include safety
devices in measurement)

49 U.S.C.S. Appx. 2316(e)(1) The Governor can petition the
Department of Transportation for
any road to be exempted from
federal width regulation for purely
safety reasons

Virginia Regulations

Va. Code Ann. 46.2-1109 (1993) Limits commercial vehicles to 102
in [2.6 m] wide and trailers to
28.5 ft [8.7 m] long

Va. Code Ann. 46.2-112 (1993) Vehicle load combinations must be
less than 60 ft [18.3 m] and
tractor-trailer combinations must
be less than 48 ft [14.6 m]

points of loading and unloading on two-lane secondary roads. In 1988,
as a result of A.B.F Freight System, Inc. v. Suthard, 681 F.Supp. 334
(1988), the Virginia statutes were found to prohibit "reasonable
access" guaranteed for single pup trailers in 49 U.S.C.S. Appx. 2312
and were struck down. A summary of the current federal and state laws
relevant to large-truck access in Virginia is given in Table 1, and a
summary of the size and weight requirements is given in reference 9.

Large-Truck Safety

Truck crashes are complex events; several factors interact to
contribute to their occurrence: the vehicle, the driver, the
environment, or another vehicle. Evidence indicates that the driver
contributes most significantly to truck crashes since driver action or
inaction can frequently precipitate or prevent the occurrence of a
crash. Driver factors (i.e., fatigue, inattention, driving

under the influence of drugs or alcohol, driving at an excessive speed
for prevailing conditions, and poor judgment) have contributed to a
large portion of truck crashes investigated by the National
Transportation Safety Board.10 Nevertheless, as with most vehicle
crashes, commercial vehicle crashes are caused by the interaction of
human, environmental, and vehicle factors. Although most crashes are
attributed to driver error, this classification often obscures the
fact that safety enhancements along the roadway and integrated into
the vehicle can decrease the probability that a driver will make a
serious error and a crash will occur.11

With the introduction of the longer and wider tractor-trailers,
questions have arisen about the safety of these vehicles. Many
secondary roadways have lane widths of only 3.1 m (10 ft) or less.7
These roads were not designed for and are not able to carry the
larger, heavier trucks that now dominate the trucking industry. A
study12 conducted by the Highway Safety Research Center at the
University of North Carolina and the Scientex Corporation found that
on high speed rural two-lane and multilane roads, the tractor-trailers
2.6 m (I 02 in) wide encroached on lane edges and operated slightly
closer to the center line than did those 2.4 m (96 in) wide. In
another study,13 Donaldson concluded that the operation of long, wide
trucks, especially on two-lane, two-way roads with substantial
geometric deficiencies, significantly compromised the safety of
automobile motorists.

Gericke and Walton10 stressed that prospective increases in the
length of trucks will correspondingly increase aborted passing
maneuvers of automobiles and will thereby increase safety hazards.
Olsen et al.11 also found that for controlled stops in which the
truck driver modulates his or her brakes to prevent spinning or
jackknifing and maintains steering control, trucks require stopping
distances that are approximately 1.4 times those required for
automobiles. Two studies13,14 concluded that many curves with lanes
less than 3.7 m (12 ft) wide on two-lane two-way roads cannot be
properly and safely negotiated by a large truck even when it is
traveling at the posted speed.

A 1992 study by Garber and Patel14 showed that large-truck crash
rates are significantly higher when lane widths are 3.1 m (10 ft) or
less on multilane highways and that steep grades and narrow lanes also
increase the probability of crashes on multilane roads. For example,
during 1987 to 1989, the crash rate of tractor-trailers with trailer
widths greater 2.4 m (96 in) in Virginia was 584 per 100 million VMT
on a sample of primary roads having lanes 3.1 m (10 ft) wide and only
203 per 100 million VMT on similar roads having lanes 3.7 m (12 ft) or
wider.14 These statistics are of particular importance for two-lane
secondary roads when it is noted that 88.9% of these roads in Virginia
have lane widths of 3.1 m (10 ft) or narrower.

There are also intrinsic characteristics of large trucks that
increase the potential of these vehicles to be involved in crashes,
particularly on two-lane secondary roads. For example, the stopping
sight distance given in AASHTO guidelines for crest vertical curves
are much shorter than the actual stopping distance for trucks while
maintaining directional control.15 The primary factors that
contribute to the longer stopping distances are inferior truck tire
properties on poor, wet roads; poor braking efficiencies of heavy
trucks; and poor driver control efficiencies in

modulating the brakes to avoid wheel lock. Fancher concluded that
vertical curves designed for speeds of 60 mph or more in accordance
with AASHTO guidelines are adequate only for trucks traveling 52 mph
or less.15 The majority of the secondary roads in Virginia have a
legal speed limit of 55 mph.

Another issue is the comparison of the crash rates of large trucks
with those of other vehicles. Based on the results of previous
studies,5,13,14 there is no consensus as to whether the crash rate of
large trucks is significantly higher or lower than that of all other
vehicles. Most truck crash studies, nevertheless, appear to indicate
that the fatal crash rate of large trucks is much higher than that for
passenger cars. For example, from 1988 to 1990, large-truck fatal
crash involvement in Virginia increased 2.1 % whereas automobile fatal
crash involvement decreased 0.8%." These statistics raise questions
regarding the overall safety of large trucks on U.S. highways,
especially on secondary roads where standards for geometric
characteristics are usually lower than those for primary and
interstate highways.

ITS and Other Advanced Technologies

ITS technologies are based on the integration of the elements of
surface transportation systems (the vehicle, the infrastructure, and
the traveler) into a single system through communication, information,
and control functions. ITS technologies use state-of-the-art
microelectronics to achieve this integration. These technologies are
capable of a wide variety of functions. Examples of advanced
technologies that have been implemented for safety reasons are
antilock brakes and airbags. ITS technologies currently in use or
being developed that could affect traffic safety include weigh-in-
motion of large trucks, automatic vehicle identification, collision
warning and/or avoidance, traveler information, traffic condition
reporting, alternate route selection, and incident management.
Research is being initiated to determine if these technologies are
economically, technically, socially, and politically feasible.

Current ITS Technologies Available

Current literature on ITS technologies indicates that research is
underway or being planned that will lead to the development of
different types of equipment that would have a significant impact on
the highway system. It is envisioned that by the year 2000 these will
include the following:17

- longitudinal and lateral collision avoidance systems

- "smart" traffic signal systems that genuinely maximize the
efficient use of roads, reducing stops and delays

- aids to tell drivers where they are when traveling in unfamiliar
areas and show the location of their destinations

- driver information systems that display congestion information
and assist the driver in selecting the best route

- a general facility for providing a variety of information to
travelers that is tailored to their locations and needs as ITS
comes to constitute an information utility

- devices to sense lapses in driver performance and aid in driving
tasks (e.g., cruise control that responds to changes in speed
and distance of the vehicle ahead)

- systems to help police, fire, ambulance, and transit services
dispatch their vehicles as quickly as possible to where they are
most needed (e.g., enhanced 91 1)

- systems to improve the efficiency of truck operations, reducing
paperwork and delays and thereby helping to reduce the cost of
all goods shipped by truck

- "may day" systems that will speed emergency response to
accidents in rural areas.

Visions of the Future

Researchers have indicated that ITS can reduce traffic fatalities
by 8% by the year 2011.18 That percentage translates to 3,300 lives
saved and 400,000 injuries prevented each year at current traffic
levels. These figures, however, could prove to be quite conservative.
Future advanced technology could "ensure the driver's own state of
fitness, enhance driver perception on a continuous basis, give warning
of impending danger, intervene with emergency control if a crash is
imminent, and perhaps eventually automate the driving process on
specialized roads."18 For example, the next generation cruise
control system will automatically slow the vehicle to maintain a safe
headway from the vehicles ahead. Further, impending departure from
the roadway will be predicted by on-board electronics using a lane
tracking system and the driver will be alerted in time to recover.
Also, a cooperative intersection will communicate data on the state of
the traffic signal and the presence of conflicting traffic so as to
avoid intersection collisions.18

Summary

The literature review showed how little is known about large-truck
safety on two-lane secondary roads. An important result that has been
shown from some studies is that large trucks are over-represented in
crashes on two-lane roads and that the severity of a crash involving a
large truck is usually greater than a similar crash involving a
passenger car on a two-lane road. The

literature also speculates that ITS and other advanced technologies
can be used to increase the safety of these large trucks.

METHODOLOGY
Data Collection

The data collection task consisted of two subtasks: collection of
field data on truck percentages within the traffic streams, and
extraction of the relevant crash data on secondary highways from
VDOT's computerized data files.

Sample Size, Site Selection, and Collection of Field Data

Currently, there are very limited data available on large-truck
AADT on secondary roads in Virginia. This deficiency necessitated the
collection of data on vehicle classification and total volumes on
these secondary roads as this information was needed for the
computation of crash rates. Unfortunately, it was not feasible for
data to be collected on each of the approximately 7,000 secondary
roads in the state. A representative sample of roads was therefore
selected for which data were obtained.

The secondary roads in each VDOT district were grouped into
clusters, with each cluster consisting of all roads having AADTs
within a specific range. It was originally intended to collect data
at a statistically selected sample of roads from each cluster, but due
to the large variation in large-truck percentages found between routes
in different districts within the same cluster, the authors determined
that the cluster methodology would not be suitable for analysis. For
example, the tractor-trailer percentages on routes in the 0-1000 AADT
cluster varied from 0.0% to 5.3% and had a standard deviation of 1.77.
This large standard deviation was unacceptable since the average
tractor-trailer percentage was only 1.35%. Therefore, the data
collection and statistical comparison were carried out for each
district rather than for the different clusters based on the AADTs.
The authors believed that this analysis procedure was better than the
original procedure based on AADT clusters as effects of variations in
land use and topography are minimized by considering needs within each
district. Consequently, the required sample size for each district
was then determined for a �1% tolerance level and a 95% confidence
level using the following equation:

Click HERE for graphic. (1)

where:

n = sample size for district i
i

t = standard two-tail t value at level of confidence (1-�)
(1-�/2)

S� = variance of truck percentages on two-lane highways in
i district i

d = tolerance level (1%)

n = total number of elements in population.

The sample size calculated and the actual sample size used for
each district are shown in Table 2.

A random selection of the calculated number of study roads required
for each district was then made, and data on volume and vehicle
classification were collected for each of those roads. A total of 124
roads were randomly selected, and data were collected for at least a
48-hour period on large trucks and passenger cars, vans and pickups
separately using Streeter Amet electronic counters. These percentages
were then used to determine actual truck volumes from the AADT on each
secondary road, which in turn were used to determine truck VMT on each
road.

Table 2. District Sample Sizes

Calculated Based on Calculated Based on
Single-Unit Truck Tractor-Trailer Actual Sample
District Percentage Percentage Size Used

Bristol 10 6 10
Salem 7 2 13
Lynchburg 11 3 12
Richmond 7 3 18
Suffolk 8 11 12
Fredericksburg 10 5 11
Culpeper 9 6 21
Staunton 7 3 9
Northern Va 11 3 18

Due to the lack of available data and time, seasonal variation of
the large-truck AADTs was not investigated. The authors believed that
this limitation would have a minimal effect on the results of the
study since the AADT data were mainly used to estimate the crash rates
for comparison purposes and for identification of large-truck crash
trends rather than for calculation of absolute crash rates. In
addition, the monthly variation of the number of large-truck crashes
was small, and all counts were made during the summer months of June
and August. Further, in order to determine the predominant causal
factors from which the countermeasures were identified, fault tree
analysis was used. This methodology is based on the proportion, not
the rate, of different types of crashes occurring under specific
conditions and was the primary tool used to identify the applicable
ITS countermeasures.

Compilation of Accident Data

Data on crashes in Virginia were obtained from the crash report
forms that are completed by police officers for each crash involving a
fatality, personal injury, or property damage exceeding a specified
dollar amount. This amount was $500 between July 1988 and June 1989,
and $750 between July 1989 and June 1992, and is currently $1,000.
Information on these reports is stored in computerized files and was
available from VDOT. Crash data for 1988 through 1990 were extracted
from these computerized files, compiled for each secondary road, and
categorized with respect to vehicle type. Specific information on
type of crash, severity, number of vehicles involved, and causal
factor was obtained for each crash from the crash data files.

The width and length of a large truck involved in a crash were not
recorded in the crash report and were, therefore, not available for
analysis purposes. This study, therefore, determined the crash rate
based on truck type (single-unit truck or tractor-trailer) and not on
truck width or length.

Computation of Accident Rates

The first step was to compute the VMT on each road for each vehicle
type using the following equation:

VMT = LENGTH x P x AADT x 365 (2)
i i

where:

VMT = VMT for vehicle type i (single-unit trucks, tractor-
i trailers, or passenger vehicles)

Length = length of road segment

P = percentage of vehicle type i in the traffic stream
i

AADT = average annual daily traffic on the road.

The VMT computed for each road was then used in the second step to
compute crash rates by type and severity in number of crashes per 100
million VMT. These crash rates were computed for each vehicle type
separately and for all vehicles together. Existing data on truck VMT
and crashes were used to determine the large-truck crash rates for
two-lane primary roads.

Statistical Analysis

The t test and proportionality test were used to find significant
differences at � = 0.05. The t test was used to test the following
null hypotheses:

1. Large-truck total (fatal, injury, and property-damage-only
[PDO]) crash rates for two-lane primary roads and for two-lane
secondary roads were not significantly different.

2. Large-truck fatal crash rates for two-lane primary roads and for
two-lane secondary roads were not significantly different.

3. Total (fatal, injury, and PDO) crash rates for two-lane
secondary roads for large trucks. and passenger cars were not
significantly different.

4. Fatal crash rates for two-lane secondary roads for trucks and
passenger cars were not significantly different.

The proportionality test was used to test the following null
hypotheses:

5. The proportion of each collision type of large-truck crashes on
two-lane secondary roads was not significantly different from
the corresponding proportion on two-lane primary roads.

6. The proportion of each collision type of large-truck crashes on
two-lane secondary roads was not significantly different from
the corresponding proportion for passenger car crashes on two-
lane secondary roads.

7. The proportion of each crash severity level in large-truck
crashes on two-lane secondary roads was not significantly
different from the corresponding proportion on two-lane primary
roads.

8. The proportion of each crash severity level in large-truck
crashes on two-lane secondary roads was not significantly
different from the corresponding proportion in passenger car
crashes on two-lane secondary roads.

The results of these analyses indicated to what extent large-truck
crash characteristics on secondary roads were different from those on
primary roads and to what extent large-truck crash characteristics
were different from those for passenger cars on secondary roads.

Identification of Significant Causal Factors

In this study, the "fault tree" approach, which is an application
of the branched-events chain theory, was used to identify the
predominant crash causes. The branched-events chain theory is based
on the assumption that the likelihood of a crash occurring can be
determined if the pathway leading to the crash can be identified.
Branched-event theory is logically adaptable to traffic crashes, and
it describes a crash phenomenon as a chain of events leading to the
top event. A fault free is a model that graphically and logically
represents the various combinations of possible events occurring in a
system that leads to the top event. The top event is an undesired
outcome of some process, which in this case was the occurrence of a
large-truck crash.

The paths of the fault tree were used to identify the sequences and
relationships between basic events and the top event. The objective
of the analysis was to determine the shortest failure path for each
type of truck crash, thereby identifying the predominant causal
factors. These shortest failure paths are known as minimum cut sets.
A cut set is a set of events whose occurrence leads to the occurrence
of the top event. A minimum cut set is obtained when a cut set cannot
be reduced further and the occurrence of the top event is still
ensured.

The significant causes were identified by first calculating the
probability of the top event on all minimum cut sets. The minimum cut
set associated with the highest probabilities for the top event
contained the significant causes.

In order to determine the probabilities of the minimum cut set, the
Boolean representation used for coherent structure by Birnbaum et al."
is used. The probabilities are calculated as follows:

Click HERE for graphic. (3)

Click HERE for graphic.

Identification of ITS and Other Advanced Technologies

ITS technologies are classified into five main categories: (1)
advanced traffic management systems (ATMS), (2) advanced traveler
information systems (ATIS), (3) commercial vehicle operations (CVO),
(4) advanced vehicle control systems (AVCS), and (5) advanced public
transportation systems (APTS). The capabilities of ITS have recently
been classified into 28 "user services" grouped into the following six
general categories: (1) travel and traffic management, (2) public
transportation management; (3) electronic payment services, (4)
commercial vehicle operations, (5) emergency management, and (6)
advanced vehicle safety systems.20

The user services that are considered to be related to this study
are two in the commercial vehicle category, i.e., automated roadside
safety inspection and on-board safety monitoring, and all services in
the advanced vehicle safety systems category, i.e.,

- longitudinal collision avoidance

- lateral collision avoidance

- intersection collision avoidance

- vision enhancement for crash avoidance

- safety readiness

- pre-crash restraint deployment

- automated vehicle operation.

A brief description of these services follows.

- Automated roadside safety inspection. This service will
facilitate roadside inspection and provide for real-time access
to the safety performance record of carriers, vehicles, and
drivers. This in turn will aid in the identification of those
vehicles and/or drivers that will be stopped for inspection.
Also, sensors and diagnostics will be used to rapidly and
accurately check vehicle systems and driver fitness for duty.
Technologies under this service will aid in significantly
reducing large-truck crashes on two-lane secondary roads in
Virginia if vehicle failure is identified as a predominant
causal factor of these crashes.

- On-board safety monitoring. This will involve the continuous
monitoring of the vehicle and driver at mainline speeds.
Vehicle monitoring will entail the collection and analysis of
data on the condition of critical vehicle components, such as
brakes, tires, lights, and the determination of thresholds of
warnings and countermeasures. It is envisioned that the
monitoring of the driver will involve the use of nonintrusive
technologies to obtain driving time and alertness, which can be
used as a basis to warn drivers or enforcement officers of
critical driver conditions. These monitoring technologies may
also be useful in reducing large-truck crashes on two-lane
highways in Virginia if vehicle failure or, to a certain extent,
driver failure is identified as a predominant causal factor of
these crashes.

- Longitudinal collision avoidance. This will help prevent head-
on and rear-end collisions by providing a means of sensing
potential or impending collisions and then prompting the driver
to take an avoidance action or temporarily control the vehicle.
Technologies developed under this user service's group may be
effective in reducing large-truck crashes on secondary roads in
Virginia if head-on and rear-end collisions are predominant in
these crashes.

- Lateral collision avoidance. This will help prevent sideswipe
collisions, which often occur as a result of vehicles leaving
their lane of travel, by providing crash warnings and controls
for lane changes and road departures. If sideswipe collisions
are identified as a predominant type of crashes of large trucks
on secondary roads in Virginia, technologies in this group may
be effective in reducing the large-truck crash rates.

- Intersection collision avoidance. This service will help
prevent crashes at intersections by informing the driver of an
imminent collision when he or she is approaching or crossing an
intersection with a traffic control device or by alerting the
driver when the right of way at the intersection is unclear or
ambiguous. If large-truck crashes at two-lane secondary road
intersections are over-represented, technologies in this area
may significantly reduce large-truck crashes on two-lane
secondary roads in Virginia.

- Vision enhancement for crash avoidance. This service will
improve the driver's ability to see more clearly the roadway and
any obstacles on or along it. Technologies in this area may aid
in significantly reducing large-truck crashes on two-lane roads
in Virginia if environmental factors such as fog, heavy storms,
snow, nighttime, etc., that inhibit driver vision are found to
be predominant causal factors for these crashes.

- Safety readiness. This service will help prevent crashes that
are mainly due to driver fatigue by the installation of in-
vehicle equipment that could unobtrusively gauge a driver's
condition and provide a warning if he or she is drowsy or
otherwise impaired. The service can also be used to monitor
critical elements of a vehicle internally and detect unsafe road
conditions, such as bridge icing and standing water on the
roadway, and then warn the driver of these conditions.
Technologies in this area may significantly reduce large-truck
crashes in Virginia if driver conditions such as fatigue and/or
environmental conditions such as ice and wet weather are found
to be predominant causal factors of these crashes.

- Pre-crash restraint deployment. This service will provide
technologies that will help reduce the severity of a crash
rather than prevent one. These technologies will identify the
velocity, mass, and dimensions of the vehicle and objects
involved in a potential crash and deploy restraint features such
as airbags, lap/shoulder belts, etc., at an optimal pressure.

- Automated vehicle operations. These technologies will provide a
fully automated "hands off" operating environment. This is,
however, a long-term goal of ITS and was not considered in this
study.

By using the information obtained from the literature review and
telephone interviews with several manufacturers, appropriate
technologies that had been developed or proposed and were related to
the user services considered in this study were identified and
classified into subcategories based on their proposed specific use in
the field. A list of these technologies and their individual status
are given in Appendix A.

Based on the predominant causal factors determined from the fault
tree analysis and the predominant collision types identified, the
appropriate user services were identified and suitable ITS
technologies within these services were selected to eliminate or
minimize the effect of these causal factors.

Since tests have not been completed, or in many cases not even
begun, to evaluate the effectiveness of these technologies in reducing
the number of crashes caused by an identified causal factor, this
criterion could not be used in the selection process. The basis for
selection of each technology was its current stage of development and
perceived effectiveness.

RESULTS

Crash Characteristics
Percentage Distribution of Crashes by Time of Day

Figure 1 shows the percentage distribution of large-truck crashes
on two-lane secondary roads by time of day. The information for
passenger cars was included to aid in the interpretation of the data.
For single-unit trucks, there was little difference in the percentage
of crashes occurring throughout the daylight hours (7 A.M.-5 P.M.).
However, a peak existed in the late afternoon (3 P.M.-4 P.M.).
Tractor-trailer crashes showed two peaks: at 11 P.M. and at 3 P.M.,
with a wide variation in the percentage of crashes occurring during
the hours between the two peaks. Crashes involving passenger cars, on
the other hand, increased from about 9 A.M. to 5 P.M. and then
dropped. A relatively low percentage of these crashes occurred during
the night period, suggesting that poor visibility due to increasing
darkness was not a significant causal factor of large-truck crashes on
two-lane secondary roads. However, it is likely that the reduction in
truck volumes resulted in a relatively lower number of crashes but a
relatively higher crash rate at nighttime. Not enough information is
available at this time, however, to draw this conclusion as the
results shown in the graph are for percentages of crashes occurring
and not for crash rates.

Click HERE for graphic.

Figure 1. Percentage Distribution of Large-Truck Crashes on Two-Lane
Secondary Roads by Time of Day

Percentage Distribution of Crashes by Month

Figure 2 shows the percentage distribution of crashes by month of
occurrence for each vehicle type. This figure does not show any
regular pattern of variation through the year for any of the vehicle
types except that the highest percentage of crashes involving single-
unit trucks occurred in June; October and June seem to be the worst
months for tractor-trailers, and May and October seem to be the worst
months for passenger cars. Due to the unavailability of large-truck
VMT for each month of the year, it was not possible to calculate the
crash rates for each month to see if the rates throughout the year
varied more significantly than the number of crashes.

Click HERE for graphic.

Figure 2. Percentage Distribution of Large-Truck Crashes on Two-Lane
Secondary Roads by Month of Year

Percentage Distribution of Crashes by Collision Type

Figure 3 shows the percentage distribution of crashes by collision
type for each vehicle type studied. Angle crashes, which are
primarily intersection crashes, seemed to be the predominant collision
type for all types of vehicles. The next order of predominance for
passenger cars and single-unit trucks was rear end, sideswipe opposite
direction (SSOD), and sideswipe same direction (SSSD), and for
tractor-trailers was SSOD, rear end, and SSSD. These predominant
collision types accounted for over 75% of crashes on two-lane
secondary roads involving single-unit trucks and tractor-trailers.
Technologies that can significantly reduce these types of crashes may
significantly reduce the potential of large-truck crashes on two-lane
secondary roads. The SSSD crashes are characteristic of passing
maneuvers, whereas the SSOD crashes are characteristic of vehicles
straying into the lane of the oncoming traffic, which in turn reflects
the characteristic of a large vehicle traveling on a narrow road.

Click HERE for graphic.

SSOD - Sideswipe opposite direction
SSSD - Sideswipe same direction
FOOR - Fixed object off roadway

Figure 3. Percentage Distribution of Large-Truck Crashes on Two-Lane
Secondary Roads by Collision Type

Percentage Distribution of Crashes by Severity

The percentage distributions of crashes by severity are shown in
Figure 4. This figure shows that tractor-trailers have a higher
percentage of fatal and injury crashes than both single-unit trucks
and passenger vehicles. This finding coincides with what has been
found on Virginia's primary highways.

Statistical Analysis

Tables 3 and 4 show the average crash rates by severity for each
district for the secondary and primary roads, respectively. These
crash rates were used to test null hypotheses 1 through 4 using the t
test. The proportionality test was used to test null hypotheses 5
through 8. These tests were carried out to determine significant
differences at a significance level of � = 0.05.

Null Hypothesis 1:
Large-truck total crash rates for two-lane primary roads and for
two-lane secondary roads were not significantly different.

Table 5 shows the results of the analysis. When the crashes were
considered using overall crash rates, the Richmond, Staunton, and
Northern Virginia districts had significantly higher crash rates for
the two-lane secondary roads than for the two-lane primary roads and
there was no significant difference in the other districts. When the
crash rates were considered for all

Table 3. Secondary Route Crash Rates for Large Trucks

Crash Rate (per 100 million VMT)
District Overall Fatal Injury PDO

Bristol 276.8 5.7 22.2 248.6
Salem 917.5 0.0 293.4 624.0
Lynchburg 1859.7 13.1 144.6 1702.2
Richmond 1427.2 4.7 332.8 1095.6
Suffolk 120.5 3.5 23.7 91.7
Fredericksburg 1386.2 0.0 539.1 847.1
Culpeper 688.5 0.0 271.7 427.7
Staunton 1145.7 15.1 556.1 574.5
Northern Va 764.8 0.0 253.9 510.9
Average 951.1 3.8 268.2 681.5

Click HERE for graphic.

Figure 4. Percentage Distribution of Large-Truck Crashes on Two-Lane
Secondary Roads by Severity

Table 4. Primary Route Crash Rates for Large Trucks

Crash Rate (per 100 Million VMT)
District Overall Fatal Injury PDO

Bristol 232.3 0.0 101.2 131.0
Salem 497.0 5.0 406.7 85.3
Lynchburg 647.1 0.0 297.6 349.7
Richmond 227.2 0.0 126.3 100.9
Suffolk 228.7 10.5 92.3 125.9
Fredericksburg 289.2 31.7 59.0 198.5
Culpeper 426.3 0.0 70.5 352.8
Staunton 103.3 0.0 65.3 38.0
Northern Va 232.4 0.0 100.2 132.2
Average 320.1 5.2 146.6 168.3

districts combined statewide, the results indicated that the large-
truck crash rates for two-lane secondary roads were significantly
higher than for two-lane primary roads. This significant difference
was, however, due to the significant difference in PDO crashes. Null
hypothesis 1 was therefore rejected when the entire state was
considered. This was expected since it was already noted that large-
truck crashes increased when the lane width fell below 3.1 m (10
ft).14 For the two-lane primary roads in Virginia, only 7.5% of these
roads have lane widths less than 3.1 m (10 ft). In contrast, 75.7% of
the two-lane secondary roads have lane widths less than 3.1 m (10 ft).
However, the significant variation found in the analysis was due
primarily to significantly higher large-truck crash rates in the
Richmond, Staunton, and Northern Virginia districts.

Null Hypothesis 2:
Large-truck fatal crash rates for two-lane primary roads and for
two-lane secondary roads were not significantly different.

Table 5 shows the results of the analysis. It can be seen that
when crashes were considered by severity, the fatal crash rates for
two-lane secondary roads were not significantly different than for
two-lane primary roads. Null hypothesis 2 could, therefore, not be
rejected. This was also true in most cases for injury and PDO crashes
with the exception of the Richmond District, where both injury and PDO
crash rates were significantly higher; the Culpeper District, where
the injury rate was significantly higher; and the Northern Virginia
District, where the PDO rate was significantly higher.

Table 5. t Values for Large-Truck Crash Rates for Two-Lane
Secondary vs. Two-Lane Primary Roads

t Value
District
Overall Fatal Injury PDO

Bristol 0.34 1.44 -2.31 0.86
Salem 0.93 -1.00 -0.34 1.65
Lynchburg 0.76 1.00 -0.48 0.90
Richmond 3.61** 1.00 2.17** 3.34**
Suffolk -0.95 -0.65 -1.31 -0.39
Fredericksburg 1.19 -1.00 1.56 1.05
Culpeper 1.05 - 2.25 0.28
Staunton 2.17 1.00 1.28 1.70
Northern Va 3.40 - 2.03 3.28
Statewide 3.24 -0.34 1.83 3.00

Note: t values that are significantly higher at a = 0.05 are
shown with **. tcrit = 2.13.

Null Hypothesis 3:
Total crash (fatal injury and PDO)rates for two-lane secondary
roads for large trucks and passenger cars were not significantly
different.

Table 6 shows that the crash rate for large trucks on secondary
roads was significantly higher than for passenger cars on the same
roads when the entire state was considered. Therefore, null
hypothesis 3 was rejected. This significant difference is due to
three of the nine districts, Richmond, Culpeper, and Northern
Virginia, which had significantly higher crash rates for large trucks.

Table 6. t Values for Total Crash Rates for Large Trucks vs.
Passenger Cars on Two-Lane Secondary Roads

t Value
District
Truck

Bristol -0.21
Salem 1.68
Lynchburg 0.87
Richmond 3.26**
Suffolk -1.84
Fredericksburg 0.07
Culpeper 2.65**
Staunton 1.42
Northern Va 2.74**
Statewide 2.62**

Note: t values that are significant at � = 0.05 are shown
in bold. tcrit = 2.13.

Null Hypothesis 4:
Fatal crash rates for two-lane secondary roads for trucks and
passenger cars were not significantly different.

The results shown in Table 7 indicate that large trucks did not
have significantly higher fatal crash rates than passenger cars on
two-lane secondary roads. Therefore, null hypothesis 4 could not be
rejected. Although these results indicate no significant difference
in the fatal crash rates between large trucks and passenger cars, it
is feasible that the proportion of fatal crashes in large-truck
crashes might have been significantly different than that for
passenger cars. This was tested with null hypothesis 8.

Table 7. t Values for Fatal Crash Rates for Large Trucks vs.
Passenger Cars on Secondary Roads

District t Value

Bristol 1.27
Salem -2.05
Lynchburg 0.42
Richmond 0.10
Suffolk 0.42
Fredericksburg -1.59
Culpeper -2.61
Staunton 0.44
Northern Va -2.59
Statewide 0.22

Note: t values that are significant at � = 0.05 are
shown with **. tcrit = 2.13.

Null Hypothesis 5:
The proportion of each collision type of large-truck crashes on
two-lane secondary roads was not significantly different from the
corresponding proportion on two-lane primary roads.

Table 8 shows the results of this test. Based on these results,
null hypothesis 5 was rejected for angle, head-on, SSOD, and backed-
into large-truck crashes.

Null Hypothesis 6:
The proportion of each collision type of large-truck crashes on
two-lane secondary roads was not significantly different from the
corresponding proportion for passenger car crashes on two-lane
secondary roads.

Table 8 shows the results of this analysis. Based on these
results, this null hypothesis was rejected for SSSD, SSOD,
noncollision, miscellaneous, and backed-into collisions. The large-
truck proportions of these collision types were significantly higher
than those for passenger cars.

Null Hypothesis 7:
The proportion of each crash severity level in large-truck crashes
on two-lane secondary roads was not significantly different from
the corresponding proportion on two-lane primary roads.

Table 8. Results of Proportionality Test for Large-Truck Crashes on
Two-Lane Secondary Roads by Collision Type

Z Value
Large Trucks vs
Collision Type Large Trucks Passenger Cars
Secondary vs Primary Secondary Roads

Rear end -1.7361 -0.8808
Angle 2.9193 -7.2039
Head on 3.3044 -1.0658
SSSD -0.2623 3.1637
SSOD 5.5943 4.7015
Fixed object on road -2.3682 0.4155
Train -0.5165 -0.1535
Pedestrian -0.9896 0.4766
Other animal -1.9347 -0.3493
Noncollision -5.2609 1.7289
Backed into 5.0577 6.6108
Fixed object off road -5.8220 1.5147
Deer -2.5028 0.1189
Not stated -0.7304 -0.4942
Miscellaneous -2.2852 4.1061

Note: Z values that are significantly higher at the 95%
confidence level are in bold. Zcrit = 1.96.

Table 9 provides the results of this analysis. Based on these
results, null hypothesis 7 could not be rejected for fatal and injury
but was rejected for PDO crashes.

Null Hypothesis 8:
The proportion of each crash severity level in large-truck crashes
on two-lane secondary roads was not significantly different from
that for the corresponding proportion in passenger car crashes on
two-lane secondary roads.

Table 9. Results of Proportionality Tests for Large-Truck
Crashes on Two-Lane Secondary Roads by Crash Severity

Z Values
Severity Large Trucks Large Trucks vs
Secondary vs Passenger Cars
Primary Secondary Roads

Fatal -3.2288 1.7914
Injury -3.9691 -1.8106
Property damage only 4.8645 1.5493

Note: Z values that are significantly higher at the 95%
confidence level are in bold. Zcrit = 1.645.

The results of this test are shown in Table 9. They indicate that
large trucks had a significantly higher proportion of fatal crashes on
two-lane secondary roads than passenger cars. Consequently, null
hypothesis 8 was rejected for fatal crashes and was not rejected for
injury and PDO crashes. This suggests that although the fatal crash
rates for large trucks were not significantly different from those for
passenger cars as shown in testing hypothesis 4, there was a
significantly higher proportion of large-truck crashes that were fatal
in comparison with passenger cars on two-lane secondary roads.

Fault Tree Analysis

In developing the fault trees, the crashes were first categorized
with respect to their major causal factor. The major causal factors
associated with large-truck crashes can be categorized as driver
related, vehicle related, and environment related (i.e., highway
related and/or weather related). Results showed that driver-related
failure was the leading cause of crashes as shown in Figure 5.
Overall, the percentages of crashes by failure type were consistent
across all vehicle types. Driver-related failures caused between 76%
and 80% of the crashes. Environment-related failures accounted for
10% to 15%, and vehicle-related failures accounted for 1% to 4%, with
single-unit trucks having the highest percentage. Figures 6, 7, and 8
show the percentage breakdown of crash severity by failure type and
vehicle type. Tractor-trailer crashes had the highest fatality
percentage when the crash was caused by a driver-related failure. In
fact, tractor trailers had higher percentages of severe crashes (i.e.,
fatal and personal injury) than both single-unit trucks and passenger
cars for all failure types on two-lane secondary roads.

Click HERE for graphic.

PDO - Property Damage Only

Figure 5. Percentage Distribution of Large-Truck Crashes on Two-Lane
Secondary Roads by Failure Type

Click HERE for graphic.

PDO - Property Damage Only

Figure 6. Percentage Distribution of Driver Related Large-Truck
Crashes on Two-lane Secondary Roads by Severity

Click HERE for graphic.

PDO - Property Damage Only

Figure 7. Percentage Distribution of Environment-Related Large-Truck
Crashes on Two-Lane Secondary Roads by Severity

Click HERE for graphic.

PDO - Property damage only

Figure 8. Percentage Distribution of Vehicle-Related Crashes on Two-
Lane Secondary Roads by Severity

Figures 9 and 10 show the fault trees for driver-related crashes
involving single-unit trucks and tractor-trailers, respectively. The
fault trees for vehicle- and environment-related crashes are shown in
Appendix B. The minimum cut sets were identified and their
probabilities calculated for each fault tree. The minimum cut sets
for driver-related failures are shown in Tables 10 and

Click HERE for graphic.

Figure 9. Fault Tree for Driver-Related Crashes Involving Single-
Unit Trucks

Click HERE for graphic.

Figure 10. Fault Tree for Driver-Related Crashes Involving Tractor-
Trailers

11 for single-unit trucks and tractor-trailers, respectively. Those
for vehicle- and environment-related failures are shown in Appendix C.

Vehicle defects included items such as brake, tire, and light
failures; environment-related failure conditions were categorized as
adverse weather or surface defects. The type of surface defects was
not specified in the Virginia crash data; therefore, this category
could not be developed any further. Adverse weather conditions were
more detailed in the reports and, consequently, could be further
categorized as nonfreezing precipitation, freezing precipitation, and
other lighting and weather problems.

Table 10. Minimum Cut Sets for Single-Unit Truck Crashes Due
to Driver-Related Failure

Path Probability

1-3-7** and 1-4** 0.4204
1-3-8** and 1-4** 0.5416
2-5-9 and 6-11-13 and 6-12->15->17 0.0001
2-5-9 and 6-11-13 and 6-12-15-18 0.0002
2-5-9 and 6-11-13 and 6-12-15-19 0.0010
2-5-9 and 6-11-13 and 6-12-15-20 0.0030
2-5-9 and 6-11-13 and 6-12-15-21 0.0002
2-5-10 and 6-11-13 and 6-12-15-17 0.0001
2-5-10 and 6-11-13 and 6-12-15-18 0.0001
2-5-10 and 6-11-13 and 6-12-15-19 0.0004
2-5-10 and 6-11-13 and 6-12-15-20 0.0012
2-5-10 and 6-11-13 and 6-12-15-21 0.0001
2-5-9 and 6-11-14 and 6-12-15-17 0.0008
2-5-9 and 6-11-14 and 6-12-15-18 0.0012
2-5-9 and 6-11-14 and 6-12-15-19 0.0048
2-5-9 and 6-11-14 and 6-12-15-20 0.0151
2-5-9 and 6-11-14 and 6-12-15-21 0.0008
2-5-10 and 6-11-14 and 6-12-15-17 0.0003
2-5-10 and 6-11-14 and 6-12-15-18 0.0005
2-5-10 and 6-11-14 and 6-12-15-19 0.0019
2-5-10 and 6-11-14 and 6-12-15-20 0.0059
2-5-10 and 6-11-14 and 6-12-15-21 0.0003

** indicates the minimum cut sets with the highest probabilities
and thus the predominant causal factors.

The driver-related failure category was broken down into auditory,
visual, and other permanent handicaps; driver inattention; fatigue;
alcohol and drugs; illness; and driver error. Examples of driver
error are improper passing, straying into the lane of oncoming
traffic, improper turns, speeding, and tailgating. Of the specific
driver factors involved, driver error had the highest frequency,
followed by speeding, impairment due to drugs and/or alcohol, and
driver

Table 11. Minimum Cut Sets for Tractor-Trailer Crashes Due to
Driver-Related Failure

Path Probability

1-3-7** and 1-4** 0.4173
1-3-8** and 1-4** 0.5377
2-5-9 and 6-11-13 and 6-12-15-18 0.0009
2-5-9 and 6-11-13 and 6-12-15-19 0.0009
2-5-9 and 6-11-13 and 6-12-15-20 0.0029
2-5-9 and 6-11-13 and 6-12-15-21 0.0007
2-5-10 and 6-11-13 and 6-12-15-18 0.0003
2-5-10 and 6-11-13 and 6-12-15-19 0.0003
2-5-10 and 6-11-13 and 6-12-15-20 0.0011
2-5-10 and 6-11-13 and 6-12-15-21 0.0003
2-5-9 and 6-11-14 and 6-12-15-18 0.0043
2-5-9 and 6-11-14 and 6-12-15-19 0.0043
2-5-9 and 6-11-14 and 6-12-15-20 0.0148
2-5-9 and 6-11-14 and 6-12-15-21 0.0034
2-5-10 and 6-11-14 and 6-12-15-18 0.0017
2-5-10 and 6-11-14 and 6-12-15-19 0.0017
2-5-10 and 6-11-14 and 6-12-15-20 0.0058
2-5-10 and 6-11-14 and 6-12-15-21 0.0014

** indicates the minimum cut sets with the highest probabilities
and thus the predominant causal factors.

handicap (which includes fatigue, illness, and sleeping). The
specific cause of driver error was not available; therefore, this
category was an undeveloped event in the fault tree analysis.

Based on the probabilities of the minimum cut sets, the following
were the most probable causes of large-truck crashes on two-lane
secondary roads:

- Driver-related failure. For all large trucks, a crash most
often occurred when the driver was not impaired but there was
error in the driver's judgment and either no evasive action was
taken or the evasive action failed. See Tables 1 0 and I 1.
These crashes represented about 74% of all single-unit truck
crashes and about 76% of tractor-trailer crashes on two-lane
secondary roads in Virginia. See Tables 12 and 13.

Table 12. Percentage of Crashes Involving Single-Unit Trucks on Two-Lane Secondary Roads Due to Causal Factors Identified

Percentage of All Single- Percentage of All Single-
Probability of Crash Unit Truck Crashes Unit Truck Crashes Due
Due to Predominant Associated with Type of to Predominant Causal
Type of Failure Predominant Causal Factor Causal Factor1 Failure Factor

Driver related Driver not impaired, but error in
driver judgment and no evasive 0.42042 76.68 32.3
action taken

Driver not impaired, but error in
driver judgment and evasive action 0.54162 76.68 41.5
failed

Vehicle related Brake failure and evasive action
failed or not taken 0.21543 3.8 0.8

Environment Faulty highway component led to
related surface defects and driver evasive
action either failed or was not taken 0.59004 13.44 7.9

1 See Figure 5.
2 See Table 13.
3 See Appendix C, Table C-1.
4 See Appendix C, Table C-3.

Table 13. Percentage of Crashes Involving Tractor-Trailers on Two-Lane Secondary Roads Due to Causal Factors Identified

Percentage of All Percentage of All Tractor-
Probability of Crash Tractor-Trailer Crashes Trailer Crashes Due to
Due to Predominant Associated with Type of Predominant Causal
Type of Failure Predominant Causal Factor Causal Factor Failure1 Factor

Driver related Driver not impaired, but error in
driver judgment and no evasive 0.41732 80.07 33.4
action taken

Driver not impaired, but error in
driver judgment and evasive action 0.53772 80.07 43.1
failed

Vehicle related Vehicle characteristics unfavorably
interacted with driver and/or 0.30753 2.50 0.8
environment

Environment Faulty highway component led to
related surface defects and driver evasive
action either failed or was not taken 0.56004 10.78 6.0

1 See Figure 5.
2 See Table 13.
3 See Appendix C, Table C-2.
4 See Table C-4.

- Vehicle-related failure

- For single-unit trucks, a crash most often occurred when
there was brake failure and the evasive action failed. These
crashes represented about 0.8% of all single-truck-involved
crashes on secondary roads in Virginia. See Table 12.

- For tractor-trailers, a crash most often occurred when the
vehicle characteristics interacted unfavorably with the
driver and/or the environment and no evasive action was
taken. These crashes also represented about 0.8% of all
tractor-trailer-involved crashes on secondary roads in
Virginia. See Table 13.

- Environment-related failure. For all large trucks, a crash most
often occurred when a faulty highway component led to surface
defects and the driver's evasive action either failed or was not
taken. These crashes represented about 8% of all crashes involving
large trucks on secondary roads in Virginia (see Table 12) and
about 6% of all crashes involving tractor-trailers on these roads
(see Table 13).

Selected Advanced Technologies

Based on the most probable causal factors and the predominant
collision types of large-truck crashes that were identified, the
different categories of advanced technologies available or proposed
that could eliminate or mitigate these causal factors and types of
crashes were selected and are described below. It should be
emphasized again that the selection of any specific technology was not
based on a knowledge of its ability to eliminate or reduce these
crashes, as no specific test for which the results are available has
been conducted. The selection was based on the availability of
equipment or its stage of development and its perceived effectiveness.

Technologies Based on Predominant Causal Factors

Driver-Related Failures

For driver-related failure, both single-unit trucks and tractor-
trailers had the same major causal factors. For all large trucks, a
crash most often occurred when there was an error in driver judgment
and the evasive action failed or no evasive action was taken. The
type of driver error was not specified in most cases in the crash data
files. If it was specified, more often that not it was simply
categorized as driver inattention. Traditional countermeasures to
driver inattention or fatigue such as "alertness maintainers" (e.g.,
coffee and loud music) are not particularly effective countermeasures.
In addition, operational rules regulating the hours of service for
commercial drivers have not appeared to eliminate the problem.
Consequently, a continuous status/performance monitoring system is
needed. Technologies under the user services area of onboard safety
monitoring may be effective in mitigating some of the effect of this
causal factor.

Such a monitoring system could either sound an alarm to alert the
driver of the need to return his or her attention to the roadway or
send a warning to enforcement officers in order for them to deal with
the problem.

Vehicle-Related Failure

Under vehicle-related failure, single-unit trucks and tractor-
trailers had two different predominant causal factors. For single-
unit trucks, a crash most often occurred when there was brake failure
and the evasive action failed. The implementation of systems that
will ensure that only single-unit trucks with adequate brakes are
permitted on the two-lane highways or will identify imminent brake
failure conditions and give adequate warning to the driver will result
in a significant reduction in the number of single-unit truck crashes
that are related to vehicle failure. The user services that are
related to this causal factor are the automated roadside inspection
and the on-board safety monitoring system in the commercial vehicle
operations area. The automated roadside inspection system would
ensure that only single-unit trucks with adequate brakes are permitted
on the two-lane highways, and the on-board safety monitoring system
would continually monitor the safety status of a vehicle, which will
include the sensing and collection of data on critical components such
as brakes, tires, and lights from which thresholds for warning and
countermeasures will be determined. Although the concept of these
technologies has been discussed, the authors are not aware of any
specific systems that have been developed. This, however, is not a
severe problem as less than 1% of single-unit truck crashes on two-
lane secondary roads were due to this causal factor.

For tractor-trailers, a crash most often occurred when the
interaction among the vehicle, driver, and environment was
incompatible and no evasive action was taken. It is likely that these
types of crashes were attributable to the incompatible physical and
operational characteristics of a large truck with the characteristics
of the two-lane secondary highway system, e.g., a long truck traveling
on a narrow lane of a roadway. Again, the authors are not aware of
any technologies available to mitigate the effect of this causal
factor. This, however, is not a serious setback as less than I% of
tractor-trailer crashes on two-lane secondary roads were due to this
causal factor.

Environment-Related Failures

Under environment-related failure crashes, both single-unit trucks
and tractor-trailers had the same most probable causal factors. For
all large trucks, a crash most often occurred when a faulty highway
component led to surface defects and the driver's evasive action
either failed or was not taken. The occurrence of these types of
crashes can be reduced considerably if drivers are made aware of the
defects in sufficient time before arriving at the defective location,
thereby providing the driver the opportunity to take the appropriate
evasive action in time. This may be achieved through a vision
enhancement system for crash avoidance that improves the driver's
ability to see the roadway and objects that are on or along the
roadway. If the driver is aware of the defects in the roadway, then
he or she will have more time to prepare for the situation and can
consider the most effective evasive action available before it becomes
necessary. One such

system is that manufactured by the Jaguar Corporation in cooperation
with Lucas Automotive. it is a computerized vision system that
provides an early warning collision avoidance system for drivers. It
uses a camera that converts the road scene into a computerized map,
identifying road edges, white line objects, and road defects. It
therefore has the ability to recognize the highway features and the
trajectory of the driven vehicle. Based on these, potential
difficulties are identified early and the driver can be warned to take
corrective action.

Technologies Based on Collision Types

As stated earlier, the predominant collision types for both single-
unit trucks and tractor-trailers are angle, rear end, SSOD, and SSSD.
These collision types accounted for over 75% of large-truck crashes on
two-lane secondary roads. The user service areas that would reduce
these types of crashes are longitudinal collision avoidance, lateral
collision avoidance, and intersection collision avoidance. A specific
technology that has been developed to reduce these types of collisions
is the vehicle on-board radar (VORAD) vehicle detection and driver
alert system manufactured by VORAD. It is a high-frequency radar
system that determines the speed and relative distance of objects from
the vehicle. Upon sensing a potential hazard, such as a vehicle
suddenly decelerating, VORAD emits a combination of lights and audible
warning tones to give drivers additional time to apply the brakes or
take evasive action. This system is currently in use.

Summary

Statistical Analysis

- Large trucks had significantly higher overall crash rates for
secondary roads than primary roads in the Richmond, Staunton,
and Northern Virginia districts.

- Large trucks had significantly higher statewide crash rates for
two-lane secondary roads than for two-lane primary roads.

- There was no significant difference between the fatal crash
rates of large trucks for two-lane secondary roads and two-lane
primary roads.

- Large trucks had significantly higher injury crash rates for
secondary roads than for primary roads in the Richmond and
Culpeper districts.

- The Richmond, Culpeper, and Northern Virginia districts had
significantly higher crash rates for large trucks than for
passenger cars for two-lane secondary roads.

- Large trucks had significantly higher crash rates statewide than
passenger cars for two-lane secondary roads.

- The crash rates for single-unit trucks were higher than the
rates for tractor-trailers for two-lane secondary roads,
although the difference was not significant.

- Large trucks did not have significantly higher fatal crash rates
than passenger cars for two-lane secondary roads.

- The predominant collision types of large-truck crashes on two-
lane secondary roads were angle, rear end, SSSD, and SSOD.
These predominant collision types accounted for over 75% of
large-truck crashes on two-lane secondary roads in Virginia.

- The proportions of angle, head-on, SSOD, and backed-into large-
truck collisions were significantly higher on two-lane secondary
roads than on two-lane primary roads.

- The large-truck proportions of SSSD, SSOD, noncollision,
miscellaneous, and backed-into collision types were
significantly higher than those for passenger cars on two-lane
secondary roads.

- Large trucks had a significantly higher fatal crash proportion
than passenger cars on two-lane secondary roads.

Fault Tree Analysis

- For single-unit trucks, a vehicle-related failure crash most
often occurred when there was brake failure and the evasive
action failed. Crashes in this category accounted for less than
I% of all single-unit truck crashes on secondary roads in
Virginia.

- For tractor-trailers, a vehicle-related failure crash most often
occurred when there was no observed vehicle defect but the
interaction among the driver, vehicle, and environment was
incompatible and no evasive action was taken. It is likely
these crashes were due to the characteristics of the large
trucks and the characteristics of the highway system. Crashes
in this category also accounted for less than I% of all tractor-
trailer crashes on two-lane secondary roads in Virginia.

- For all large trucks, an environment-related failure crash most
often occurred when a faulty highway component led to surface
defects and the driver's evasive action either failed or was not
taken. These crashes represented about 8% of all single-unit
truck crashes and about 6% of all tractor-trailer crashes on
two-lane secondary roads in Virginia.

- For all large trucks, a driver-related failure crash most often
occurred when there was an error in driver judgment and the
evasive action either failed or was not taken. These crashes
represented over 70% of all large-truck crashes on secondary
roads.

CONCLUSIONS

- The lack of adequate data on large-truck volumes by type on
secondary roads makes it essential that a significant effort be
made to obtain these data at regular intervals.

- The lack of data on the characteristics (width, trailer length,
tractor length) of large trucks involved in crashes on two-lane
secondary roads inhibits research in this area.

- Because SSSD, SSOD, and angle collisions are significant
problems for large trucks on two-lane secondary roads, it is
essential that a significant effort be made to reduce these
types of collisions.

- Because large-truck crash rates for two-lane secondary roads
were significantly higher than for two-lane primary roads in
Virginia, the effort placed in reducing large-truck crashes on
two-lane secondary roads must be at least the same as that put
on primary roads.

- Because driver error was the single highest causal factor of
large-truck crashes on two-lane secondary roads, significant
effort must be placed in eliminating or reducing large-truck
driver errors or mitigating the effect of such errors.

RECOMMENDATIONS

- Enforcement officers in Virginia should be better trained to
provide more details on crash and vehicle characteristics in
their crash reports to facilitate more in-depth analysis of
large-truck crashes.

- A statistical sampling system for collecting traffic data,
including vehicle classification on two-lane secondary roads,
should be developed and implemented as soon as possible.

- As it has been shown that the predominant causal factors of
large-truck crashes on two-lane secondary roads are driver
related and the predominant types of crashes are angle, rear
end, SSSD, and SSOD, a pilot study using an identified ITS
technology should be conducted to evaluate the effectiveness of
such a technology in reducing these types of large-truck crashes
on these roads. The most appropriate technology identified is
the VORAD system and should be used in the study. Such a test
should be conducted under the cooperative effort of VDOT and the
Department of Motor Vehicles in partnership with a selected
number of private large-truck companies, under the sponsorship
of the Federal Highway Administration and the National Highway
Traffic Safety Administration. Data should be collected for a 3
year period on the characteristics of crashes involving
different types of large trucks on two-lane highways in the
Richmond, Staunton, and Northern Virginia districts, which had
significantly higher crash rates for large trucks than for
passenger cars on the secondary

roads. Such a study should also evaluate the feasibility of
public/private partnership in the implementation of ITS
technologies.

REFERENCES

1. Transportation Research Board. 1986. Twin Trailer Trucks:
Effects on Highways and Highway Safety. Special Report No. 21 1.
Washington, D.C.

2. Highway Users Federation. 1988. Beyond Gridlock. Washington,
D.C.

3. Chirachavala, T. et al. 1984. Severity of Large-Truck and
Combination-Vehicle Crashes in Over-the-Road Service: A Discrete
Multivariate Analysis. Report No. TRR 975.
Washington, D.C.: Transportation Research Board.

4. Hedlund, J. 1977. The Severity of Large-Truck Crashes. DOT HS-
802-332. NHTSA Technical Note. Washington, D.C.: U.S.
Department of Transportation.

5. Black, K. 1992. Tractor-Trailer Crash Trends on Two-Lane Primary
Roads in Virginia. Bachelor of Science Thesis, University of
Virginia, Charlottesville.

6. Cleveland, D.E.; Kostynuik, L.P.; and Ting, K.L. 1985. Design
and Safety on Moderate Volume Two-Lane Roads. Report No. TRR
1026. Washington, D.C.: Transportation Research Board.

7. Transportation Research Board. 1989. Providing Access for Large
trucks. Special Report 223. Washington, D.C..

8. Khasnabis, S. 1986. Operational and Safety Problems of Trucks in
No-Passing Zones on Two-Lane Rural Highways. Report No. TRR
1052. Washington, D.C.: Transportation Research Board.

9. Virginia Department of Transportation, Department of Motor
Vehicles, Department of State Police, State Corporation
Commission. 1989. Size Weight Equipment and Other Requirements
for Trucks, Trailers and Towed Vehicles. Richmond.

10. Gericke, O.F., and Walton, C.M. 1981. Effect of Increased Truck
Size and Weight on Rural Highway Geometric Design (and Redesign)
Principles and Practices. Report No. TRR 806. Washington, D.C.:
Transportation Research Board.

11. Olsen, P.L. et al. 1984. Parameters Affecting Stopping Sight
Distance. NCHRP Report 270. Washington, D.C.: National Research
Council.

12. Harkey, D.L.; Zegeer, C.V.; Stewart, J.R.; and Reinfurt, D.W.
1991. Operational Impacts of Wider Trucks on Narrow Roadways.
Highway Safety Research Center, University of North Carolina.

13. Donaldson, G. 1986. Safety of Large Trucks and the Geometric
Design of Two-Lane, Two-Way Roads. Report No. TRR 1052.
Washington, D.C.: Transportation Research Board.

14. Garber, N.J., and Patel, S.T. 1992. The Effect of Trailer Width
and Length on Large-Truck Crashes. Charlottesville: Virginia
Transportation Research Council.

15. Fancher, P.S. 1989. Turner Truck Handling and Stability
Properties Affecting Safety. Washington, D.C.: Transportation
Research Board.

16. Virginia Department of Motor Vehicles. 1988 through 1990.
Virginia Traffic Crash Facts. Richmond.

17. Transportation Research Board. 199 1. Advanced Vehicle and
Highway Technologies. Special Report No. 232. Washington, D.C.

18. IVHS America. 1992. Strategic Plan for Intelligent Vehicle-
Highway Systems in the United States. Washington, D.C.

19. Birnbaum, Z.W.; Esary, J.D.; and Saunders, S.C. 1961.
Multicomponent Systems and Structures and Their Reliability.
Technometrics, 12(2): 55-57.

20. IVHS America and U.S. Department of Transportation. 1994. IVHS
Architectural Development Program: Interim Status Report.
Washington, D.C.

Appendix A
ADVANCED TECHNOLOGIES FOR LARGE-TRUCK SAFETY

Table A-1. Advanced Technologies for Large-Truck Safety

Crash Causal Factor User Service Advanced Technologies Status

Vehicle equipment Automatic roadside Automatic commercial Operation test in
failure and evasive safety inspection vehicle operation safety progress
action failed or was inspection system ADVANTAGE 1-75
not taken
On-board safety Vehicle monitoring Operational test
monitoring systems: Automated bus
diagnostic system,
Michigan Mass Transit
Authority

Inattentive Safety readiness Pattern recognition Proposed
driver error continuous driver
status/performance
monitoring system

Impairment Safety readiness Galvanic skin detectors Automobile manufacturers have the
technology

Steering wheel motion Proposed
pattern recognition
system

Roadway hazardous Vision enhancement Hazard-mounted beacons Proposed
conditions for crash avoidance
Infrared laser scanning Proposed
system

Machine-vision guidance Operational test
system by U.S. Army

Interaction between Longitudinal col- Headway detection Available;
vehicles lision avoidance, system, VORAD, lateral operational test
lateral collision encroachment warning
avoidance, and system, ultrasound
interaction backing sensors
collision avoidance

Appendix B
FAULT TREE FOR VEHICLE- AND ENVIRONMENT-RELATED CRASHES

Click HERE for graphic.

Figure B-1. Fault Tree for Vehicle-Related Crashes Involving
Single-Unit Trucks

Click HERE for graphic.

Figure B-2. Fault Tree for Vehicle-Related Crashes Involving
Tractor-Trailers

Click HERE for graphic.

Figure B-3. Fault Tree for Environment-Related Crashes
Involving Single-Unit Trucks

Click HERE for graphic.

Figure B-4. Fault Tree for Environment-Related Crashes
Involving Tractor-Trailers

Appendix C
MINIMUM CUT SETS AND THEIR ASSOCIATED PROBABILITIES

Table C-1. Minimum Cut Sets for Single-Unit Truck Crashes Due
to Vehicle-Related Failure

Path Probability

1-3-7 and 1-4-9-17 0.0244
1-3-7 and 14-9-18 0.0244
1-3-8 and 1-4-9-17 0.0250
1-3-8 and 1-4-9-18 0.0250
1-3-7 and 14-10-19 0.1051
1-3-7 and 14-10-20 0.1051
1-3-8 and 1-4-10-19** 0.1077
1-3-8 and 1-4-10-20** 0.1077
1-3-7 and 14-11-21 0.0582
1-3-7 and 14-11-22 0.0582
1-3-8 and 1-4-11-21 0.0596
1-3-8 and 1-4-11-22 0.0596
2-5-12 and 2-6-14 0.0590
2-5-12 and 2-6-15 0.0885
2-5-12 and 2-6-16 0.0173
2-5-13 and 2-6-14 0.0269
2-5-13 and 2-6-15 0.0403
2-5-13 and 2-6-16 0.0079

** indicates the minimum cut paths with the highest
probability and thus significant causal factors.

5 9

Table C-2. Minimum Cut Sets for Tractor-Trailer Crashes Due to
Vehicle-Related Failure

Path Probability

1-3-7 and 1-4-9-17 0.0210
1-3-7 and 1-4-9-18 0.0210
1-3-8 and 1-4-9-17 0.0215
1-3-8 and 1-4-9-18 0.0215
1-3-7 and 1-4-10-19 0.0408
1-3-7 and 1-4-10-20 0.0408
1-3-8 and 1-4-10-19 0.0417
1-3-8 and 14-10-20 0.0417
1-3-7 and 1-4-11-21 0.0618
1-3-7 and 1-4-11-22 0.0618
1-3-8 and 1-4-11-21 0.0633
1-3-8 and 1-4-11-22 0.0633
2-5-12 and 2-6-14** 0.1230
2-5-12 and 2-6-15** 0.1845
2-5-12 and 2-6-16 0.0361
2-5-13 and 2-6-14 0.0560
2-5-13 and 2-6-15 0.0840
2-5-13 and 2-6-16 0.0164

** indicates the minimum cut paths with the highest
probability and thus significant causal factors.

Table C-3. Minimum Cut Sets for Single-Unit Truck Crashes
Due to Environment-Related Failure

Path Probability

1-3-8 and 1-4** 0.29500
1-3-9 and 1-4** 0.29500
2-5-10-16 and 5-11-18 0.0128
2-5-10-17 and 5-11-18 0.0859
2-5-10-16 and 5-11-19 0.0128
2-5-10-17 and 5-11-19 0.0859
2-5-10-16 and 5-11-20 0.0132
2-5-10-17 and 5-11-20 0.0885
2-6-12-21 and 6-13-23 0.0044
2-6-12-22 and 6-13-23 0.0294
2-6-12-21 and 6-13-24 0.0044
2-6-12-22 and 6-13-24 0.0294
2-6-12-21 and 6-13-25 0.0045
2-6-12-22 and 6-13-25 0.0303
2-7-14-26 and 7-15-28 0.0004
2-7-14-27 and 7-15-28 0.0024
2-7-14-26 and 7-15-29 0.0004
2-7-14-27 and 7-15-29 0.0024
2-7-14-26 and 7-15-30 0.0004
2-7-14-27 and 7-15-30 0.0024

** indicates the minimum cut paths with the highest
probability and thus significant causal factors.

Table C-4. Minimum Cut Sets for Tractor-Trailer Crashes Due to
Environment-Related Failure

Path Probability

1-3-8 and 1-4** 0.2800
1-3-9 and 1-4** 0.2800
2-5-10-16 and 5-11-18 0.0143
2-5-10-17 and 5-11-18 0.0960
2-5-10-16 and 5-11-19 0.0143
2-5-10-17 and 5-11-19 0.0960
2-5-10-16 and 5-11-20 0.0148
2-5-10-17 and 5-11-20 0.0922
2-6-12-21 and 6-13-23 0.0045
2-6-12-22 and 6-13-23 0.0303
2-6-12-21 and 6-13-24 0.0045
2-6-12-22 and 6-13-24 0.0303
2-6-12-21 and 6-13-25 0.0047
2-6-12-22 and 6-13-25 0.0312

** indicates the minimum cut paths with the highest
probability and thus significant causal factors.

(435.html)