Data Design
Resume:
ASSESSING BENEFITS IN VEHICLE SPEED AND LATERAL POSITION
WHEN CHEVRONS WITH FULL RETROREFLECTIVE SIGN POSTS ARE
IMPLEMENTED ON RURAL HORIZONTAL CURVES
A Thesis
by
JONATHAN MICHAEL R
Submitted to the Office of Graduate Studies of
Texas A&M University
in partial fulfillment of the requirements for the degree of
MASTER OF SCIENCE
May 2009
Major Subject: Civil Engineering
ASSESSING BENEFITS IN VEHICLE SPEED AND LATERAL POSITION
WHEN CHEVRONS WITH FULL RETROREFLECTIVE SIGN POSTS ARE
IMPLEMENTED ON RURAL HORIZONTAL CURVES
A Thesis
by
JONATHAN MICHAEL R
Submitted to the Office of Graduate Studies of
Texas A&M University
in partial fulfillment of the requirements for the degree of
MASTER OF SCIENCE
Approved by:
Chair of Committee H. Gene Hawkins, Jr.
Committee Members, Yunlong Zhang
Eric Dumbaugh
Head of Department, David Rosowsky
May 2009
Major Subject: Civil Engineering
iii
ABSTRACT
Assessing Benefits in Vehicle Speed and Lateral Position when Chevrons with Full
Retroreflective Sign Posts are Implemented on Rural Horizontal Curves. (May 2009)
Jonathan Michael R, B.S., Michigan State University
Chair of Advisory Committee: Dr. H. Gene Hawkins, Jr.
Driving a horizontal roadway curve requires a change in vehicle alignment and a
potential reduction in speed. Curves may present a challenging situation during adverse
conditions or to inattentive drivers. Chevron signs provide advanced warning and
positive guidance throughout the curve. Some agencies place supplemental
retroreflective material on sign posts to enhance the signs conspicuity and visibility.
The objective of this study was to determine any incremental benefits in vehicle speed
and lateral lane position when retroreflective material was applied to Chevron sign posts
(ChevFull). This study analyzed three separate evaluation scenarios in a before, after,
and after-after experimental design. There was an existing Baseline evaluation with no
vertical delineation, a standard Chevron evaluation, and an experimental ChevFull
treatment evaluation. Data collection measured vehicle speed and lateral position data at
the point of curvature and mid-point on two separate curves. Findings showed that both
Chevrons and the ChevFull treatment moved vehicles away from oncoming traffic by
about 15 inches. Overall, there was little difference between the lateral position findings
of the two Chevron treatment scenarios. Chevrons achieved a 1.28 MPH reduction in
mean vehicle speed from the Baseline evaluation and the ChevFull treatment obtained a
2.20 MPH reduction. The findings determined that the benefits of the ChevFull
treatment were not substantial. The author recommends that the MUTCD should
continue to present the ChevFull treatment as an optional delineation tool. Based on this
research, the author does not recommend any changes to the MUTCD.
iv
DEDICATION
This thesis is dedicated to my parents for their unending love and support.
v
ACKNOWLEDGEMENTS
This thesis could not have been completed without the help and support from so
many exceptional people. I would like to thank my committee chair Dr. Hawkins for his
direction and guidance in my research endeavors. I am grateful to my committee
members Dr. Zhang and Dr. Dumbaugh for their time and patience.
I have worked with many gifted and knowledgeable colleagues at the Texas
Transportation Institute and I have thoroughly enjoyed my time at the institution. I
would especially like to thank Sue Chrysler, Beverly Kuhn, Keith Knapp, and Dillon
Funkhouser. It has been a pleasure and a privilege to have worked with them.
I would like to thank Kirk Barnes and Michael Jedlicka at the Texas Department
of Transportation. Implementing field treatments was made possible with their help and
cooperation. I am also appreciative of all the TxDOT staff that went out to the test sites
and put on a hardhat.
Studying at Texas A&M University has been an enriching experience. I am
grateful to all of my wonderful friends that I have met during my excursion down south.
Thank you for the memories and the experiences.
Finally, thanks to my family for their love and support.
vi
NOMENCLATURE
ANOVA Analysis of Variance
AASTHO American Association of State Highway and Transportation
Officials
ChevFull Chevrons with Full Length Retroreflective Material on Sign Post
FHWA Federal Highway Administration
FM Farm to Market Road
GIS Geographic Information System
HSD Honestly Significant Difference
ITE Institute of Transportation Engineers
K-S Kolmogorov-Smirnov Test
LED Light Emitting Diode
MANOVA Multivariate Analysis of Variance
MDOT Michigan Department of Transportation
MOE Measures of Effectiveness
MP Mid Point of Curve
MPH Miles per Hour
MUTCD Manual on Uniform Traffic Control Devices
PC Point of Curvature
PT Point of Tangent
PMD Post-Mounted Delineators
PMD Full PMD with Full Length Retroreflective Material
RPM Raised Pavement Markers
SUV Sports Utility Vehicle
TTI Texas Transportation Institute
TxDOT Texas Department of Transportation
vii
TABLE OF CONTENTS
Page
ABSTRACT iii
DEDICATION iv
ACKNOWLEDGEMENTS v
NOMENCLATURE vi
TABLE OF CONTENTS vii
LIST OF FIGURES x
LIST OF TABLES xi
CHAPTER
I INTRODUCTION 1
Problem Statement 3
Research Objectives 3
II BACKGROUND 6
Driver Error on Curves 6
Improper Speed Selection 6
Inadequate Lateral Position 8
Delineation Treatments 9
Delineation Research 10
Enhanced Delineation Treatments 12
Background Summary 14
III STUDY DESIGN 16
Study Approach 16
Experimental Design 16
Measures of Effectiveness 17
Site Selection 19
Site Selection Criteria 19
Site Selection Process 20
Test Site Characteristics 21
Delineation Treatment Application 23
Studied Delineation Treatments 23
Application and Installation 25
viii
CHAPTER Page
IV DATA COLLECTION AND ANALYSIS 27
Data Collection Procedure 27
Data Collection Equipment 27
Collection Locations 28
Data Collection Schedule 30
Data Processing and Screening 30
Free-Flowing Vehicles 31
Vehicle Type 31
Weather Analysis 32
Time Classification 32
Control Point Speed Assessment 33
Functional Data Formatting 34
Encroachment Rates 35
Vehicle Tracking Through Curve 35
Analysis Methods 36
Multivariate Analysis of Variance 36
Tukey s Honestly Significant Difference 37
Z-test 38
F-test 39
Normality of Data 39
V FINDINGS 41
Lateral Position 41
Site Findings 41
Location Findings 43
Lateral Position Tracking 45
Lateral Position Standard Deviation 46
Encroachment Rates 47
Lateral Position Summary 49
Speed 49
Site Findings 50
Location Findings 51
Speed Tracking 53
Speed Standard Deviation 54
High Speed Findings 55
Speed Summary 56
Findings Summary 57
ix
CHAPTER Page
VI CONCLUSIONS AND RECOMMENDATIONS 58
Chevron Treatment Conclusions 58
Treatment Application 58
Lateral Position Conclusions 59
Speed Conclusion 60
Recommendations 61
REFERENCES 63
APPENDIX A 69
VITA 72
x
LIST OF FIGURES
Page
Figure 1 Retroreflective Sign Post Examples (3) 2
Figure 2 PMD and Chevron Sign (1, 16) 10
Figure 3 Treatment Effect on MOE and Traffic Safety 18
Figure 4 Map of Test Sites 22
Figure 5 Wedge Anchor Post Assembly (29 24
Figure 6 Chevron Spacing on a Curve (16 25
Figure 7 ChevFull Treatment on Site 1 26
Figure 8 ChevFull Treatment on Site 2 26
Figure 9 Z-Configuration Sensor Layout (30) 28
Figure 10 Data Collection Location Diagram 29
Figure 11 FM 974 Curve Schematic 70
Figure 12 FM 50 Schematic 71
xi
LIST OF TABLES
Page
Table 1 Delineation Treatment Matrix 17
Table 2 Test Site Characteristics 23
Table 3 Percentage of Heavy Vehicles 31
Table 4 Control Point Speed Assessment 34
Table 5 Overall Sample Size Summary 40
Table 6 MANOVA Lateral Position Results 42
Table 7 Tukey's HSD Mean Lateral Position Results 43
Table 8 Location MANOVA Lateral Position Results 43
Table 9 Location Tukey's HSD Mean Lateral Position Results 44
Table 10 Nighttime and Daytime Mean Lateral Position Results 45
Table 11 Tukey's HSD Lateral Position Tracking Results 46
Table 12 Lateral Position Standard Deviations 47
Table 13 Encroachment Percentages 48
Table 14 MANOVA Speed Results 50
Table 15 Tukey's HSD Mean Speed Results 51
Table 16 Location MANOVA Speed Results 52
Table 17 Location Tukey's HSD Mean Speed Results 52
Table 18 Nighttime and Daytime Mean Speed Position Results 53
Table 19 Tracking Tukey's HSD Mean Speed 54
Table 20 Speed Standard Deviation 54
Table 21 High Vehicle Curve Speeds 56
1
CHAPTER I
INTRODUCTION
A horizontal curve requires a change in vehicle path alignment and a potential
reduction in vehicle speed. The change from tangent alignment may present a
challenging task during adverse driving conditions or to inattentive drivers. Delineation
devices and horizontal curve treatments aid and assist drivers in safe and efficient
horizontal curve negotiation. Delineation treatments provide advanced warning on the
approach tangent and positive guidance throughout the curve. Chevron signs are a
common type of delineation treatment and are widely utilized. Chevron signs are
classified as a warning sign (1) and are placed on the outside of a curve.
Some agencies have been placing supplemental retroreflective material on the
Chevron sign posts to enhance the conspicuity and visibility of the sign. Figure 1
illustrates examples of current uses for the retroreflective material on warning and
regulatory sign posts. The retroreflective material is applied with either adhesive
backing or attached on a flat panel. There are commercial venders that sell such
treatments, which are marketed as Sign Post Covers or Reflective Panels.
The practice of placing supplemental retroreflective material on sign posts
became common at passive at-grade railroad crossings. As of 1990, many states began
placing a strip of retroreflective material on the front and back of Crossbuck sign posts
when there was not an automatic gate that notified drivers of an approaching train (2).
The retroreflective material is intended to alert drivers of the critical crossing situation
and help the drivers to detect the presence of a crossing train. The Federal Highway
Administration (FHWA) recommended this treatment for all passive at-grade rail
crossings (1) and the practice of placing retroreflective material on sign posts spread to
other warning and regulatory sign applications.
This thesis follows the style of the Transportation Research Record.
2
Figure 1 Retroreflective Sign Post Examples (3)
The only standards that govern the application of retroreflective material on sign
posts are contained in the 2003 edition of Manual on Uniform Traffic Control Devices
(MUTCD). The MUTCD states in section 2A.21 that Where engineering judgment
indicates a need to draw attention to the sign during nighttime conditions, a strip of
retroreflective material may be used on regulatory and warning sign supports (1). The
MUTCD specifies that the retroreflective material shall be at least 2 inches in width and
shall extend from the bottom of the sign to 2 feet above the roadway surface (1).
Agencies may utilize the additional retroreflective material as an optional treatment and
it is not required by the MUTCD.
The specifications in section 2A.21 covering retroreflective sign posts first
appeared in the 2003 edition of the MUTCD. A Notice of Proposed Amendment on
May 21, 2003 discussed the new specifications, but it did not provide justification for the
added option nor did the final rule provide a research basis for the benefits of using
retroreflective post treatments (4). The author believes that the standard was added
without extensive support or rigorous testing.
Since the addition in the 2003 MUTCD, agencies have been placing
retroreflective material on Chevron sign posts (ChevFull) at select locations as an
additional treatment to the standard Chevron sign. The ChevFull treatment is intended
3
to increase the visibility and conspicuity of the Chevron sign. Curve negotiation and
driver safety may improve as a result of earlier detection and enhanced guidance. The
ChevFull treatment is relatively inexpensive, easy to install, and requires no maintenance
cost. The additional retroreflective material may be an attractive option due to its
simplicity, low-cost, and practicality.
PROBLEM STATEMENT
There has been a recent study that analyzed the change in vehicle speed when the
ChevFull treatment was implemented at one curve (5). This thesis evaluated both
vehicle speed and lateral lane position at two curves to determine the effects of the
ChevFull treatment. If the current practice of applying retroreflective material to
Chevron sign posts is going to continue, then it is a worthy endeavor to investigate the
treatment in a more comprehensive study to ascertain if there are additional benefits in
both speed and lateral position.
RESEARCH OBJECTIVES
The objective of this thesis was compare the effects of Chevrons and the
ChevFull treatment to a Baseline condition with no treatment in a before and after
experimental design and to determine if the ChevFull treatment achieved additional
benefits to the Baseline and Chevron evaluations. The research data for this thesis came
from a Texas Transportation Institute (TTI) study that was conducted between the fall of
2007 and summer of 2008. The TTI study analyzed multiple delineation treatments in a
closed-course test track portion, a laptop survey, and an open-road field evaluation. This
thesis focused specifically on the Chevron and the ChevFull treatment results from the
field evaluation of the TTI study.
The objective was accomplished by analyzing three separate evaluation scenarios
in a before and after experimental design. A before, an after, and an after-after design
was used to isolate the specific effects of the treatments. Evaluation scenarios consist of
an existing Baseline evaluation (before), a standard Chevron evaluation (after), and an
4
experimental ChevFull treatment evaluation (after). The Baseline evaluation employed
no existing vertical delineation, such as Chevron signs and Post-Mounted Delineators
(PMD). The analysis compared both Chevron treatment evaluations to the Baseline
evaluation to identify any changes in vehicle operations. The results from the ChevFull
treatment were compared to the Chevron results to determine if the added retroreflective
material achieved additional benefits.
This thesis evaluated the treatments at two test curves. Vehicle speed and lateral
position was measured at each curve in the Baseline evaluation. The Texas Department
of Transportation (TxDOT) then installed Chevron signs on both test curves. Site 1 had
the ChevFull treatment and Site 2 employed standard Chevron signs. The first after
analysis repeated the data collection process in an identical manner to the Baseline
evaluation. Afterwards, researchers removed the ChevFull treatment from Site 1 and
placed it on Site 2. The after-after evaluation completed the final data collection
scenario.
Measures of Effectiveness (MOE) assessed the change in vehicle operations
between the three evaluation scenarios. A background review of past literature helped to
identify appropriate MOE for assessing the benefits of Chevron signs and the ChevFull
treatment. Analyzed MOE for both speed and lateral position included the mean, the
standard deviation, and the change in individual vehicle data from the PC to the MP.
Line lane encroachments and high speed percentages were also assessed.
The data collection process obtained vehicle speed and lateral position data at the
Point of Curvature (PC) and at the Mid Point (MP) on both curve approaches of each
site. Roadway sensors recorded both vehicle speed and lateral position for around 4 to 7
days during each evaluation scenario. A control speed was measured approximately one
mile upstream from each curve approach to determine if vehicle speeds considerably
changed between evaluation scenarios. A screening and formatting process transformed
the raw data into working vehicle data for the statistical analysis.
5
Statistical techniques analyzed and determined if the change in MOE amongst
the three evaluation scenarios were significantly different. The general testing
hypothesis stated that if the treatments did achieve a significant difference, then the null
hypothesis was rejected and the alternative hypothesis was accepted. The statistical
analysis employed the Multivariate Analysis of Variance (MANOVA), Tukey s
Honestly Significant Difference (HSD), the Z-test, and the F-test to test for significance.
The statistical analysis performed all tests at a confidence interval of 95 percent or
higher.
The author extracted meaningful trends and findings from the statistical analysis.
The recommendations addressed the benefits for both Chevrons and the ChevFull
treatment over the Baseline Evaluation. This study determined if the ChevFull treatment
did or did not achieve significant and substantial benefits over standard Chevrons. In
summation, the need for changes in language or treatment practices in the MUTCD was
discussed.
6
CHAPTER II
BACKGROUND
This chapter contains a background review of past studies and practices.
Previous research served as a guide and indicated how this study could contribute
knowledge to the current transportation practice. This background review started
general and then narrowed the focus to establish suitable methods for investigating the
ChevFull treatment effects, analyzing the data, and interrupting the results.
A driver must guide his or her vehicle safely through a horizontal curve. Vehicle
guidance involves maintaining proper speed and lane placement that does not conflict
with roadway constraints or regulations. Vehicle guidance is one of the fundamental
tasks in Alexander and Lunenfeld s positive guidance framework. Positive guidance
tasks include vehicle control, guidance, and navigation (6). Negotiating a horizontal
curve involves all three driving tasks. Driver error on a horizontal curve is typically a
result of a breakdown in one of the positive guidance tasks.
DRIVER ERROR ON CURVES
Driver error in vehicle guidance is typically attributed to improper vehicle speed
or lateral lane position selection. Driver error and inadequate vehicle guidance may
increase the chance of a potential hazard.
Improper Speed Selection
Appropriate curve speed is critical for safe vehicle guidance. In a fundamental
study, Solomon identified several significant relationships between speed and safety on
rural roadways (7). The study examined data from 10,000 crashes before the year of
1964. One of the main discoveries revealed that crash rates were significantly higher for
vehicles traveling at speeds that were considerably above or below the roadway mean
speed. The relationship between vehicle speed and crash rates resembled a U-shape
curve. Crash rates were lowest for vehicles traveling near the roadway mean speed and
7
highest when there was a large disparity between the vehicle speed and the roadway
mean speed (7). The study concluded that variance in speed and speed differential were
significant factors that increased the likelihood of a crash.
A study by Nicholas and Ehrhart reconfirmed Solomon s variance in speed and
crash relationship (8). The study evaluated 15 two-lane rural highways in Virginia
between the years of 1993 and 1995. The researchers created a model to determine if the
mean speed, speed standard deviation, flow per lane, lane width, or shoulder width were
significant contributors to increased crash rates. The results from the model showed that
speed standard deviation had the greatest influence on crash rates (8). It was determined
that crash rates increased exponentially as the standard deviation of travel speed
increased. A study in a different part of the county revealed more relationships between
speed and crash rates.
At comprehensive crash investigation by the Michigan Department of
Transportation (MDOT) explored crash rates on horizontal curves and speed
characteristics (9). The investigation involved an extensive literature review, an
examination of crash data, and a field evaluation of six rural horizontal curves. The
most reoccurring speed characteristic that was associated with crash rates was speed
differential between the tangent and the curve speed (9). The data showed that as the
speed differential increased, then so did the crash rates. For instance, crash rates were
higher at a curve that required drivers to reduce vehicle speed by 15 MPH, as opposed to
5 MPH. The investigation determined that increased speed differential was strongly
correlated to both head-on and single-vehicle crashes (9).
Anderson and Krammes further built upon MDOT s speed differential and crash
rate relationship (10). The researchers developed a model that quantified and illustrated
the relationship. The model incorporated speed differential and geometric characteristics
from 1,126 rural horizontal curves. A linear regression line plotted the relationship
between speed differential and crash rates (10). The regression line showed that crash
rates were significantly higher on a curve with a 20 MPH speed differential, as opposed
to a curve that required a 10 MPH speed reduction. Liner relationships exhibited R2
8
values greater than 0.90 and statistical analysis proved that speed differential was a
significant contributor to increased crash rates (10).
Speed alone is not the only cause or contributor to driver error on a horizontal
curve. Vehicle speed and lateral position are related and it is typically a breakdown in
both that leads to driver error.
Inadequate Lateral Position
Speed selection greatly influences a vehicle s lateral lane position within a
horizontal curve. Centripetal force pushes a vehicle to the outside of a curve when the
operating speed exceeds the curve design speed. Moving outwards will increase a
vehicle s radius path to compensate for the excessive speed. Adopting a larger radius
than the road s intended design radius is called curve flattening. Zador et al. revealed in
study that the curve flattening was common at 46 rural horizontal curves in two states
(11). Researchers collected vehicle speed and lateral lane position at several points
along each of the horizontal curves. The results showed that many vehicles shifted
towards the edgeline on an outside curve (left-handed curve) and closer to the centerline
on an inside curve (right-handed curve).
Spacek determined that curve flattening was more prevalent at curves with large
speed differential between the tangent and curve speed (12). The study monitored
vehicle speed and lateral position at twelve points within a horizontal curve. The
researcher also identified another inadequate vehicle path, which was referred to as
curve cutting. Vehicles shifted towards the center of a curve during curve cutting. The
researcher observed that 37 percent of the total vehicles in the study displayed an
inadequate vehicle path (12). Spacek determined that the curves with the high rates of
improper vehicle paths also exhibited high crash rates (12). The study concluded that
abruptly overcorrecting for poor lane position was a significant contributor in horizontal
curve crashes.
Besides a specific vehicle path, a study in Pennsylvania established a relationship
between the lateral position standard deviation and crash rates (13). Taylor et al.
9
evaluated nine rural two-lane curves that exhibited high crash rates. The nine curves
varied in crash rates, traffic volumes, geometric characteristics, and driver types. Lateral
position data were collected at the PC and at the MP on both directional approaches at
each curve. The lateral position mean and standard deviation were generated for each
data collection location. The standard deviation indicated the variation in lane position
amongst the sampled vehicles. A model determined that crash rates were associated
with high lateral position standard deviation values (13).
One indication of improper vehicle paths is lane line encroachments. The
comprehensive MDOT crash investigation also examined lateral position data and crash
rates (9). Along with speed differential, the researchers established that lane line
encroachments were also a significant contributor to crash rates (9). The relationship
determined that crash rates increased when total lane line encroachments increased. The
correlation was very strong for single-vehicle crashes and edgeline encroachments.
DELINEATION TREATMENTS
Delineation devices are placed on a horizontal curve to curtail improper speed
and lateral position by providing advanced warning and guidance. Delineation devices
include Raised Pavement Markers (RPM), barrier reflectors, PMD, and Chevron Signs.
Implementation is based upon roadway geometry, speed differential, sight distance, and
crash history (1, 14).
Curve warning and guidance is achieved through the delineation devices size,
color contrast, and retroreflectivity (1). The Retroreflectivity of a delineation device is
critical during nighttime or adverse driving conditions. Retroreflection is the physical
principle of returning light back to its source (15). Light from a vehicle s headlight is
redirected back to the driver by means of a retroreflective device. The MUTCD states
that delineation shall be retroreflective devices mounted above the roadway surface and
along the side of the roadway in a series to indicate the alignment of the roadway (1).
Two commonly utilized horizontal curve delineation devices are Chevron signs
and PMD. Figure 2 depicts an image of both delineation treatments. PMD are
10
approximately 4 foot in length and 4 inches in width with retroreflective material applied
at the top of the post. The Chevron signs (W1-8) are classified as a warning sign in the
MUTCD and are comprised of a pointed black arrow on a yellow background that
indicates the direction of the roadway. Both devices are placed on the outside curve
shoulder. Chevrons should be placed so that at least two signs are in the drivers view
throughout the curve (1, 16).
Figure 2 PMD and Chevron Sign (1, 16)
DELINEATION RESEARCH
In 1983 Niessner summarized several field studies that evaluated PMD and
Chevrons impacts on vehicle safety and crash rates (17). The summary analyzed results
from eight different states. All of the reviewed studied were conducted in a before and
after experiment design. Each study evaluated crash rates before and after the
installation of delineation treatments. Niessner extracted from the studies that Chevrons
significantly reduced the fatal crash rate and PMD significantly lowered run-off-the road
11
crashes (17). The study concluded that both Chevron and PMD were adequate devices
for delineating horizontal rural curves.
The reduced crash rates may be attributed to improved vehicle operations, which
were shown in an Australian study conducted in 1983 (18). Johnston evaluated the
benefits of Chevrons and PMD in a closed-course test track. Delineation treatments
were assessed by measuring vehicle lateral position, encroachment rates, and speed. The
results showed that the curves without delineation treatments exhibited the least
desirable vehicle operations (18). Curves employing Chevron signs achieved
significantly lower vehicle speed during nighttime and superior lateral position results
compared to curves with PMD. It was found that Chevrons moderately increased the
mean speed during daytime conditions. Johnston attributed the small speed increase to
enhanced driver confidence and comfort (18). Nevertheless, mean speeds were
significantly lower and vehicles followed a better path on curves with Chevrons, as
opposed to PMD.
A study in Virginia conducted an open-road field evaluation that was similar to
Johnston s study (19). Jennings and Demetsky compared the effects of Chevrons, PMD,
and a road edge delineator on horizontal curves. Each treatment was placed individually
on five curves and vehicle speed and lateral position data were collected at the PC and
MP. Results determined that none of the treatments achieved a significant reduction in
speed, but benefits were obtained in lateral position (19). All treatments shifted drivers
away from the edgeline on an outside curve. The researchers concluded that Chevrons
promoted a more centralized vehicle path, reduced encroachment rates, and lowered the
lateral position variance (19).
A study by Agent and Creasey evaluated Chevrons and PMD in a slightly
different approach (20). The researchers studied the delineation treatments in two parts:
a subjective laboratory evaluation and a field evaluation. In the laboratory evaluation,
forty subjects were shown curve photographs with PMD and Chevrons varied in spacing,
offset, and height. The researchers found that curves were perceived sharper when
delineation treatments were taller (20). The second part of the study evaluated PMD and
12
Chevrons at increased heights on the open-road. RPM and pavement markings were
also evaluated in the field investigation. The researchers measured vehicle speed and
lane line encroachments as treatment MOE. The field investigation concluded that
Chevrons achieved a greater reduction in vehicle speeds and lowered centerline
encroachment rates then did PMD (20).
In 1987, a study by Zador et al. evaluated the short and long-term effects of
Chevrons, PMD, and RPM. The study analyzed on vehicle operations at 51 rural curves
in Georgia and New Mexico (11). Speed and lateral position were measured in a before
and after experimental design for short-term and long-term effects. The speed results
showed that PMD and RPM generally increased vehicle speeds by 1 to 3 feet per second
(11). Chevrons did not produce a significant change in vehicle speed in Georgia, but
increased vehicle speed by approximately 3 feet per second in New Mexico. Chevrons
shifted vehicles away from the centerline in both curve directions and PMD moved them
closer to the centerline. Neither treatment significantly corrected the curve cutting. The
researchers concluded that the data suggested that short-term changes did not erode over
time (11). In the end, the researchers could not definitively support one delineation
treatment over the other.
ENHANCED DELINEATION TREATMENTS
Most of the reviewed literature dealt with conventional or standard delineation
devices. Enhanced or modified delineation devices may be beneficial in certain
situations. A study on peripheral visual detection concluded that where there is a need
for early detection, the reflectivity of the target should be increased to assure timely
recognition, information processing, decision making, and appropriate control actions
(21). Another study that assessed roadway delineation for older drivers also determined
that enhanced delineation treatments should be considered in areas with a large
population of older drivers or at roadway locations with sharp horizontal curves (22).
A study by Pietrucha et al. in 1996 investigated older driver curve perception
when standard and enhanced delineation treatments were implemented on horizontal
13
curves (23). The objective of the study was to identify effective delineation treatments
that increased perception distance and heighten awareness for older drivers. The
researchers initially formulated 25 different delineation combinations. The treatments
consisted of pavement markings, RPM, Chevrons, standard PMD, PMD with fully
retroreflective post, and T-post PMD (23). The T-post PMD was an experimental
treatment that employed a thin strip of retroreflective material that ran the length of the
post from the standard PMD material to the bottom of the device. The PMD with fully
retroreflective post was also an enhanced experimental treatment. There were 45
subjects and each was placed in one of three age categories; youth, middle-age, and
older drivers.
The first portion of the older driver study evaluated each treatment combination
in a driving simulation by measuring the rate of deceleration on the upstream curve
approach (23). Subjects were also asked to subjectively rank the advanced warning
ability of each treatment combination. Treatment combinations that included Chevrons,
PMD, and T-post PMD achieved earlier deceleration than curves with just pavement
markings or RPM. Chevrons and the T-post PMD were subjectively ranked high by all
age groups (23). The second portion of the study assessed the 12 most promising
treatment combinations by evaluating curve perception distance, cost, and ease of
implementation. The treatment combinations that provided the longest perception
distance consisted of the T-post PMD and Chevrons (23). The effective treatments
exhibited large retroreflective targets and retroreflective material that extended from the
top of the device to the ground.
The ChevFull treatment was first assessed in a study conducted in 2003 at TTI
(5). Gates et al. evaluated a 4 inch wide strip of fluorescent yellow prismatic sheeting
that extended the entire length of a Chevron sign post. The treatment was placed on all
Chevron signs at one rural horizontal curve. Vehicle speed was measured at two
upstream tangent points, the PC, and the MP. Overall the ChevFull treatment achieved a
slight speed decrease of 1.7 MPH during twilight and 1.6 MPH during nighttime (5).
The researchers concluded that the use of fluorescent yellow microprismatic materials
14
on Chevron posts or other curve delineation is recommended on an as-needed basis at
spot locations where additional delineation is desired (5).
The parent study to this thesis evaluated the ChevFull treatment in a closed-
course test track setting (24). Chrysler et al. evaluated five different treatments on four
test track curves. The treatments consisted of a Baseline condition with no vertical
delineation, PMD with standard retroreflective material, PMD with full length
retroreflective material (PMD Full), Chevrons, and the ChevFull treatment. Twenty
subjects drove ten laps on the track and saw each treatment at each curve in both
directions. An instrumented vehicle measured foot pedal displacement, lateral
acceleration, specific Global Positioning System (GPS) location, and vehicle speed.
The PMD Full and the ChevFull treatments showed the most promising results
and the least desirable vehicle operations were observed for the Baseline condition (24).
Subjects were able to detect curves at a greater distance when PMD Full and the
ChevFull treatments were implemented on curves (24). Subjects also released the
acceleration pedal and initiated the brake pedal earlier when PMD Full and ChevFull
treatments were present (24). It was reasoned that the enhanced delineation achieved a
greater detection distance, which allowed drivers to decelerate earlier and minimize
lateral acceleration on the vehicle.
BACKGROUND SUMMARY
This background review established that drivers must maintain proper vehicle
guidance when traversing a horizontal curve. A driver must select an adequate speed
and sustain a lateral lane position that complies with the roadway environment and
geometrics. Safety issues may occur if curve speed exceeds the roadway design speed or
if lateral position deviates considerably from a centralized lane position.
Appropriate vehicle speed is critical for safe curve negotiation. Solomon showed
that crash rates significantly increased when the vehicle speed greatly exceeded the
mean roadway speed (7). The background review also determined that speed standard
deviation and the speed differential were significant contributors to crash rates (7, 8, 9).
15
Promoting more uniform speeds and lowering excessive vehicle speed were deemed to
be beneficial safety measurements. Improving vehicle speed on a curve may also be
advantageous for lateral lane position and vehicle path.
Excessive speed may force a vehicle to the outside of the curve requiring the
driver to adopt a curve flattening strategy to minimize the centrifugal force. Curve
flattening was associated with high crash rates (12). Curtailing excessive curve speed
may mitigate curve flattening and reduce the chance of single-vehicle or head-on
crashes. Another improper vehicle path that was linked to crash rates was curve cutting
where the driver will shift towards the inside of the curve. Both improper curve
flattening and curve cutting may be negated with lowered lateral position standard
deviation values and reduced lane line encroachment rates (12, 13). It is ideal to achieve
a more uniform and centralized lane position at the PC and at the MP.
Past research showed that Chevron signs have achieved beneficial vehicle
operations (18, 19, 20) and reduced crash rates on horizontal curves (17). Some
researchers recommended placing enhanced delineation treatments at locations where
early curve detection is critical or where there are large populations of older drivers (21,
22). Placing retroreflective material on sign posts or on the entire length of the PMD has
shown great promise in past studies (5, 23, 24). These past studies focused on
performance measures of curve detection distance and vehicle speed. Two of the studies
were conducted at a close-course test track (23, 24) and one evaluated enhanced
treatments on a single roadway curve (5). This background review determined that there
is a need to assess the effects of the ChevFull treatment on both vehicle speed and lateral
position in an open-road study with more than one test curve.
16
CHAPTER III
STUDY DESIGN
This chapter documents the study design and methods utilized for evaluating the
effectiveness of the Chevron and the ChevFull treatments on rural horizontal curves.
The study design provides the foundation for the data collection and analysis procedures.
This chapter documents the study approach, the site selection, and delineation treatment
application.
STUDY APPROACH
The study approach details the fundamental structure for the treatment
evaluation. It indicates how treatments were assessed and what were the specific criteria
used to determine any incremental benefits that were associated with both Chevron
treatments.
Experimental Design
This thesis measured vehicle operations in a before and after experimental
design, which identified changes in vehicle speed and lateral position that could be
attributed to the Chevron treatments. The study design consisted of three separate
evaluation scenarios: a before, after, and after-after. The before scenario was an existing
Baseline evaluation with no vertical delineation treatment. There were two treatment
scenarios that consisted of a standard Chevron evaluation (after) and an experimental
ChevFull treatment evaluation (after). Researchers collected vehicle speed and lateral
position data at a test site before the addition of a study treatment in the Baseline
evaluation. After the Baseline evaluation, Chevron treatments were installed and vehicle
data were collected at the same site in an identical manner. Researchers switched the
Chevron treatments and the after-after evaluation was conducted. The comparison of
vehicle speed and lateral position data between the three evaluation scenarios determined
the effects and value of the experimental treatment. The Institute of Transportation
17
Engineers (ITE) Manual of Transportation Engineering Studies acknowledged that
before and after experiments are effective and practical for eliminating site-to-site
comparisons, reducing the number of sites, and are easily comprehended by engineers
and non-technical readers (25). Table 1 displays the treatment matrix.
Table 1 Delineation Treatment Matrix
Selected Sites Before After After - After
Site 1 Baseline ChevFull Chevrons
Site 2 Baseline Chevrons ChevFull
Measures of Effectiveness
Safety benefits can be directly observed with a reduction in crash rates, but in
some cases sufficient crash data may not always be accessible. Through years of
research, studies have been able to identify surrogates for crashes. Safety surrogate
measures establish a relationship between vehicle operations and crashes rates.
Surrogates are an accepted intermediate in lieu of the absence or lack of sufficient crash
data, but they are not a substitute (26).
The background literature review identified suitable MOE that were associated
with safety surrogate measures. MOE define the vehicle operations for a given scenario.
A comparison between the MOE of two different scenarios reveals the change in vehicle
operations. The general testing hypothesis states that if there is relationship between the
ChevFull treatment and a beneficial change in MOE, then it is possible to associate the
treatment with traffic safety. Figure 3 illustrates the logic and reasoning behind the
general hypothesis.
18
Reduce the
Enhance Improve
Install the
Probability of
ChevFull Detection and Vehicle
a Crash
Treatment Guidance Operations
Experimental Goal
Effect Change in
Treatment MOE
Figure 3 Treatment Effect on MOE and Traffic Safety
MOE included both longitudinal components (speed) and lateral components
(lateral lane position). MOE in this thesis were:
mean lateral position,
mean change in lateral position from the PC to the MP,
lateral position standard deviation,
lane line encroachment rates,
mean speed,
mean change in speed from the PC and the MP,
speed standard deviation, and
high speed percentages.
Lateral Position Measures of Effectiveness
Justification for the lateral position MOE was derived from the background
review. Previous studies determined that high crash rates were associated with
overcorrecting improper lateral position due to the curve flatting or curve cutting (12).
Both incorrect vehicle paths involved vehicles deviating from a centralized path and
moving close to or encroaching onto a line lane. Line lane encroachments (9) and large
variation in lateral position (13) also led to higher crash rates. Achieving a more
centralized and uniform lateral lane position throughout the curve would reduce
improper vehicle paths, line lane encroachments, and lateral position standard deviation.
19
Therefore, a reduction in improper lateral position characteristics may reduce the
likelihood of a crash and ultimately improve safety.
Speed Measures of Effectiveness
Previous research also validated speed MOE. Studies acknowledged that high
speeds increased the probability of crashes, such as single-vehicle crashes (7, 9). Large
disparity between vehicle speed and the roadway mean speed significantly contributed to
higher crash rates (7, 8). Specifically for a horizontal curve, crash rates were shown to
decrease when the tangent speed on the upstream approach was closer to the curve
negotiation speed (9, 10). Promoting more uniform curve speed close to the appropriate
advisory curve speed may reduce the probability of a crash.
SITE SELECTION
The site selection portion of this thesis involved a great deal of effort and focus.
The TTI research project had the resources and time to assess the Chevron treatments on
two rural horizontal curves. It was highly important that both selected test sites were
ideal and satisfactory.
Site Selection Criteria
TxDOT and TTI staff assisted in creating a preliminary list of potential
horizontal curve sites. Potential site criteria stated that:
the roadway shall be classified as a high-speed rural highway with a posted
tangent speed of 55 MPH or greater,
curves should warrant a reduction in speed from the posted speed limit,
curves shall be located on the TxDOT roadway system, and
curves should yield volumes of approximately 1,000 or more vehicles per
day.
20
The analysis identified 170 potential curves near Bryan, Texas. Researchers resided
within the Bryan area and local sites minimized travel time and conserved resources.
Local agencies provided roadway information and curves were plotted on a
comprehensive regional map. TTI personnel visited each potential site and digitally
filmed the curve for later evaluation. Geometric characteristics, traffic control devices,
roadway features, and other relevant information were recorded in a spreadsheet for each
curve. The author generated a list of site selection criteria to systematically eliminate
any curves that were not ideal. Site selection criteria were that chosen curves:
shall have edgeline, centerline, and a total travel width greater than 20 feet,
shall have Curve Warning signs (W1-1 or W1-2) and Advisory Speed
plaques (W13-1),
shall have the same posted tangent speed limit and advisory curve speed on
both directional approaches,
should have minimal interference from intersecting roadways or driveways in
the immediate area,
should all exhibit si
Contact this candidate