U.S. Department
of Transportation
National Highway
Traffic Safety
Administration
DOT HS 809 561
March 2003
NHTSA Light Vehicle Antilock Brake System
Research Program Task 5.2/5.3:
Test Track Examination of Drivers' Collision
Avoidance Behavior Using Conventional and
Antilock Brakes
This document is available to the public from the National Technical Information Service, Springfield,
Virginia 22161.
DISCLAIMER
This publication is distributed by the U. S. Department of
Transportation, National Highway Traffic Safety Administration, in the
interest of information exchange.
The opinions, findings, and
conclusions expressed in this publication are those of the author(s) and
not necessarily those of the Department of Transportation or the National
Highway Traffic Safety Administration. The United States Government
assumes no liability for its contents or use thereof.
If trade or
manufacturers’ names or products are mentioned, it is because they are
considered essential to the object of the publication and should not be
construed as an endorsement. The United States Government does not
endorse products or manufacturers.
ii
NOTE
REGARDING COMPLIANCE WITH
AMERICANS WITH DISABILITIES ACT SECTION 508
For the convenience of visually impaired readers of this report using
text-to-speech software, additional descriptive text has been provided for
graphical images contained in this report to satisfy Section 508 of the
Americans With Disabilities Act (ADA). This descriptive text can be
found both within the body of the document and as alternate text.
iii
Technical Report Documentation Page
1. Report No.
2. Government Accession No.
3. Recipient's Catalog No.
DOT HS 809 561
4. Title and Subtitle
5. Report Date
NHTSA Light Vehicle Antilock Brake System Research Program Task 5.2/5.3: Test
Track Examination of Drivers' Collision Avoidance Behavior Using Conventional and
Antilock Brakes
March 2003
6. Performing Organization Code
7. Author(s)
8. Performing Organization Report No.
NHTSA/NRD-22
Elizabeth N. Mazzae, Frank S. Barickman, Garrick Forkenbrock, NHTSA
G. H. Scott Baldwin, Transportation Research Center Inc.
9. Performing Organization Name and Address
10. Work Unit No. (TRAIS)
National Highway Traffic Safety Administration
Vehicle Research and Test Center
P.O. Box 37
East Liberty, OH 43319
11. Contract or Grant No.
12. Sponsoring Agency Name and Address
13. Type of Report and Period Covered
National Highway Traffic Safety Administration
400 Seventh Street, S.W.
Washington, D.C. 20590
Final Report
14. Sponsoring Agency Code
15. Supplementary Notes
The authors acknowledge the support of this effort by W. Riley Garrott of NHTSA VRTC, Mark Flick of RAI, and Heath Albrecht,
Adam Andrella, Jan Cooper, David Dashner, Mark Gleckler, Lyle Heberling, Larry Jolliff, Ed Parmer, Jim Preston, and Judy
Weiser of TRC
16. Abstract
Numerous crash data statistical analyses conducted over the past few years suggest that, for automobiles, the introduction of fourwheel antilock brake systems (ABS) has produced net safety benefits much lower than originally expected. The studies indicate the
apparent increase in single-vehicle crashes involving passenger cars equipped with four-wheel ABS almost completely offsets the
safety advantage such vehicles have over their conventionally-braked counterparts. The National Highway Traffic Safety
Administration (NHTSA) has developed its Light Vehicle Antilock Brake Systems (ABS) Research Program in an effort to
determine the cause(s) of the apparent increase in fatal single-vehicle run-off-road crashes as vehicles undergo a transition from
conventional brakes to ABS. As part of this program, NHTSA conducted research examining driver crash avoidance behavior and
the effects of ABS on drivers' ability to avoid a collision in a crash-imminent situation. The study described here was conducted on
a test track under dry and wet pavement conditions to examine the effects of ABS versus conventional brakes, ABS brake pedal
feedback level, and ABS instruction on driver behavior and crash avoidance performance. This study found that drivers do tend to
brake and steer in realistic crash avoidance situations and that excessive steering can occur. However, a significant number of road
departures did not result from this behavior for either pavement condition. ABS was found to reduce crashes significantly on wet
pavement as compared to conventional brakes.
17. Key Words
18. Distribution Statement
Antilock Brake Systems, ABS, light vehicle braking, driver behavior, collision
avoidance systems
19. Security Classif. (of this report)
Unclassified
Form DOT F 1700.7 (8-72)
20. Security Classif. (of this page)
Unclassified
Document is available to the public from the
National Technical Information Service
Springfield, VA 22161
21. No. of Pages
22. Price
161
Reproduction of completed page authorized
iv
TABLE OF CONTENTS
1.0 INTRODUCTION to the NHTSA LIGHT VEHICLE ABS RESEARCH PROGRAM .......................1
1.1.
CRASH DATA...........................................................................................................................1
1.2.
NHTSA’S LIGHT VEHICLE ABS RESEARCH PROGRAM.................................................2
1.3.
INTENT OF THE RESEARCH PROGRAM ............................................................................5
1.4.
TASK 5: DRIVER CRASH AVOIDANCE BEHAVIOR USING CONVENTIONAL AND
ANTILOCK BRAKES – BACKGROUND AND PURPOSE................................................................5
2.0 RELATED RESEARCH..............................................................................................................................7
2.1.
DRIVING SIMULATOR STUDY OF EMERGENCY BEHAVIOR .......................................7
2.2.
OTHER RELEVANT STUDIES ...............................................................................................7
2.3.
Comments on Related Research .................................................................................................10
3.0 METHOD......................................................................................................................................................11
3.1.
SUBJECTS .................................................................................................................................11
3.2.
PILOT TESTING FOR SCENARIO REFINEMENT ...............................................................12
3.3.
EXPERIMENTAL DESIGN......................................................................................................12
3.4.
INSTRUMENTATION..............................................................................................................16
3.5.
TEST SCENARIO......................................................................................................................20
3.6.
3.6 RUSE.................................................................................................................................26
3.6.
3.6 RUSE.................................................................................................................................27
3.7.
TEST PROCEDURE..................................................................................................................27
3.8.
DATA ANALYSIS ....................................................................................................................29
4.0 RESULTS......................................................................................................................................................31
4.1.
Useful Definitions ......................................................................................................................31
4.2.
OVERALL .................................................................................................................................35
4.3.
BRAKE SYSTEM: ABS VS. CONVENTIONAL ...................................................................44
4.4.
ABS BRAKE PEDAL FEEDBACK..........................................................................................66
4.5.
VEHICLE ...................................................................................................................................72
4.6.
ABS INSTRUCTION.................................................................................................................79
4.7.
BRAKING PRACTICE..............................................................................................................84
4.8.
INTERACTION OF INSTRUCTION AND PRACTICE..........................................................89
4.9.
TIME-TO-INTERSECTION (TTI)............................................................................................90
4.10. GENDER....................................................................................................................................92
4.11. INTERACTION OF GENDER AND BRAKE SYSTEM .........................................................98
5.0 EXAMINATION OF OBSERVED ROAD DEPARTURES ....................................................................99
5.1.
FULL ROAD DEPARTURES ...................................................................................................99
5.2.
PARTIAL ROAD DEPARTURES ............................................................................................102
6.0 QUESTIONNAIRE RESULTS ...................................................................................................................107
6.1.
DRIVER DEMOGRAPHICS.....................................................................................................107
6.2.
Driving Habits and Experience ..................................................................................................107
6.3.
DRIVER BEHAVIOR................................................................................................................109
6.4.
VEHICLE INFORMATION ......................................................................................................112
6.5.
PAST BRAKING EXPERIENCES ...........................................................................................112
6.6.
RESPONSES TO QUESTIONS REGARDING THE INCURSION SCENARIO ...................114
6.7.
QUESTIONS Regarding DRIVER KNOWLEDGE OF ABS ...................................................117
7.0 DISCUSSION................................................................................................................................................119
7.1.
REALISM OF THE SCENARIO...............................................................................................119
7.2.
RESEARCH QUESTIONS ........................................................................................................119
v
8.0 CONCLUSIONS...........................................................................................................................................124
9.0 REFERENCES .............................................................................................................................................126
10.0 APPENDICES.............................................................................................................................................128
10.1. APPENDIX A: Subject Recruitment Newspaper Advertisement ..............................................129
10.2. APPENDIX B: Subject Recruitment Flier .................................................................................130
10.3. APPENDIX C: Information Summary and Informed Consent Form........................................131
10.4. APPENDIX D: Subject Data Form ............................................................................................134
10.5. APPENDIX E: Outline of Points Covered in ABS Instruction Video .......................................135
10.6. APPENDIX F: Script for Pre-Recorded CD Instructions In Dry Pavement Testing ................137
10.7. APPENDIX G: Script for Pre-Recorded CD Instructions in Wet Pavement Testing ...............138
10.8. APPENDIX H: In-Vehicle Log Sheet ........................................................................................139
10.9. APPENDIX I: Post-Drive Debrief ............................................................................................140
10.10. APPENDIX J: Post-Drive Questionnaire ..................................................................................141
vi
LIST OF FIGURES
Figure 1.
Quad-frame DASCAR video showing a subject approaching the simulated intersection prior to
incursion. .....................................................................................................................................19
Figure 2.
Quad-frame DASCAR video showing a subject approaching the simulated intersection during
the incursion. ...............................................................................................................................19
Figure 3.
Illustration of location and layout of test courses used in the dry and wet pavement testing [14].
21
Figure 4.
Front view of incursion vehicle. ..................................................................................................22
Figure 5.
Rear view of incursion vehicle showing truss support structure. ................................................22
Figure 6.
Illustration of intersection layout and vehicle positioning prior to the incursion. .......................24
Figure 7.
Illustration of incursion vehicle position after the incursion. ......................................................24
Figure 8.
Photograph of intersection with actual vehicles in position prior to incursion scenario. ............26
Figure 9.
Photograph of intersection with foam vehicles after incursion. ..................................................26
Figure 10. Sensor data plot showing evidence of wheel lockup. ..................................................................30
Figure 11. Sensor data plot showing evidence of ABS activation. ...............................................................30
Figure 12. Illustration of steering input definitions for a subject whose initial steering input was in the
opposite direction of the avoidance steering input. .....................................................................33
Figure 13. Illustration of steering input definitions for a subject whose initial steering input was the same
as the avoidance steering input. ...................................................................................................34
Figure 14. Maximum steering input rate frequency distribution by pavement condition. ...........................41
Figure 15. Maximum brake pedal force frequency distribution by pavement condition. .............................42
Figure 16. Brake pedal application duration frequency distribution by pavement condition. ......................43
Figure 17. Successful avoidance maneuver with conventional brakes in which the subject steered left
around the incursion vehicle........................................................................................................46
Figure 18. Brake line pressure and wheel speed data illustrating lockup of the right front wheel. ..............47
Figure 19. Successful avoidance maneuver with ABS in which the subject steered left around the incursion
vehicle..........................................................................................................................................48
Figure 20. Brake line pressure and wheel speed data illustrating ABS activation........................................49
Figure 21. Successful avoidance maneuver with conventional brakes in which the subject braked the
vehicle to a stop. ..........................................................................................................................50
Figure 22. Successful avoidance maneuver with ABS in which the subject braked the vehicle to a stop....50
Figure 23. Frequency distribution of avoidance steering input magnitudes by brake system for dry
pavement......................................................................................................................................54
Figure 24. Frequency distribution of avoidance steering input magnitudes by brake system for wet
pavement......................................................................................................................................55
Figure 25. Frequency distribution of avoidance steering input range by brake system as a function of
whether ABS was activated or whether wheels were locked with conventional brakes for wet
pavement......................................................................................................................................56
Figure 26. Frequency distribution of maximum steering wheel input range by brake system and whether
ABS was activated or whether wheels were locked in the conventional brake system condition.
57
Figure 27. Frequency distribution of maximum steering input rates by brake system for dry pavement.....57
Figure 28. Frequency distribution of maximum steering input rates by brake system for wet pavement. ...58
Figure 29. Frequency distribution of maximum steering wheel input rate by brake system and whether
ABS was activated or whether wheels were locked with conventional brakes. ..........................58
Figure 30. Frequency distribution of first lane recovery steering input range by brake system and whether
ABS was activated or whether wheels were locked with conventional brakes. ..........................59
Figure 31. Frequency distribution of maximum brake pedal force by brake system for dry pavement........61
Figure 32. Frequency distribution of maximum brake pedal force by brake system for wet pavement. ......62
Figure 33. Brake pedal application duration frequency by brake system for dry pavement.........................63
Figure 34. Brake pedal application duration frequency by brake system for wet pavement. .......................63
vii
Figure 35.
Figure 36.
Figure 37.
Figure 38.
Figure 39.
Figure 40.
Figure 41.
Figure 42.
Figure 43.
Figure 44.
Figure 45.
Figure 46.
Figure 47.
Figure 48.
Figure 49.
Figure 50.
Figure 51.
Figure 52.
Figure 53.
Frequency distribution graph for brake pedal application duration for ABS by vehicle for dry
pavement. (Lumina = light feedback, Taurus = heavy feedback)...............................................70
Frequency distribution graph for brake pedal application duration for ABS by vehicle for wet
pavement......................................................................................................................................70
Frequency distribution graph for brake pedal application duration for conventional brake system
by vehicle for dry pavement. (Lumina = light feedback, Taurus = heavy feedback) .................76
Frequency distribution graph for brake pedal application duration for the conventional brake
system condition by vehicle for wet pavement............................................................................76
Maximum steering wheel input range frequency distributions by vehicle. .................................79
Frequency distribution of avoidance steering input magnitudes by gender for dry pavement. ...95
Frequency distribution of avoidance steering input magnitudes by gender for wet pavement....95
Frequency distribution of avoidance steering input rate by gender for dry pavement. ...............96
Frequency distribution of avoidance steering input rate by gender for wet pavement. ...............96
Illustration of the avoidance maneuver of a female subject that ended in a four-wheel, left side
road departure with ABS on dry pavement. ..............................................................................100
Measured applied brake pedal force and associated brake line pressures during an avoidance
maneuver in which the vehicle fully departed the roadway and ABS activation is believed to
have occurred.............................................................................................................................100
Illustration of the avoidance maneuver of a male subject that ended in a four-wheel, right side
road departure with ABS on dry pavement. ..............................................................................101
Measured applied brake pedal forces and associated brake line pressures during an avoidance
maneuver in which the vehicle fully departed the roadway and ABS activation is believed to
have occurred.............................................................................................................................102
Illustration of the avoidance maneuver of a male subject that ended in a two-wheel, right side
road departure with conventional brakes on dry pavement. ......................................................103
Measured applied brake pedal forces and associated brake line pressures during an avoidance
maneuver in which the vehicle partially departed the roadway and wheel lockup with
conventional brakes is believed to have occurred. ....................................................................103
Illustration of the avoidance maneuver of a female subject that ended in a two-wheel, left side
road departure with ABS on dry pavement. ..............................................................................105
Measured applied brake pedal forces and associated brake line pressures during an avoidance
maneuver in which the vehicle partially departed the roadway and ABS activation is believed to
have occurred.............................................................................................................................105
Illustration of the avoidance maneuver of a female subject that ended in a one-wheel road
departure with conventional brakes on wet pavement...............................................................105
Measured applied brake pedal forces, wheel speeds, and associated brake line pressures during
an avoidance maneuver in which the vehicle partially departed the roadway and wheel lockup
with conventional brakes is believed to have occurred. ............................................................106
viii
LIST OF TABLES
Table 1.
Table 2.
Table 3.
Table 4.
Table 5.
Table 6.
Table 7.
Table 8.
Table 9.
Table 10.
Table 11.
Table 12.
Table 13.
Table 14.
Table 15.
Table 16.
Table 17.
Table 18.
Table 19.
Table 20.
Table 21.
Table 22.
Table 23.
Table 24.
Table 25.
Table 26.
Table 27.
Table 28.
Table 29.
Table 30.
Table 31.
Table 32.
Table 33.
Table 34.
Table 35.
Table 36.
Table 37.
Experimental Design for Main Test (shaded cells represent conditions not tested)....................14
Measured parameters (sensor data channels)...............................................................................17
Derived parameters......................................................................................................................18
Crash avoidance strategy prevalence and related crash results. ..................................................36
Percent of subjects that released throttle and percent that steered as initial reaction. .................36
Brake Input Transition Times. .....................................................................................................37
Steering Input Transition Times. .................................................................................................37
Crash avoidance strategy prevalence and related crash results. ..................................................38
Percentage of responses by steering input direction and related crashes by pavement condition.
38
Percentage of avoidance steering inputs in a particular direction based on initial steering
direction and related crashes for the dry pavement condition.....................................................38
Percentage of avoidance steering inputs in a particular direction based on initial steering
direction and related percent of crashes for the wet pavement condition....................................38
Selected steering input results. ....................................................................................................39
Average of avoidance steering input magnitudes and related crash results by crash avoidance
strategy.........................................................................................................................................40
Average avoidance steering input ranges and crash results by crash avoidance strategy............40
Selected braking input results......................................................................................................42
Scenario Entrance Speed by Brake System. ................................................................................44
Reaction Time Measures by Brake System. ................................................................................44
Crash avoidance strategy prevalence and related crash percentages for dry pavement...............45
Crash avoidance strategy prevalence and related crash percentages for wet pavement. .............45
Transition Time Measures by Brake System. (P-values correspond to the comparison of the 2
rows they span. Non-significant results are denoted by p-values that are shaded.)....................51
Percentage of responses for steering input direction by pavement condition..............................51
Steering input measures by brake system (P-values correspond to the comparison of the 2 rows
they span. Non-significant results are denoted by p-values that are shaded.) ............................52
Steering input measures by brake system (Non-significant results denoted by shaded p-values.
53
Braking input measures by brake system (P-values correspond to the comparison of the 2 rows
they span. Non-significant results are denoted by p-values that are shaded.) ............................60
Percent ABS activations and conventional wheel lockups by brake system. ..............................64
Percentage of brake pedal misses and braking application techniques by brake system.............64
Deceleration by brake system and wheel lockup / ABS activation. ............................................64
Lateral Acceleration by brake system and wheel lockup / ABS activation. ................................65
Percent of subjects who crashed into the incursion vehicle as a function of brake system and
pavement condition. (Value pairs marked with one or more asterisks were significantly
different.) .....................................................................................................................................65
Scenario entrance speed by ABS brake pedal feedback level. ....................................................66
Reaction times by ABS brake pedal feedback level. ...................................................................66
Steering behavior measures by ABS brake pedal feedback level................................................68
Braking behavior measures by ABS brake pedal feedback level. ...............................................69
ABS activations and conventional wheel lockup percentages by ABS brake pedal feedback
levels, brake system, and vehicle.................................................................................................71
Percentages of brake pedal misses and braking application techniques by brake system. ..........71
Accelerations by ABS brake pedal feedback level. .....................................................................71
Percent Crashes by ABS brake pedal feedback levels, brake system, and whether ABS was
activated.......................................................................................................................................72
ix
Table 38.
Table 39.
Table 40.
Table 41.
Table 42.
Table 43.
Table 44.
Table 45.
Table 46.
Table 47.
Table 48.
Table 49.
Table 50.
Table 51.
Table 52.
Table 53.
Table 54.
Table 55.
Table 56.
Table 57.
Table 58.
Table 59.
Table 60.
Table 61.
Table 62.
Table 63.
Table 64.
Table 65.
Table 66.
Table 67.
Table 68.
Table 69.
Table 70.
Comparisons for scenario entrance speed and reaction times by brake system, vehicle, and
pavement condition. (Non-significant results are denoted by shaded p-values.).......................73
Comparisons for steering and brake input measures by brake system, vehicle, and pavement
condition. (Non-significant results are denoted by p-values that are shaded.) ..........................74
Comparisons for brake input magnitudes by brake system, vehicle, and pavement condition
(cont.). (Non-significant results are denoted by p-values that are shaded.)................................75
Comparisons for lane recovery, lateral acceleration, and crashes by brake system, vehicle, and
pavement condition. (Non-significant results are denoted by p-values that are shaded.)..........77
Road departures by brake system, vehicle and pavement condition. ..........................................77
ABS activations and conventional wheel lockup percentages by ABS brake pedal feedback
levels, brake system, and vehicle.................................................................................................78
Brake pedal misses and braking application techniques by vehicle and gender..........................78
Scenario Entrance Speeds by ABS instruction. ...........................................................................80
Reaction times by ABS instruction..............................................................................................80
Steering behavior by ABS instruction. ........................................................................................81
Braking behavior by ABS instruction (P-values correspond to the comparison of the 2 rows they
span. Non-significant results are denoted by p-values that are shaded.) ....................................82
Percentage of ABS activations by ABS instruction and conventional brake system wheel lockup
cases on dry pavement. ................................................................................................................83
Percentage of ABS activations by ABS instruction and conventional brake system wheel
lockups on wet pavement.............................................................................................................83
Percentage of missed and non-standard brake pedal application techniques by ABS instruction
and pavement condition...............................................................................................................83
Percent Crashes by ABS instruction............................................................................................84
Reaction times by braking practice for dry pavement. ................................................................85
Steering behavior by braking practice for dry pavement (P-values correspond to the comparison
of the 2 rows they span. Non-significant results are denoted by p-values that are shaded.).......86
Braking behavior by braking practice for dry pavement (P-values correspond to the comparison
of the 2 rows they span. Non-significant results are denoted by p-values that are shaded.).......87
Percentage of conventional brake system wheel lockups and ABS activations by braking
practice for dry pavement. ...........................................................................................................87
Percentage of missed and non-standard brake pedal application techniques by brake system and
braking practice. ..........................................................................................................................88
Deceleration results by braking practice for dry pavement. ........................................................88
Percent ABS activations, cases of conventional brake system wheel lockup, and associated
crashes by braking practice for the dry pavement condition. ......................................................89
Crash results (in percentages) based on the two training methods of instruction and practice, and
by brake system. (Value pairs marked with asterisks are significantly different.) ......................89
Scenario entrance speed by TTI. .................................................................................................90
Reaction times by TTI for dry pavement.....................................................................................90
Steering behavior by TTI (P-values correspond to the comparison of the 2 rows they span.
Non-significant results are denoted by p-values that are shaded.)...............................................91
Braking behavior by TTI (P-values correspond to the comparison of the 2 rows they span. Non-
significant results are denoted by p-values that are shaded.).......................................................91
Comparison of selected measures for the 2.5 second TTI condition dry versus wet pavement. .92
Scenario entrance speed by gender..............................................................................................93
Reaction times by gender for dry and wet pavements. ................................................................93
Steering behavior by gender (P-values correspond to the comparison of the 2 rows they span.
Non-significant results are denoted by p-values that are shaded.)...............................................94
Braking behavior by gender (P-values correspond to the comparison of the 2 rows they span.
Non-significant results are denoted by p-values that are shaded.)...............................................97
Crashes as a function of gender, brake system, and pavement condition....................................98
x
Table 71.
Table 72.
Table 73.
Table 74.
Table 75.
Table 76.
Table 77.
Table 78.
EMPLOYMENT STATUS: Response Percentages. ................................................................107
DRIVING EXPERIENCE: Response Percentages...................................................................108
DRIVING BEHAVIOR: Response Percentages. .....................................................................110
DRIVING BEHAVIOR: Response Percentages (Continued). .................................................111
VEHICLE INFORMATION: Response Percentages...............................................................112
PAST BRAKING / OFF ROAD EXPERIENCES: Response Percentages..............................113
Summary of crash results by condition for dry pavement. ........................................................122
Summary of crash results by condition for wet pavement.........................................................122
xi
EXECUTIVE SUMMARY
In 1997, NHTSA initiated a three-year Light Vehicle ABS Research Program to examine all
plausible reasons why crash data studies had not shown that ABS had improved automobile safety
by producing a net reduction in fatal crashes. In fact, ABS was associated with a statistically
significant increase in the frequency of single-vehicle, run-off-road (rollovers or impacts with fixed
objects) fatal crashes, as compared to cars without ABS. One hypothesis for this phenomenon was
based on the idea that, if drivers tend to "oversteer" during a crash avoidance maneuver, ABS may
give them the ability to steer their vehicles off-road in cases in which a vehicle equipped with
conventional brakes would experience wheel lockup and skid in the direction of the vehicle's
momentum with little directional control.
To investigate this hypothesis, an experiment was conducted in which drivers' collision avoidance
behavior in a simulated right-side intersection incursion scenario was examined as a function of
vehicle brake system (conventional, ABS) and pavement condition (dry, wet). A crash avoidance
scenario was staged in which a stopped vehicle would suddenly move across the path of a subject
vehicle at an intersection causing the subject to take some evasive action to avoid colliding with the
incursion vehicle. This scenario was run with a large number of subjects under both dry and wet
pavement conditions to examine drivers' behavior in a crash-imminent situation and evaluate their
crash avoidance performance with ABS versus conventional brakes. In addition to brake system and
pavement condition, independent variables examined included time-to-intersection, ABS brake
pedal feedback level, gender, and the effects of ABS instruction and braking practice.
Results of this study found that nearly all subjects both braked and steered during their crash
avoidance maneuvers. In fact, subjects in these studies demonstrated the capability to make
aggressive steering and braking inputs. Some evidence of driver oversteering was seen. However,
despite the high magnitudes and rates of many steering inputs observed, very few road departures
occurred. Those road departures that were observed could not be judged attributable to ABS
performance nor driver interaction with ABS. Although these data suggest that oversteering with
ABS may not be responsible for the increase in single-vehicle road departure crashes, it is not clear
whether the extent to which oversteering was seen in this study is comparable in proportion to that
associated with the road departure crash trend phenomenon.
ABS was found to have beneficial effects on crash rates for some conditions in this research. On
wet pavement, 97 percent of subjects activated ABS in the intersection incursion scenario. ABS
was associated with significantly fewer crashes on wet pavement as compared to conventional
brakes.
With no ABS instruction or braking practice, subjects in the ABS condition crashed 50% less than
those in the conventional brake system condition on wet pavement. No significant reduction in
crashes was seen on dry pavement for ABS versus conventional brakes regardless of training
provided. Providing subjects with video instruction on the proper use of ABS did not produce a
significant reduction in crashes for either pavement condition. For subjects in the dry pavement
study who received braking practice prior to the incursion event, those with ABS crashed half as
much as those with conventional brakes. Although providing ABS instruction did not reduce
crashes in this research, there was evidence that ABS instruction may reinforce proper braking
techniques.
xii
Heavy ABS brake pedal feedback was associated with fewer crashes on wet pavement than was
light ABS brake pedal feedback, however, not at a significant level. No evidence of subjects being
startled by ABS brake pedal feedback and removing their foot from the brake pedal was seen in this
research.
In conclusion, the results of this study do not appear to indicate that a problem exists due to driver
crash avoidance behavior or driver interaction with ABS that would contribute to the apparent
increase in fatal single-vehicle crashes as identified in conjunction with vehicles transitioning from
conventional to antilock brake systems. Results from this study will be examined in conjunction
with the results of other tasks included in NHTSA's Light Vehicle ABS Research Program to
determine whether the collective results viewed as a whole provide some insight into the cause of
the increase in fatal single-vehicle crashes observed in conjunction with the implementation of
ABS.
The authors acknowledge that these results are specific to the particular intersection incursion
scenario used in this study. The results may not apply to other types of crash avoidance scenarios.
In addition, the authors feel testing of this sort involving higher travel speeds (greater than 45 mph
on dry and 35 mph on wet pavement) should be investigated. Additional insight may be obtained
by conducting similar research using different crash avoidance scenarios and vehicle travel speeds.
xiii
1.0 INTRODUCTION to the NHTSA LIGHT VEHICLE ABS RESEARCH PROGRAM
Since 1985, antilock brake systems (ABS) have been increasingly available on many passenger
car and light truck make/models. ABS has been sold in four-wheel and two-wheel versions, with
four-wheel ABS being found primarily on passenger cars and two-wheel ABS being prevalent on
light trucks. These systems have been marketed as an added safety feature designed to enhance
drivers' ability to control a vehicle.
With the introduction of ABS, the National Highway Traffic Safety Administration (NHTSA)
undertook a series of investigations to determine the potential benefits of ABS and the effect of
ABS on crash rates. Test programs have shown that these systems appear to be very promising
safety devices when evaluated on a test track. Under many braking conditions on paved
surfaces, four-wheel ABS allows the driver to stop a vehicle more rapidly than with conventional
brakes while maintaining steering control even during situations of extreme, panic braking.
Brake experts anticipated that the introduction of ABS on passenger vehicles would reduce both
the number and severity of crashes. However, a number of crash data analyses have been
performed in recent years by NHTSA, automotive manufacturers, and others which have shown
for passenger cars that the introduction of ABS has not been associated with a net reduction in
crashes to the expected extent.
1.1. CRASH DATA
Kahane [1] found that, for passenger cars, involvements in multi-vehicle crashes on wet roads
were significantly reduced for cars equipped with ABS: fatal crashes were reduced by 24
percent, and nonfatal crashes by 14 percent. A significant 27 percent decrease in fatal collisions
with pedestrians and bicyclists was also found to be associated with ABS. However, these
reductions were offset by a statistically significant increase in the frequency of single-vehicle,
run-off-road crashes, as compared to cars without ABS. Run-off-road crashes, as considered in
this report, included rollovers, side impacts with fixed objects, and frontal impacts with fixed
objects. Fatal run-off-road crashes were up by 28 percent and nonfatal crashes by 19 percent
with ABS. On wet roads, fatal run-off-road crashes increased 17 percent and non-fatal
run-off-road crashes increased by 24 percent. On dry roads, fatal run-off-road crashes increased
by 29 percent while non-fatal crashes increased by 17 percent. It is unknown to what extent, if
any, this increase is due to ABS or other causes. It is also unknown to what extent, if any, this
increase is due to drivers incorrect usage of ABS or incorrect responses by drivers to their ABS.
Hertz, Hilton, and Johnson [2] presented results for passenger car run-off-road crashes according
to the following crash types: rollovers, side impacts with parked vehicles or fixed objects, and
frontal impacts with parked vehicles or fixed objects. For dry roads, ABS was found to be
associated with a 17 percent decrease in all rollover crashes, a 13 percent decrease in all frontal
impacts with parked vehicles or fixed objects, and a 7 percent increase in all side impacts with
parked cars or fixed objects. For all pedestrian crashes, ABS was associated with a 30 percent
reduction on dry roads and a 10 percent reduction in unfavorable road conditions (i.e., wet,
snowy, icy, gravel). In regards to only those crashes involving fatalities, ABS was found to be
associated with a 51 percent increase in fatal rollover crashes on dry roads. For fatal side impact
crashes, ABS produced a 69 percent increase for unfavorable road conditions, and a 61 percent
increase for favorable road conditions. ABS was associated with a 38 percent decrease in fatal
1
pedestrian crashes in unfavorable road conditions. Fatal frontal impact crashes in unfavorable
road conditions were also decreased by 40 percent with the introduction of ABS.
In comparison, some benefits were observed for light vehicles other than automobiles (pickup
trucks, sport utility vehicles, and vans), equipped with two-wheel ABS (instead of the four-wheel
ABS used on automobiles). Rear-wheel antilock brake systems have been effective in reducing
the risk of nonfatal run-off-road crashes for almost every type of light truck [3]. Nonfatal
rollovers were reduced by 30 to 40 percent. Side impacts with fixed objects were reduced by 15
to 30 percent. Frontal impacts with fixed objects were reduced by 5 to 20 percent.
1.2. NHTSA’S LIGHT VEHICLE ABS RESEARCH PROGRAM
In an effort to investigate possible causes of the crash rate phenomena identified, NHTSA
developed its Light Vehicle ABS Research Program. This program contained nine separate tasks
which address potential theories as to the cause of the lack of net crash benefits such as driver
behavior in a crash-imminent situation, driver response to ABS activation, ABS hardware
performance, and environmental factors (as outlined in [7]). To date, NHTSA research has
found no systematic hardware deficiencies in its examination of ABS hardware performance (as
documented in [8]). It is unknown, however, to what extent the increase in run-off-road crashes
may be due to drivers’ incorrect usage of ABS, incorrect response to ABS activation, incorrect
instinctive driver response (e.g., oversteering), changes in driver behavior (i.e., behavioral
adaptation) as a result of ABS use, and/or some other factor.
Task 1 of NHTSA’s Light Vehicle ABS Research Program, performed by Hertz in 2000 [5] as
mentioned in the previous section, involved performing a new crash data study of the effect on
safety of adding four-wheel ABS to automobiles. This study differed from those previously
conducted [1, 2, 3, 4] in that it focused on newer vehicles and antilock brake systems and
included some methodological improvements. This study endeavored to address whether
whatever problem may have caused the apparent increase in single-vehicle crashes for ABSequipped automobiles still existed following the introduction of newer generation ABS
hardware.
Task 2 [9] of this program involved conducting a national telephone survey to determine drivers’
knowledge and expectations about ABS. The purpose of this 1998 survey was to assess whether
the apparent increase in single-vehicle crashes for automobiles might be due to drivers’
misunderstanding of ABS functionality. Results of the survey showed that, although most
drivers had heard of ABS, many did not know what it did or how it affected vehicle
performance, when it functioned, or even if their vehicle was so equipped. Certain types of
brake pedal feedback from an activated ABS were often misinterpreted, making driver reaction
inappropriate and in some cases potentially dangerous. There was also some evidence drivers of
ABS-equipped vehicles placed more confidence in ABS and what it could do for them than the
non-ABS owners did. Lastly, this survey also found that information imparted at the time of
purchase was the means by which the majority of drivers find out about the brakes on their
vehicle. However, approval ratings for lengthy or mandatory information sessions were not well
received, though some methods held promise.
Task 3 involved the examination of 257 selected single-vehicle 1996 crash reports collected by
the National Automotive Sampling System (NASS). The goal of this work was to determine
2
what differences could be identified in the characteristics of single-vehicle crashes incurred by
ABS-equipped versus non-ABS-equipped automobiles using NASS Crashworthiness Data
System (CDS) cases. Results of this examination of crash cases did not provide conclusive
evidence that ABS had a significant effect on crash rates for the time period covered.
Task 4 [8] measured the braking performance of a group of model year 1993-97 production
ABS-equipped vehicles over a broad range of surfaces and maneuvers. While ABS stopping
performance has been measured by many groups over many years, there is a possibility that poor
performance on some unusual surface or during some maneuver may have been overlooked. If
such could be found, this might explain the apparent increase in single-vehicle crashes of ABSequipped automobiles. Results of this 1997-98 study showed that for most maneuvers, on most
surfaces, ABS-assisted stops yielded distances shorter than those made with the ABS disabled.
The one exception was on loose gravel where stopping distances increased by an average of 27.2
percent overall. Additionally, the vehicular stability observed during testing was almost always
superior with ABS. For the cases in which instability was observed, ABS was not deemed
responsible for its occurrence.
Task 5 examined the hypothesis that the apparent increase in single-vehicle crashes with ABSequipped vehicles is due to driver “oversteering” in crash-imminent situations. In a crash
imminent situation, a driver’s first action is expected to be a very hard application of the brake
pedal. Oversteering occurs when the driver, possibly believing that the hard braking input is
insufficient to avoid the upcoming obstacle (such as another vehicle), rapidly turns the steering
wheel by a large amount. For conventionally braked or rear-wheel ABS only vehicles, this
oversteering has little effect, since the initial driver brake pedal activation is likely to lock the
vehicle’s front wheels. However, for a vehicle equipped with four-wheel ABS (where the ABS
minimizes front wheel lockup and allows the driver to maintain steering capability), the
oversteering may result in the vehicle missing the upcoming obstacle, going off of the roadway,
and being involved in a single-vehicle crash.
Task 5 was divided into multiple subtasks to examine driver crash avoidance behavior with and
without ABS. This task sought to assess the prevalence of driver oversteering and examined the
effects of ABS instruction and braking practice on successfully avoiding a crash. Task 5.1 used
a driving simulator to address this issue. Task 5.2 examined driver crash avoidance behavior in a
test track environment on a dry, high coefficient of friction road surface. Task 5.3 also studied
driver crash avoidance behavior in a test track environment but on a wet, low coefficient of
friction road surface. Results of the 1997-98 test track studies, Tasks 5.2 and 5.3 [10], showed
that drivers do tend to brake and steer in realistic crash avoidance situations and that excessive
steering can occur. However, a significant number of road departures did not result from this
behavior for dry or wet pavement. ABS was found to significantly reduce crashes on wet
pavement as compared to conventional brakes. Results of the 1997 simulator study (Task 5.1)
[11] also showed that excessive steering can occur during realistic crash avoidance situations.
However, this steering was not found to result in a significant number of road departures.
In 2000, Task 6 investigated the effects of ABS during road recovery maneuvers (i.e., when a
driver attempts to maneuver an automobile back onto the roadway after a departure). Many road
departures occur when the driver drives the vehicle in an essentially straight line that leaves the
road. This action may be due to driver inattention, sleepiness, or intoxication. None of these
3
causes are related to the presence or absence of ABS. However, the presence of ABS may or
may not influence the ability of the driver to safely maneuver the vehicle back onto the roadway.
Task 7 involved two separate studies that examined the issue of ABS and behavioral adaptation.
Several studies have found that people drive faster or more aggressively on test tracks in ABSequipped vehicles than with conventionally braked vehicles. The goal of this task was to try to
determine if these trends occur during typical driving on actual public roads.
Task 7 was divided into multiple subtasks. Task 7.1 [12] involved remote, unobtrusive
observation methods to collect data about the behavior (e.g., speed) of drivers. Although a
consistent trend was seen in mean speed by brake system for each site with slightly higher speeds
being observed for drivers of ABS-equipped vehicles, this trend was not statistically significant.
This study showed that type of brake system (ABS or conventional) had no significant effect on
driving speed under the conditions examined. Task 7.2, the subject of this report, sough to assess
possible ABS-related behavioral adaptation through the collection of more detailed data about
the driving behavior of a small number of subjects using instrumented vehicles in a naturalistic
research setting.
Task 8 involved the integration of data from all of the preceding tasks in an attempt to infer why
the crash data studies did not find the anticipated increase in safety for ABS-equipped
automobiles.
Task 9 involved the dissemination of task results. NHTSA has shared knowledge gained through
the program’s research efforts by reporting its findings with interested parties within NHTSA
and the public at large. Summaries of current research efforts and results-to-date have been
presented for discussion.
NHTSA’s Light Vehicle ABS Research Program has only been a first step in assessing the
anticipated safety benefits from ABS. This program deals solely with trying to learn why the
crash data studies did not find the anticipated increase in safety (i.e., reduction in crashes) for
ABS-equipped automobiles. The development of countermeasures to resolve any problems
discovered is left to future research.
4
1.3. INTENT OF THE RESEARCH PROGRAM
NHTSA’s Light Vehicle ABS Research Program is only a first step in assessing the anticipated
safety benefits from ABS. This program deals solely with trying to learn why the crash data
studies did not find the anticipated increase in safety for ABS-equipped automobiles. The
development of countermeasures to resolve any problems discovered is left to future research.
1.4. TASK 5: DRIVER CRASH AVOIDANCE BEHAVIOR USING CONVENTIONAL
AND ANTILOCK BRAKES – BACKGROUND AND PURPOSE
To determine whether some aspect of driver behavior in a crash-imminent situation may be
counteracting the potential benefits of ABS, NHTSA embarked on a series of human factors
studies. These studies, which compose Task 5 of the research program, focus on the examination
of driver crash avoidance behavior as a function of brake system and various other factors.
One theory, which Task 5 sought to address, was whether the apparent increase in fatal
single-vehicle crashes involving ABS-equipped vehicles may be due to characteristics of driver
steering and braking behavior in crash-imminent situations. According to this theory, in
situations of extreme, panic braking, drivers may have a tendency to brake hard and make large,
potentially excessive, steering inputs in an attempt to avoid a crash.
In a crash-imminent situation, a driver’s initial reaction may be either to steer or release the
throttle. If a driver steers as the initial reaction, the secondary response may be to release the
throttle and then apply the brakes (third response). In rare circumstances, a driver may steer
initially, release the throttle, and then never apply the brakes. If the driver releases the throttle as
the initial reaction, the secondary response may be to either brake or steer. If a driver then steers
as the secondary response after having released the throttle, the third response may be to apply
the brakes; likewise, if the secondary response was to brake, the third response may be to steer.
Depending on which combination of reaction possibilities the driver exercises, the implications
of oversteering occurring may be more or less severe.
If the driver brakes hard and steers, oversteering may occur when the driver, possibly believing
that the hard braking input will be insufficient to avoid the obstacle, rapidly turns the steering
wheel by a large amount. For conventionally braked or rear-wheel ABS vehicles, aggressive
braking may lock the front wheels of the vehicle, eliminating directional control capability,
rendering the driver's steering behavior irrelevant. However, with four-wheel ABS, wheel
lockup is minimized. As a result, the vehicle does not lose directional control capability during
hard braking and driver’s steering inputs are then effective in directing the vehicle's motion.
This directional control could result in drivers avoiding multi-vehicle crashes by driving off the
road and, instead, experiencing single-vehicle crashes.
To investigate this theory, Task 5 sought to address issues such as whether:
• Drivers tend to both brake and steer (as opposed to only braking or only steering) during
crash avoidance maneuvers;
• Drivers tend to make large, potentially excessive, steering inputs during crash avoidance
maneuvers;
5
• Drivers' crash avoidance maneuvers in ABS-equipped vehicles result in road departures
more often than in vehicles with conventional brakes; and
• Drivers avoid more crashes in ABS-equipped vehicles than in vehicles with conventional
brakes.
Task 5 of NHTSA's Light Vehicle ABS Research Program includes three studies. Two studies
were conducted on a test track (one on dry pavement, Task 5.2; and one on wet pavement, Task
5.3) and one on the University of Iowa’s Iowa Driving Simulator (IDS) (Task 5.1).
These studies used a right-side intersection incursion scenario to elicit a crash avoidance
response from human subjects. This scenario was chosen because it was likely to induce steering
behavior and had the potential for subjects driving the vehicle off of the road. This obstacle
avoidance scenario is not responsible for all, or even most, run-off-road crashes and results may
not be representative of driver behavior in all situations leading to vehicle road departure. Many
run-off-road crashes occur when drivers are unable to maneuver through a curve in the roadway
or when they are drowsy or under the influence of alcohol. However, it is believed that the
results of this study will be useful in determining not only the extent to which drivers are able to
maneuver a vehicle, but also drivers' physical capacity to supply control inputs to the vehicle.
Insight into drivers' ability to maintain vehicle control during a panic maneuver and ability to
avoid a collision can also be gained from this research.
Although the same intersection incursion scenario was involved in each of these experiments,
each experimental venue provided unique advantages for observing driver behavior. The test
track experiments allowed driver behavior to be examined in a realistic environment at moderate
speeds in real vehicles with simulated obstacles on both dry and wet pavement. The IDS study
allowed for driver behavior to be examined using a highly repeatable test method in a simulated
environment at higher travel speeds and with no chance of actual physical collision or injury.
The fundamental knowledge gained through these tests will aid researchers in assessing the
extent of drivers’ abilities in crash imminent situations. Through this assessment of driver
behavior with both conventional and antilock brakes, researchers will be able to infer whether
ABS enhances drivers’ collision avoidance capabilities over those attainable with conventional
brakes without increasing the probability of roadway departure. This final report discusses the
methods used and results obtained from both the dry and wet pavement test track studies and
attempts to address these issues.
6
2.0 RELATED RESEARCH
This study was not the first to examine driver behavior in an obstacle avoidance scenario. The
following are descriptions of studies of driver behavior in crash-imminent situations including
one study involving ABS.
2.1. DRIVING SIMULATOR STUDY OF EMERGENCY BEHAVIOR
A previous NHTSA study [7,8] performed under Contract No. NRD-20-95-08086 on the Iowa
Driving Simulator (IDS) utilized a crash scenario which was very similar to the one employed in
this research. The IDS study was conducted to examine the collision avoidance behavior and
reaction time to an unexpected intersection incursion while driving a vehicle equipped with
conventional brakes only, since at the time the IDS did not have the capability to simulate ABS.
This study involved an intersection incursion, which occurred at the intersection of two, two-lane
rural highways. Traffic on the crossing road was controlled by stop signs, while the roadway on
which the subject was traveling had no traffic signals and thus had the right of way through the
intersection. The speed limit on this roadway was 55 mph. At one of three possible time-tointersection (TTI) values an incursion vehicle began moving into the intersection in front of the
subject vehicle. The incursion vehicle could intersect from either the driver’s left or right side,
and either completely blocked the driver’s lane or blocked one-half of the driver’s lane.
The principle results of the IDS study indicate drivers are most successful at avoiding an
incursion vehicle when the escape gap is large and/or the required steering magnitude is small.
Drivers appear to use this information, along with the time-to-intersection in the formulation of
an avoidance strategy. The study demonstrated that drivers have more difficulty avoiding an
incursion vehicle when it intrudes into the driver’s lane from the left than from the right.
During a severe obstacle avoidance maneuver, drivers will often lock the brakes and attempt to
steer (with the brakes locked). Drivers may input large amplitude steering movements, or
“oversteer”, in emergency avoidance situations. With conventional brakes locked (the vehicle in
a longitudinal skid with minimal lateral control), the oversteering does no harm. However, for
an ABS-equipped vehicle, a large amplitude steering input is likely to increase the chance of
lateral skidding, roadway departures, and subsequent roll-over crashes. In this study, more than
80% of subjects in this study locked wheels when attempting to avoid the incursion vehicle.
Over 70% of subjects attempted to steer when the wheels were locked. Over 80% of the subjects
who attempted to steer when wheels were locked collided with the incursion vehicle. Thus, it
appears that the hypothesis that drivers tend to exhibit instinctive steering inputs of significant
magnitudes in panic situations is a plausible one and that ABS, by allowing the vehicle’s wheels
to continue to roll during aggressive braking, may be allowing drivers to maneuver their vehicles
into potentially dangerous off-road situations.
2.2. OTHER RELEVANT STUDIES
Several field studies in which an object was projected into the subject vehicle’s path have been
performed to examine drivers’ obstacle avoidance behavior. The method has been successfully
used to investigate driver reaction times and behavior in response to an incursion obstacle. The
following are brief summaries of three such studies.
7
2.2.1. Prynne, K. and Martin, P.: Braking Behavior in Emergencies (1995)
This study [10], performed by Lucas Industries, endeavored to determine in what way, if any, a
typical driver’s emergency braking behavior was inadequate. The experiment created an
unexpected, sudden, obstacle avoidance situation so as to generate driver panic braking behavior.
Seventy-seven subjects were recruited. Subjects were characterized by gender, age, and level of
driving experience. Tests were initially conducted using a polystyrene obstacle that was
propelled into the path of the subject vehicle. The polystyrene obstacle did not, however, elicit a
sufficiently authentic crash avoidance response from the drivers and was eventually replaced
with life-size painted figures representing children running out into the road. These child-like
obstacles were judged to be successful in achieving a higher degree of realism. The life-size
painted figures proved to be realistic enough to elicit a genuine crash avoidance response from
subjects but not so life-like that the subjects feared that they were facing a potential collision
with an actual child.
This study found that drivers typically exhibit a two-stage braking behavior in response to
potential frontal crash conflicts. Most subjects initially applied the brake moderately and then
hesitated momentarily, holding the brake pedal steady at a moderate application level. Then,
when subjects perceived the threat could not be avoided by a moderate brake application and
required an all-out avoidance response, they continued their brake application to achieve full
brake application. For tests involving the child-like obstacles, 66 of the 77 subjects
demonstrated a pause or break in their brake application. Fifteen subjects collided with the
obstacles.
Prynne, et al. attempted to create a realistic driving event by projecting obstacles in the path of
unprepared drivers. In this regard, the study is very similar to that used in Task 5 of NHTSA’s
Light Vehicle ABS Research Program. Unfortunately,