three stories from the ltccs: using the ltccs for safety research
DESCRIPTION
Three Stories from the LTCCS: Using the LTCCS for Safety Research. Daniel Blower June 4, 2008. Purpose of Presentation. Illustrate the strengths and limitations of the LTCCS using experience from three research projects: Truck mirrors and blind zones. - PowerPoint PPT PresentationTRANSCRIPT
Three Stories from the LTCCS:Three Stories from the LTCCS:Using the LTCCS for Safety ResearchUsing the LTCCS for Safety Research
Daniel Blower
June 4, 2008
Purpose of PresentationPurpose of Presentation
Illustrate the strengths and limitations of the LTCCS using experience from three research projects: Truck mirrors and blind zones. Evaluating effectiveness of advanced stability
technologies. Effect of truck mechanical condition on truck
crashes.
Crash Data OverviewCrash Data Overview
Trucks Involved in Fatal Accidents (TIFA) Fatal crashes only; census file; survey to
supplement FARS; since 1980. General Estimates System (GES)
All police-reported crashes; sample file; nationally representative; since 1987.
Large Truck Crash Causation Study (LTCCS) Fatal, A-, B-injury crashes; sample file; nationally-
representative; 2001-2003
Story One:Story One:
Blind Zone Crashes and Truck MirrorsBlind Zone Crashes and Truck Mirrors
National crash data show that mirror-relevant crashes overinvolved on the right. No detail on truck mirror configuration in any
available crash data. Large Truck Crash Causation Study data
offers more detailed information. Presence of side & fender mirrors. Available materials such as scene diagram,
photos, extensive narrative, interviews with all parties to code additional data.
Mirror and Direct Fields of View for One Truck & Driver
Blocked by left planar mirror Blocked by A Pillar
Source of drawing: Matthew Reed, UMTRI
Note car fits neatly in area visible only using fender-mirror
Area visible using fender mirror
Area visible using planar mirror
HypothesisHypothesis
Fender-mounted mirrors reduce the over-representation of right-side blind zone crashes.
MethodMethod
Reproduce classification algorithm in LTCCS data.
Review scene diagrams/narrative to verify blind zones.
Code position of conflict vehicles.► Are the vehicles in the blind zones?
Compare crash configuration with mirror configuration. ► Do right fender-mounted mirrors help?
Mirror-relevant Crash InvolvementsMirror-relevant Crash InvolvementsLTCCSLTCCS
0.0 1.0 2.0 3.0 4.0
Backing
Start up, nonmotorist
Left turn, other in blindzone
Right turn, other in blind zone
Lane change/merge left
Lane change/merge right • Mirror-relevant are 7.2% of all.
• LCM Right 1.9 times LCM Left
• Right turn 0.7 times left turn
• Backing ~0.3%
Mirror-relevant Crash InvolvementsMirror-relevant Crash InvolvementsLTCCSLTCCS
0.0 1.0 2.0 3.0 4.0
Backing
Start up, nonmotorist
Left turn, other in blindzone
Right turn, other in blind zone
Lane change/merge left
Lane change/merge right • Mirror-relevant are 7.2% of all.
• LCM Right 1.9 times LCM Left
• Right turn 0.7 times left turn
• Backing ~0.3%
Comparison of Mirror-Relevant Comparison of Mirror-Relevant Crashes in LTCCS, TIFA/GESCrashes in LTCCS, TIFA/GES
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Backing
Start up, nonmotorist
Left turn, other in blind zone
Right turn, other in blind zone
Lane change/merge left
Lane change/merge right
LTCCS
TIFA/GES
• Less over-representation on the right.
• Relationship in turns not observed.
• Overall proportion of involvements similar.
Left planar mirror A Pillar
Source of drawing: Matthew Reed, UMTRI
4.4
14.3
1.7
79.524.8
38.8
27.8
8.6
LCM Crashes, Position of Conflict LCM Crashes, Position of Conflict Vehicle and Blind ZonesVehicle and Blind Zones
Association of Fender/hood Mounted Association of Fender/hood Mounted Mirrors with LCM CrashMirrors with LCM Crash
Percentage with Left Fender Mirrors
Percentage with Right Fender Mirrors Trucks in LCM Right
crashes somewhat more likely to have right fender mirrors.
Trucks in LCM Left crashes somewhat less likely to have left fender mounted mirrors.
Neither difference is statistically significant.
29 unweighted LCM right cases.
23 unweighted LCM left cases.
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0
All crashes
LCM Right
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0
All crashes
LCM left
CautionsCautions
Small sample sizes limit conclusions: the size of the effects are not statistically significant.
LTCCS not well-suited for analysis of infrequent events.
Use of researcher’s summary, diagrams, photographs is labor intensive but worth the effort.
Story Two:Story Two:
Evaluating Technology Evaluating Technology to Reduce Rolloverto Reduce Rollover
and Loss of Control Crashesand Loss of Control Crashes
Current Project to Estimate Effectiveness of Current Project to Estimate Effectiveness of Advanced Stability TechnologiesAdvanced Stability Technologies
Roll stability control & electronic stability control technologies available for trucks.
Goal is to reduce rollover & loss of control crashes.
Current project to estimate effectiveness in real-world crashes.
MethodMethod
Identify scope of crashes where the technology is relevant.
Identify vehicle & environmental factors that contribute to the crashes.
Engineering evaluation of real-world crashes to estimate effect.
Hardware-in-the-loop simulation of technologies in real-world crashes.
Relevant Crashes in TIFA, GES, & LTCCS Relevant Crashes in TIFA, GES, & LTCCS
Single-vehicle
14.6
30.0
16.5
0.0
5.0
10.0
15.020.0
25.0
30.0
35.0
40.0
TIFA GES LTCCS
pe
rce
nt
Rollovers
12.2
28.0
14.0
0.0
5.0
10.0
15.020.0
25.0
30.0
35.0
40.0
TIFA GES LTCCS
pe
rce
nt
(K) (K, A, B) (K, A, B) (K) (K, A, B) (K, A, B)
Use of LTCCS CasesUse of LTCCS Cases
Specific limitations: Relatively few cases: 963 crashes, 1123 vehicles. Questions about national representativeness. (E.g.,
about twice as many rollovers as expected) Therefore, in light of the limitations:
Use TIFA/GES to determine national distribution of yaw & roll instability crash scenarios
Apply algorithm to classify yaw and roll cases developed in TIFA/GES to LTCCS to establish crash types
Exploit superior detail of LTCCS to produce exemplars of the crash types for engineering review and simulation
Review of LTCCS Yaw & Roll Instability Review of LTCCS Yaw & Roll Instability CrashesCrashes
Review of: 83 yaw instability cases 81 roll instability cases
Assess for Accuracy of coding Suitability for Hardware in the Loop simulation Crash details (radius of curvature, curve entry speed) Relevance of roll & yaw control technologies (likely,
probably, unlikely, unknown)
•Road curved
•Dry surface
•Cargo: loaded
3-axle tractor pulling bottom dump.
31,000 lbs cargo (dirt)
61,800 gross weight
Est. 40 mph
LTCCS Rollover Case for Simulation
A single vehicle crash occurred on the southbound on-ramp for a four-lane northbound divided state highway in the early afternoon hours. This transition ramp comes off of a two-lane state highway. The ramp is one-lane with a -4/122cm (-3%) grade and a super-elevation of approximately 8/122cm (+7%). Narrow paved shoulders bordered each side of the on-ramp. Paved shoulders bordered the highway as well. The posted speed limit for the northbound highway was 65 mph (105 kmph). The on-ramp speed limit is unknown. There are no curve warning signs posted from the overpass all the way through the ramp curve. This is inadequate roadway signage for the curvature and ramp speed, and considered an associative factor for this crash. There was no traffic congestion, adverse weather conditions or sight restrictions at the time of the crash. Vehicle one, a 1995 Freightliner FLD-120 tractor pulling a 1999 Red River bottom-dump trailer, loaded with dirt was southbound on the on-ramp as it entered the curve. According to the police report, the truck was traveling at about 40 mph (64 kmph) as it drove down the ramp. As the truck approached the end of the curve, the truck began to skid in a clockwise direction, leaving 77 feet (23 meters) of scuffmarks. The trailer rolled to its left, pulling the tractor with and both units fell onto their left sides in the first lane of the eastbound highway. The units slid in an eastbound direction for another 64 feet (20 meters) before coming to rest in the first and second lanes, facing eastbound. The load of dirt was spilled into the second lane. … At the time of the crash the Freightliner was hauling 31,140 lbs (14,125 kgs) of dirt. The entire unit weighed 61,879 lbs (28,067 kgs). … This driver drives this vehicle daily and over this route daily. This driver stated that he down shifted into fourth as he was negotiating the curve when he noticed that the trailer started to rollover. The driver did not realize that he was traveling to fast for the curve with the load that he was carrying.
Researcher’s Narrative from LTCCS
Engineering assessment: Roll control likely relevant; Yaw control unlikely relevant
LessonsLessons
Very rich crash investigations, scene & vehicle documentation support engineering analysis.
Both uniform and of high quality. Reasons to question some national
estimates from LTCCS. No current alternative for rich detail
available from LTCCS.
Story Three:Story Three:
Vehicle Mechanical Condition Vehicle Mechanical Condition and Crash Riskand Crash Risk
State-level Research Showed Vehicle State-level Research Showed Vehicle Mechanical Condition Related to Crash RoleMechanical Condition Related to Crash Role
FACT data from Michigan, 1996-2001 CVSA Level 1 inspection & detailed crash events Rear-end crashes:
Higher rates of brake violations when the truck is the striking vehicle.
Higher rates of light violations when truck is the struck vehicle
FACT program a forerunner of LTCCS
Brake Violations Increase Risk of Striking in Brake Violations Increase Risk of Striking in Rear-end CrashesRear-end Crashes
FACT data FACT data
Brakeinspection results
Rear-end type
TotalTruck striking Truck struck
0 violations 15 26 41
1 or more violation 17 11 28
Total 32 37 69
proportion with violations
0 violations 46.9 70.3 59.4
1 or more violation 53.1 29.7 40.6
Total 100.0 100.0 100.0
chi=3.89, p=0.05
Lights Violations Increase Risk of Being Lights Violations Increase Risk of Being Struck in Rear-end CrashesStruck in Rear-end Crashes
FACT dataFACT data
Inspection resultsRear-end type
TotalTruck striking Truck struck
0 violations 28 23 51
1 or more violations 4 14 18
Total 32 37 69
Proportion of lighting violations
0 violations 87.5 62.2 73.9
1 or more violations 12.5 37.8 26.1
Total 100.0 100.0 100.0
chi=5.71, p=0.02
LTCCS Vehicle Inspection Data Much LTCCS Vehicle Inspection Data Much More ComprehensiveMore Comprehensive
Inspection category
LTCCS
FARS/TIFA OOS Violation
Brakes 21.7 42.1 1.9
Log 8.0 23.0 n/a
Lights 4.1 22.4 0.3
Load securement 3.4 4.6 n/a
Tires 3.3 14.3 0.9
Suspension 2.3 3.5 0.1
HOS 1.4 4.3 n/a
Steering 1.2 1.8 0.1
Frame 1.0 3.1 0.0
Wheels 0.8 3.5 incl. w/tires
Any OOS or violation 35.1 70.5 6.8
Brake Violations More Likely When Brake Violations More Likely When Truck is Striking Vehicle Truck is Striking Vehicle
in Rear-end Crashes in LTCCSin Rear-end Crashes in LTCCSBrake OOS
0.0 10.0 20.0 30.0 40.0
Rear-endstriking
Rear-endstruck
In rear-end crashes, trucks that are the striking vehicle have higher rates of brake violations and out-of-service than as the struck vehicle.
Neither difference is statistically significant.
Same pattern as Michigan FACT data
Brake violations
0.0 10.0 20.0 30.0 40.0 50.0 60.0
Rear-endstriking
Rear-endstruck
Mixed Results for Light System Mixed Results for Light System Violations in Rear-end Crashes, LTCCSViolations in Rear-end Crashes, LTCCS
In rear-end crashes, trucks that are the struck vehicle have higher rates of light system OOS.
Rates the same for all lights violations.
Neither difference is statistically significant.
Different pattern from the Michigan FACT data
Lights OOS
0.0 5.0 10.0 15.0 20.0
Rear-endstriking
Rear-endstruck
Lights violations
0.0 10.0 20.0 30.0 40.0
Rear-endstriking
Rear-endstruck
Lessons to ApplyLessons to Apply
LTCCS invaluable as source of detailed crash data. Engineering reviews of crash narratives, etc. Availability of critical information such as
mirror configuration and mechanical condition. Restricted sample sizes pose problems Care must be used in choosing truck
safety issues and methods to address. Representativeness can be an issue.
Thank you.Thank you.
Questions?Questions?