Factors that Increase and Decrease Motorcyclist Crash Risk...• 55% of these 53 single-vehicle (motorcycle) events involved the participant negotiating a curve leading into the crash

©2016MotorcycleSafetyFoundation

Factors that Increase and Decrease Motorcyclist Crash RiskVicki Williams, Shane McLaughlin, and Jon AtwoodVirginia Tech Transportation InstituteTim BucheMotorcycle Safety Foundation


Organizational Mission• Virginia Tech Transportation Institute

– Conducts research to save lives, save time, save money, and protect the environment

– Develops the techniques and technologies to solve transportation challenges from vehicular, driver, infrastructure, and environmental perspectives

– Specifically for motorcycle safety, uses the collection of real-world driving/riding data and analysis/data mining to improve safety, with a focus on the user


Organizational Mission• Motorcycle Safety Foundation

– Mission: To make motorcycling safer and more enjoyable by ensuring access to lifelong quality education and training for current and prospective riders, and by advocating a safer riding environment.

– Vision: The MSF is an internationally recognized not-for-profit foundation, supported by motorcycle manufacturers, that provides leadership to the motorcycle safety community through its expertise, tools, and partnerships.


Understanding Crash Risk• Traditional methods to understand crash risk rely on

post-event analyses• Other methods include simulators and controlled

experimentation• Observance of crash events via naturalistic study

reveals conditions that would otherwise remain unknown

• In addition, near-crash events (surrogates for crashes) are observed as never before


Training Systems Development


The Study

• Sponsored by MSF, who assisted with day-to-day operations

• Instrumentation of 100 riders’ personal motorcycles (riding as they normally do)

• Recorded video and kinematic data (collected 366,667 miles)

• First large-scale naturalistic motorcycle study to provide this type of unique and complex data

The MSF 100 Motorcyclists Naturalistic Study


MSF 100 Motorcyclists Naturalistic Study


• GPS• Machinevisionlane

tracker• Accelerometers(3axes)• Gyro(3axes)• Forwardradar

• TurnSignals• Brakeleverinputs• Continuouscollection• 8-12mocapacity• Cellular

communicationfrombikesbacktoVTTI

• Fivecolorcameras• forward• rear• lefthand• righthand• ridertorso



Frequency

Female 22

Male 78

Cruiser 41Sport 21

Touring 38



Collected Data• Studyparticipation

• Range:2monthsto2years• Totalof30,844trips• Totalof366,667miles• Totalinstalledtimeof100.6years

• Collectedevents• 30crashes• 122near-crashes• Eventsperriderrangedfrom0to13• 55ridersexperiencedatleastoneevent


Descriptive Statistics for CNCsCNC Mean Median Std.Dev. Minimum Maximum

CountperParticipant 1.54 1 2.18 0 13Rateper1000Miles

perParticipant 0.87 0.18 2.85 0 27.03

• Sampleparticipantsaveraged1.5CNCeventsperrider• Whenexpressedasarate,theaverageparticipantnoticedaCNC

rateof0.87per1,000milestraveled• 34%oftheridersinthestudyaccountedfor86%ofthecrashes

andnear-crashes


Crash Descriptions• “Crash,”asdefinedbythisstudy,includes:

• Anycontactthatthesubjectvehiclehaswithanobject,eithermovingorfixed,atanyspeed.

• Non-premeditateddeparturesoftheroadwaywhereatleastonetireleavesthepavedorintendedtravelsurfaceoftheroad.

• Anycontactbetweenthegroundandthebike(otherthantires/stands)orgroundandrider(otherthanfoot).

• 57%ofthe30crasheswerelow-speed“capsizes”• Othercrasheswereofvarioustypes,asindicatedinthenextslide


Crash Descriptions


• ExampleofGroundImpact– LowSpeed(“capsize”)Crash Descriptions

Towatchvideo:https://youtu.be/C6r3vnebK5k


• ExampleofOtherVehicleTurnAcrossPath

Crash Descriptions

Towatchvideo:https://youtu.be/I0AvOgciHD8


Single-Vehicle Crash and Near-Crash Descriptions

• Thenextslideprovidesanindicationofthetypesofcrashesandnear-crashes(wherearapid,evasivemaneuverisrequired)thatinvolvedonlytheparticipantbike(noothervehicles,objects,orpedestrianswereinfluentialintheevent)

• Therewere53casesofthesesingle-vehicleevents(involving29differentriders)

• 55%ofthese53single-vehicle(motorcycle)eventsinvolvedtheparticipantnegotiatingacurveleadingintothecrashornear-crash

• Theremainderoftheseeventsinvolvedvariousscenarios,asindicatedinthefollowingtable


Single-Vehicle Crashes(SVCs)

&Near-

Crashes

PrecipitatingEvent Pre-incidentManeuver NumberofEvents

PercentageofSVCs

Subjectoverleftlaneline Negotiatingacurve 18 34.0%Subjectoverleftedgeofroad Turningright 1 1.9%

SubjectoverrightedgeofroadNegotiatingacurve 4 7.5%Goingstraight,butwithunintentional"drifting"withinlaneoracrosslanes 1 1.9%

Subjectoverrightlaneline Negotiatingacurve 2 3.8%Thisvehiclelostcontrol-excessivespeed

Goingstraight,constantspeedordecelerating 4 7.5%Negotiatingacurve 3 5.7%

Thisvehiclelostcontrol-insufficientspeed

Entering/leavingaparkingposition,movingforward 3 5.7%Goingstraight,constantspeedordecelerating 3 5.7%Turningright 2 3.8%Turningleft 1 1.9%Backingup(otherthanforparkingpurposes) 1 1.9%MakingU-turn 1 1.9%Negotiatingacurve 1 1.9%Startingintrafficlane 1 1.9%Stoppedintrafficlane 1 1.9%

Thisvehiclelostcontrol- othercause

Backingup(otherthanforparkingpurposes) 1 1.9%Negotiatingacurve 1 1.9%

Thisvehiclelostcontrol- poorroadconditions

Goingstraight,constantspeedordecelerating 2 3.8%Turningright 2 3.8%


Single-Vehicle Crash and Near-Crash Descriptions• Exampleofsubjectoverleftlanelinewhilenegotiatingacurve

Towatchvideo:https://youtu.be/xXnPQese0hU


Crash and Near-Crash Descriptions Involving Other Vehicles or Objects• Theremainderofthecrashesandnear-crashesinvolvedat

leastoneothervehicleorobject(e.g.,pedestrian,animal,cyclist)

• Therewere99oftheseevents(involving44differentriders)• 35%ofthese99eventswerecasesofthesubjectbikerear-

endingaleadvehicle• Therestoftheeventsincluded13categoriesofIncidentType,

andareincludedinthefollowingtable


Crash and Near-Crash

Descriptions Involving

Other Vehicles or

Objects

PrimaryIncidentType NumberofEvents

PercentageofMulti-VehicleConflicts

Rear-end,striking 35 35.4%Sideswipe,samedirection(leftorright) 21 21.2%Othervehicleturnacrosspath 8 8.1%Oppositedirection(head-onorsideswipe) 7 7.1%Animal-related 6 6.1%Othervehicleturnintopath(oppositedirection) 6 6.1%Othervehicleturnintopath(samedirection) 5 5.0%Pedestrian-related 3 3.0%Backingintotraffic 2 2.0%Rear-end,struck 2 2.0%Subjectvehicleturnintopath(samedirection) 1 1.0%Other 1 1.0%Pedalcyclist-related 1 1.0%Othervehiclestraight,crossingsubjectpath 1 1.0%


Method of Evaluating Crash/Near-Crash (CNC) Risk

• Videoverificationofcrashandnear-crashevents• Videoanalysisusinga95-variabledatadictionary,VTTIdeveloped/tested

o 7,028baselineevents(“eventless”riding),randomlyselectedperrider,numberbasedonridermileage

o 152crashandnear-crashevents• Oddsofbeinginvolvedinacrashornear-crash(CNC)givenexposuretoa

factorarecalculatedo BasedonoddsofCNCoccurrencewhenexposedtofactor

comparedtooddswhennotexposedtofactoro Factorscanberelatedtotherider,environment,orroadway(these

arethedictionaryvariables)


Results: Factors

that IncreaseCNC Risk

Variable Level Odds Ratio Reference

Exposure to this increases

risk by this many times

compared to:

Intersection Influence Yes, Uncontrolled 40.7 None

Intersection Influence Yes, Parking lot, driveway entrance/exit 8.5 None

Intersection Influence Yes, Traffic signal 2.9 None

Rider Behavior Aggressive riding (only) 17.9 None

Rider Behavior Lack of knowledge or skill/Inattention (only) 9.3 None

Rider Behavior Combination of behaviors 30.4 None

Pre-incident Maneuver Maneuvering to avoid object 11.8 Going straight, constant speed

Surface Type Gravel/Dirt road 9.4 Paved, smoothRoadway Grade Grade down 4.3 LevelRoadway Grade Grade up 1.9 LevelTraffic Density Unstable 3.6 StableRoadway Alignment Curve right 2.1 Straight


Results: Factors that Decrease CNC Risk

Variable Level Odds Ratio Reference

Exposure to this is associated

with a risk that is this many times

the risk for:

Locality Urban 0.1 Open country/ Open residential

Locality Highway 0.2 Open country/ Open residential

Locality Miscellaneous/Other 0.2 Open country/ Open residential

Locality Moderate residential/ Business/Industrial 0.4 Open country/

Open residential

Pillion Riders 1 0.3 0


Contribution to Motorcycle Safety Research

• Discovered12factorsthatincreasetheriskofCNCs• Discovered5factorsthatdecreasetheriskofCNCs• Provideddetailedguidancebasedonriskfactorsthatcanbeincorporatedintotrainingprograms

• Producedalarge,richdatabaseofnaturalisticridinginformationthatwillbeusedforyearstouncovercrashandnear-crashmechanismsandsupportsafety-relatedmotorcycleresearch

• Developedandtestedadatareductiondictionaryspecificallyfornaturalisticmotorcycleanalysisthatcanbeappliedconsistentlyacrossfuturestudies


Contributions to Motorcycle Safety Research

• Observedsomegoodthings,too• Usefulinsupportingtheemphasisofpropertechniqueandexecution

Towatchvideo:https://youtu.be/-5Y0u_Hokhc


Contact InfoVTTI Motorcycle Research Group http://www.motorcycle.vtti.vt.edu/

Motorcycle Safety Foundationhttp://www.msf-usa.org

TimBuchePresident&CEO(949)[email protected]

VickiWilliamsHumanFactorsEngineer(540)[email protected]

ShaneMcLaughlinGroupLeader(540)[email protected]


Descriptive Statistics for CNCs

R²=0.25933

0

2

4

6

8

10

12

14

0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000

Num

bero

fCNCs

TotalMileage

NumberofCrashes/Near-CrashesbyTotalMileageperRider

Documents

Factors that Increase and Decrease Motorcyclist Crash Risk...• 55% of these 53 single-vehicle (motorcycle) events involved the participant negotiating a curve leading into the crash