FESTA Experimental ProceduresDeliverable 23.docx

FESTA Experimental ProceduresDeliverable 23.docx

《FESTA Experimental ProceduresDeliverable 23.docx》由会员分享，可在线阅读，更多相关《FESTA Experimental ProceduresDeliverable 23.docx（59页珍藏版）》请在冰豆网上搜索。

FESTAExperimentalProceduresDeliverable23

Primeronexperimentalprocedures

1May2008

Project:

FieldopErationalteStsupporTAction（FESTA）

Grantagreementno.:

214853

Workpackage:

WP2Methodology

Task:

T2.3Experimentalprocedures

Deliverableno.:

D2.3

Documenttitle:

Primeronexperimentalprocedures

Deliverablenature:

Public

Deliverablestatus:

Draft

Author:

FESTAconsortium

FESTAConsortium:

CentroRicercheFiat,UniversityofLeeds,BMWForschungundTechnikGmbH,DaimlerAG,GieRecherchesetetudesPSARenault,VolvoCarCorporation,VolvoTechnologyCorporation,RobertBoschGmbH,A.D.C.AutomotiveDistanceControlSystemsGmbH,DelphiFranceSAS,LoughboroughUniversity,ChalmersUniversityofTechnology,InstitutNationaldeRecherchesurlesTransportsetleurSécuritéINRETS,StatensVäg-ochTransportforskningsinstitutVTI,NederlandseOrganisatievoortoegepast-natuurwetenschappelijkonderzoekTNO,BundesanstaltfuerStrassenwesenBASt,ValtionTeknillinenTutkimuskeskusVTT,INFOBLUSPA,OrangeFrance,EuropeanRoadTransportTelematicsImplementationCoordinationOrganisationERTICO,UniversitaetzuKoeln

Disclaimer:

TheFESTASupportActionhasbeenfundedbytheEuropeanCommissionDGInformationSocietyandMediainthe7thFrameworkProgramme.ThecontentofthispublicationisthesoleresponsibilityoftheprojectpartnerslistedhereinanddoesnotnecessarilyrepresenttheviewoftheEuropeanCommissionoritsservices.

TableofContents

1Introduction1

2Participants2

2.1Demographics2

2.2Drivingexperience3

2.3Personalityandattitudes4

2.4FOTSamplesizeandpoweranalysis9

3Studydesign12

3.1Hypothesisformulation12

3.2Experimentaldesign15

3.3Threatstovalidity:

Confoundsandotherinterferingeffects18

3.4Integratedfunctions20

4Experimentalenvironment22

4.1Geographicallocation23

4.2Roadtype29

4.3Weatherconditions35

4.4Timeofdayandseasonaleffects40

5References44

1Introduction

ThisdeliverableprovidesguidanceontheoverallexperimentaldesignofFOTsinordertoensureexperimentalrigourandscientificquality.Thefirstsection“Participants”providesadviceonparticipantselection,includingdemographics,drivingexperience,personalityandattitudes,alongwithconsiderationtosamplesize.Thesecondsection“Studydesign”providesguidanceoftheformulationofhypothesis,experimentaldesignandpossibleconfounds.Thethirdsection“Experimentalenvironment”suggestshowtheroadenvironment（roadtype,weatherconditionsetc.）playsapartinthedesignofanFOTandthesubsequentdataanalysis.

2Participants

Dependingupontheresearchquestions,thereisaneedtoselectcertaingroupsofparticipantsforanFOT,oroneneedstoassurethattheselectedparticipantsrepresentthetotalpopulation.Itisclear,forexample,thatifoneisinterestedintheeffectsofacertainsystemspecificallydesignedtoaidelderlydrivers,oneneedstoselectasubgroupofthetotalpopulationconsistingofelderlydrivers.Howeverifoneneedstohavearepresentativesampleofthetotalpopulation,onealsoneedstoknowwhetherthesampledoesnotcontain,forexample,toomanyelderly,ortoomanymales.Inthissection,differentaspectsofparticipantselectionaredescribedaswellastheestimationofthenumberofparticipantsthatasamplemustcontaininorderforthestudytoberobust.

2.1Demographics

Participantdemographicsareoftendefinedbyage,gender,income,disabilities,mobility（intermsoftraveltimetoworkornumberofvehiclesavailable）,educationalattainment,homeownership,employmentstatusandlocation.Inthefollowingsection,someguidelinesareprovidedforselectingparticipants.TobeabletogeneralizefromtheFOTresults,thereisaneedtopayattentiontothetypeofparticipantsandthedataobtained.Inaddition,theselectionofparticipantsshouldbearrelationtothesystemtobetested,tothepopulationthesystemistestedfor,andtotheidentifiedhypothesesandperformanceindicators.

2.1.1Permanentvariables

Ageandgenderarethemostuseddemographicfactors.Itisnecessarytokeepinmindthatifasystemisevaluatedforan“average”male,theresultsmaynotbetransferabletoanelderlyfemale.Thereforeageandgenderneedtobedefinedinanearlyphase.

2.1.2Socioeconomicvariables

Amongthesocioeconomicfactors,income,education,employmentandmaritalstatusareoftenused.Thesefactorscaninfluencenotonlytheeffectsofthesystemdescribedintermsofdriver/drivingbehaviour,butalsoinfluencetheexposuretodifferentsituationsandtheacceptanceandwillingnesstopayforaspecificsystem.Hereitisnecessarytoeithermakesurethatthereishomogeneityamongtheparticipantsortomakesurethatthereisarandomsamplethatrepresentsthedriverpopulationinfocus.

2.1.3Permanentdriverimpairment

Permanentimpairmentscouldberelatedtotheeyesortheearsbutalsotoimpairmentsthatreduceaparticipant’smobilityorcognitivefunctions.Inordertoreducetheriskofconfound,driverswiththistypeofimpairmentshouldnotbeincluded.Ontheotherhand,iftheevaluationhasaninterestin“workingforall”participantswithpermanentimpairmentsthatarecommoninthepopulation（e.g.colourblindness）,theFOTshouldconsiderthese.

2.1.4Temporarydriverimpairment

Insomecases,thesystemunderexaminationcouldbeonethatsupports,forexample,driverfatigue.Itisthencriticaltoselectparticipantsthatdriveunderrelevantconditions,suchasnightshiftworkers,driverswithsleepdisorders,professionaldrivers,commutersandyoungdrivers.Ontheotherhand,thereisariskforconfoundingeffectsifaparticipantisusinglegaldrugswhichcaninducetemporaryimpairmentsandaffectthedatacollected.

2.2Drivingexperience

WhenconductinganFOT,itisalsoimportanttotakeintoaccountthedrivingexperienceoftheparticipants.Specifically,whentheFOTincludesintelligentvehiclesystemsorin-cardevices,suchasAdaptiveCruiseControl（ACC）,brakeassistingsystemsorlanedeparturewarningsystems,itisimportanttobespecificaboutwhichparticipantstoinvite.

2.2.1Overalldrivingexperience

Thedefinitionofthedrivingexperiencerecommendedfortheparticipantsdependsofthetestedsystems.Forexample,whentestingarelativelynewin-cardeviceforusability,itisimportantthatparticipantshavenoexperiencewiththesystem.However,drivingexperienceingeneralandwithotherdevicescanbequiteimportant,asinterferencewiththedrivingtaskmayoccur（whichcouldresultinunreliabletestresultsandevenhigheraccidentriskwithinexperienceddriversduetodistraction）.Ithasbeenshownthatnovicedrivershavehighercognitiveworkloadlevelsthanexperienceddrivers.Thislattergroupitbetterabletoautomatethedrivingtask,whichcausesthemto‘saveattention’morethantheinexperienceddrivers.Thus,itcanbesaidthatwhenatestconcernsin-cardevicesthatmaycausedistractionfromtheprimarydrivingtask,participantsshouldbeexperienced,i.e.moderatetohigh-mileagedrivers.

2.2.2Experiencewithvarioussystems

WhenitcomestoanFOTthatisbeingconductedtoinvestigatethepossibleeffectsofasystemlikeACC,itcanbeinterestingtoinvitedriversthatalreadyhavesomeexperiencewiththesystem.Thesedriversaremostlikelytoperformthedrivingtaskwithouthavingtogetusedtoitfirst.Otherwise,driversthatdonothaveexperiencewithasimilarsystemmightinfluencethedatabecausetheywillnotbedrivingastheywouldnormallydo.Inanycase,thelearningeffectsthatmightbeobservedinanFOTareusefultounderstandtheeaseoflearningofasystem.

2.2.3Accidenthistory

Dependingonthepurposeofthestudy,itcanbeinterestingtoselectdriverswhohaveanaccidenthistoryastheymaydisplaycertaindrivingcharacteristics.Similarly,participantswithoutanaccidenthistorycanalsoprovideinterestinginformation,asacontrolgroup.Accidenthistoryoftheparticipantsneedstobespecifiedinordertodiminishside-effectsandtohaveequalgroups.

2.2.4Timeandroad

Toenabletheresultsofthedifferentparticipantstobecompared,itisimportanttoleteveryparticipant（inatestgroup）drivethesame（orsimilar）routesandatthesametimeoftheday.Forthisaspect,itcanbeimportanttobecriticalinselectingparticipants.Driversthathaveneverdrivenacertainroute（orroadtype）beforemightdisplaymorecarefuldrivingbehaviourthanparticipantsfamiliarwithit.Thismightcausemajordifferencesin,forexample,speedandmostprobablyalsointhecognitiveworkloadoftheparticipants.Thesameisthecasewhenitcomestotimerelatedissues;driversthatareusedtodrivingindaylightmightnotdriveinanaturalwaywhentheyhavetodriveinthedark.

2.3Personalityandattitudes

2.3.1Theinfluenceofpersonalityandattitudesondrivingbehaviour

Giventhelargebodyofliteraturethathassoughttoassociatepersonalitydimensionsandattitudeswithaberrantdrivingbehaviours,manyFOTsincorporateabatteryofpsychometricmeasures.Suchmeasuresaregenerallyincludedinordertorelatepsychologicalfactorstodrivingbehaviour,ratherthanselectparticipantsperse,butwhentestingcertainintelligenttransportsystemstheremaybesomebenefitinbasingrecruitmentonpersonalityandattitudes.Sincedriversexhibitingcertaintraitsorattitudesareknowntoengageinriskierdrivingbehaviours,itwouldseemimportantthatsystemsunderinvestigationinFOTsaretrialledamongstarangeofdriverstoensurethatthesystemsworksforthosewhoneeditmost.Moreover,arecentqualitativetheoryof