Content uploaded by Marco Dozza
Author content
All content in this area was uploaded by Marco Dozza on Dec 20, 2015
Content may be subject to copyright.
Proceedings,InternationalCyclingSafetyConference2013
20‐21November2013,Helmond,TheNetherlands
ABSTRACT
Safetyisaconcernjeopardizingcyclingsustainability.InEurope,morethan2.000cyclistsdie
everyyearintrafficaccidents.Electricalbicyclesamplifythisconcernbecauseoftheirhigh
speedandincreasingprevalence.Previousstudies,suchasthenaturalisticcyclingstudy
BikeSAFER,showedthatinteractionamongcyclistsandotherroadusers(e.g.,drivers,pedes‐
triansorotherbicyclists)iscrucialtocyclingsafety.Aselectricalbicyclesbecomeincreasingly
popular,otherroadusersmayneedtorecalibratetheirexpectationstomaintainasafeinter‐
actionwiththisnewtypeofbicycle.However,theextenttowhichelectricalbicyclesfailto
meetotherroadusers’expectationsbecauseoftheirhigherspeedisyettobeconfirmed.Fur‐
thermore,whetherthegrowingnumberofelectricalbicyclesinfluencescyclingbehaviourand
resultsinnewsafetychallengesisstillunknown.
Thee‐BikeSAFEprojectiscurrentlycollectingnaturalisticcyclingdatainGothenburg.Equipped
bicyclesine‐BikeSAFEpresentlyrecordcyclistbehaviourinrealtrafficfromacameraGPS,and
kinematicssensors.Thisdatawillbeusedto1)understandhowbicyclistswithelectricalbicy‐
clesbehaveintrafficand2)theextenttowhichsafetycriticalsituations(crashandnear‐
crashes)aredifferentforelectricalbicyclescomparedtotraditionalones.Forthislateranaly‐
sis,datafromBikeSAFEwillbeusedasareference.Thispaperexplainshowthenaturalistic
methodologywasadaptedtotheelectricalbicyclesandshowshowe‐BikeSAFEdatawillbe
combinedwithexistingdatafromBikeSAFEforsafetyanalysesincludingcyclingbehaviourand
accidentcausation.
Keywords:trafficsafety,roaduserbehaviour,naturalisticdata,bicycledynamics.
1INTRODUCTION
RenewedinterestforcyclingiscurrentlyraisingimportantsafetyquestionsinseveralEurope‐
ancountries,suchastheNetherlands,Germany,Denmark,andSweden.Inaddition,theintro‐
ductionofnewbicycles,poweredbyanelectricalmotorwiththeabilitytomaintaina25‐km/h
speedindependentlyfromtheweatherandroadgeometry,changescyclingdynamicsandin‐
fluencestheinteractionbetweencyclistsandotherroadusers.Nevertheless,cyclingisavery
appealingactivitypromotedinnumerousEuropeancitiessuchasGothenburg[1]asitde‐
creasescongestionandpollution,whileincreasinghealth[2].
ThenumberofelectricalbicyclesisrapidlyrisinginEurope.IncountriessuchasAustriaand
Germanysalesofelectricalbicycles(pedalecandS‐pedalec)havedoubledfrom2010to2012,
whileinSwitzerlandsalesincreasedby50%andinItalyandFranceby20%.Furthermore,ac‐
cordingtoEurostat,in2012theimportofelectricalbicyclestotalled180.000unitsandcon‐
tributedtoaEuropeanmarketofapproximately1.2millionunits.Theincreasingnumberof
electricalbicyclesontheroadenablesalargernumberofcycliststorideforlongerstretchesof
road,andthusgreatlyincreasesexposuretobicycleaccidents.
e‐BikeSAFE:ANaturalisticCyclingStudytoUnderstandhowElectrical
BicyclesChangeCyclingBehaviourandInfluenceSafety.
M.Dozza1,J.Werneke1,M.Mackenzie1
1AppliedMechanics–VehicleSafety
ChalmersUniversityofTechnology
41296,Gothenburg,Sweden
e‐mail:marco.dozza@chalmers.se
2
Whilewaitingforaccidentdatabasestohavethefinalwordonthesafetyimpactofelectrical
bicyclesinEurope,thee‐BikeSAFEproject[3]startedcollectingnaturalisticcyclingdatafrom
electricalbicyclesinGothenburg.Naturalisticcyclingdataisdatacollectedfrominstrumented
bicyclesriddenbyvolunteersduringtheirdailyactivities.Thisdatacapturesbothcyclingbe‐
haviourandbicycledynamicscontinuouslyandcanprovidein‐depthinformationaboutacci‐
dentcausationandcyclistbehaviourwhichisnotavailableinaccidentdatabases.Infact,acci‐
dentdatabasesonlyincludedatacollectedpost‐crash,thusanyinformationaboutthe
situationbeforethecrashanditscausesisnotmeasuredorrecorded,butcanonlybesimulat‐
edoracquiredfrominterviewsaposteriori.Inaddition,naturalisticdatarecordwithcameras
theinteractionbetweentheegobicycleandtheotherroadusersinthesurroundings.
Theabsolutesafetyimpactofelectricalbicyclesisveryimportantbutnotsufficienttodeter‐
minetheextenttowhichsafetycriticalsituations(crashandnear‐crashes)aredifferentfor
electricalbicyclescomparedtotraditionalonesand,asaconsequence,whetherelectricalbi‐
cyclesaremoreorlesssafethantraditionalones.Toaddresstheseissues,thee‐BikeSAFEpro‐
jectwillcomparethedatacurrentlyundercollectionfromelectricalbicycleswithapre‐existing
datasetofnaturalisticcyclingdatafromtraditionalbicycles[4].Evenifthesizeofthesenatu‐
ralisticcyclingdatasetsisrelativelylimitedandnotcomparabletothepresentnaturalisticdriv‐
ingdatasets[5],thedatacollectedine‐BikeSAFEisuniqueandpromisestoinforminfrastruc‐
tureandbicycledesign,technologicalapplications,andregulationsforsafercycling.
2DATACOLLECTION
Naturalisticcyclingdataiscollectedbyinstrumentedbicyclesriddenbyvolunteersduringtheir
dailyactivities.Instrumentingelectricalbicyclespresentsnewchallengesincomparisonwith
traditionalbicyclesbecausetherearemorewaystouseanelectricalbicyclethanatraditional
one.Infact,anelectricalbicyclecanbestillusedasatraditionalone,differentsettingsenable
thecyclisttodecidehowthebicyclewillbehave,andanelectricalbicyclecanbecommanded
Figure1.A:TraditionalbicycleinstallationfromBikeSAFE.B:Electricalbicycleinstalla‐
tionfrome‐BikeSAFE.
3
bothwithpedals,andtosomeextent,withahandlebarthrottle.Thissectiondescribeshow
datacollectionwasadaptedfromtraditionalbicyclestoelectricalbicycles,includingthenewly
collecteddata’sabilitytocapturetheinteractionandinterplaybetweenthecyclistandthe
electricalbicycle.
2.1Cyclists’recruitment
e‐BikeSAFEaimsatcollectingdatafrom1)thesamecyclistswhoparticipatedinBikeSAFEand
2)duringthesameperiodoftheyear.Thus,allcyclistswhoparticipatedintheBikeSAFEcollec‐
tionwerecontactedandaskedtoparticipatetothee‐BikeSAFEstudyaswell.However,new
volunteerswerealsorecruitedtoguaranteeasufficientlylargesampleofparticipants.Never‐
theless,cyclistswhoparticipatedinBikeSAFEareprioritizedtoenablewithin‐subjectanalyses
possiblyavoidingbiasesfromdifferentbicycleusagesorcyclingperformances.
2.2Installation
PreviousexperiencefrompreBikeSAFE[6]wasthebasisforinstrumentingelectricalbicyclesin
e‐BikeSAFE.Traditionalbicycleinstallationsmadeuseofacustomizablelogger[7]tocollect
datafromavarietyofsensorsincludingGPS,inertialmeasurementunits,brakeforce,and
cameras(Fig.1A).Allcomponentswerefoundofftheshelfwithexceptionofthepressure
brakesensors.However,allsoftwareforoperationoftheloggerandcalibrationofthesensors
wasdevelopedintheBikeSAFEprojectandisdetailedinDozza&Fernandez,2013[8].Thetra‐
ditionalinstallationalsoincludedacircuitryforautomaticstartandstopofthedatalogging
andasimplehumanmachineinterfacecomprisedofanLEDlightandapushbutton.TheLED
lightsignalledthestatusoftheloggerbylightingupduringrecording,thepushbuttonenabled
thecyclisttotimestampeventsduringtheirjourney.Anytimethebuttonwaspressed,atime
stampwassavedintheloggertoenabletheanalysttoquicklyfindthedata(includingvideo)
correspondingtothatpointintimeduringtheanalysis.
Figure2.A:Throttle.B:Pedalrotationsensor.C:Controlunit.D:Brakesensor.
4
2.3Objectivedata
Anelectricalbicycleisnotjustatraditionalbicyclewithanelectricalmotorandabattery;an
electricalbicyclealsoincludesothercomponentssuchasathrottle(tobringthebicycleupto
a6‐km/hspeed;Fig.2A),apedalsensor(whichisusedasadigitalindicatorofthecyclist’swill‐
ingnesstomoveforward;Fig.2B),acontrolunit(enablingthecyclisttochangethebicycle
speed;Fig.2C),andtwobrakeswitches(toautomaticallydisconnectthemotoranytimethe
cyclistbrakes;Fig.2D).Allthesecontrolsenable“novel”cyclingbehavioursandwereconsid‐
eredaspotentialsignalsfordatacollection.AstheinstrumentationfromBikeSAFEalreadyhad
twobrakepressuresensors,thebrakeswitchessignalswerecombined.Thewirefromthepe‐
dalsensorwassplitandconnectedtotheloggersothatthisnewsignalcouldbeusedtode‐
terminethecyclist’sintentiontomoveforward(Fig.1B).Acurrentsensorwasalsoinserted
betweentheelectricalbicyclecontrolunitandthemotorinordertomeasurethepowerthat
themotorwasinstantaneouslyusingtopropelthebicycle(Fig.1B).Asspeediscontinuously
collectednoinformationaboutthecontrolunitsettingswascollected.Further,asthethrottle
isonlyintendedforstartingthebicycleanditsstatuscouldbeinferredbycombiningspeed
withpedalvelocity,noinformationaboutthethrottlepositionwasacquired.Allsignalswere
collectedwitha100HzsampleratewithexceptionfortheGPS(10Hz)andthevideo(30fps).
2.4Subjectivedata
Beforedatacollection,thecyclistswereaskedtosignaconsentformandtofillinaquestion‐
nairecovering1)demographicdata(e.g.,age,gender,occupationalfield,typeofbicyclethey
mainlyride,etc.)2)theircyclingbehaviourandusageoftheirbicycles(e.g.,overtheyear,for
whichpurpose,helmetuse,etc.),aswellas3)theiropiniontowardselectricbicycles,inde‐
pendentlyiftheyalreadyhaveordon’thaveexperiencewithelectricbicycles(e.g.,levelof
agreementtostatementssuchas,e‐bikesareratherforolderpeople).Thisquestionnairewas
againinspiredbytheBikeSAFEproject.Duringcollection,thecyclistswereaskedtofillinatrip
diary,detailingtimeandpurposeforeachtriptakenwiththeelectricalbicyclesimilarlyto
BikeSAFE.Anytimethecyclistsexperienceasafetycriticalsituation,i.e.asituationwhich
madethemfeelsafetyuncomfortable,theywereaskedtopressthepushbuttononthebicycle
handlebarandtoincludethisinthetripdiary.Additionally,thecyclistswereaskedtofillina
protocolforeachsafetycriticaleventinordertodescribethesituationindetail.Furthermore,
theprotocolincludedquestionsconcerningtheexpectationoftheupcomingevent,howcom‐
fortabletheyfeltinthesituation,anyvisualocclusions,andalsotheroad,lightandweather
conditions.Afterthetwo‐weeksofdatacollection,thecyclistswereinterviewedconcerning
eachsafetycriticaleventwithananalystfromtheproject,basedontheprotocols.Duringthe
interviewthetimestampsfromthepushbuttonwereusedtopresentthecyclistwiththevid‐
eocorrespondingtothesafetycriticalsituationunderdiscussion.Finally,allcyclistsalsofilled
inaquestionnairetoreportontheirimpressionsofridingtheelectricalbicycle.Thisquestion‐
nairewasdevelopedine‐BikeSAFEtodeterminetheextenttowhichcyclistsfeltsafewhilerid‐
ingtheelectricalbicycle.
3DATAANALYSIS
Oneoftheadvantagesofnaturalisticdatasetsisthepossibilitytoreusethesamedatasetfor
differenttypesofanalysis.Further,bycombiningnewandexistingdatasetsnovelanalysescan
beenabled.Thisisindeedthecaseine‐BikeSAFEwherecomparisonswiththeexisting
BikeSAFEdatasetarethebasistodeterminewhetherelectricalbicyclesaremoreorlesssafe
thantraditionalones.
3.1BikeSAFEdataset
TheBikeSAFEdatasetwascollectedin2012fromfiveinstrumentedbicycles(Fig.1A)and20
cyclistswhotookturnsinridingthebicyclesfor2weeks.Allbicyclistsalsocompiledatripdia‐
ry,twoquestionnaires(onebeforeandoneafterdatacollection)andunderwentaninterview
afterthetwoweeksofdatacollection.Duringtheinterview,thecyclistsdiscussedwithanana‐
5
lystthesafetycriticalsituationstheyencountered.Intotal,332trips,covering1549kmover
114hourswerecollected.Averagespeedwas13.6km/h(SD=3.2km/h),andaveragetripdu‐
rationwas22min(SD=8.6min).Intotal,63criticalsituationswerereportedbythepartici‐
pantsduringcollection.Sixofthesecriticalsituationswereconsideredascrashesastheyin‐
volvedsomephysicalcontact,inthreecriticalsituationsthecyclistfelloffthebicycle.All
criticalsituationswerereviewedbytwoanalystsandcodedaccordingtotheBikeSAFEcoding
scheme[9].ObjectivedataintheBikeSAFEdatasetincludeddatafromthesensorsinFig.1A.
Alldatawascollectedwitha100HzsampleratewithexceptionfortheGPS(10Hz)andthe
video(30fps).
3.2Cyclingbehaviour
Cyclingbehaviourcanbeassessedinseveralways,oneofwhichisbymeasuringbicycledy‐
namics.Forinstance,distributionsofspeedanddistributionsofaccelerationsanddecelera‐
tionsoverspeedintervalscanidentifythecyclistcomfortboundaries.Acomparisonofspeed
distributionbetweentraditionalbicyclesandelectricalbicyclescanhelpindeterminingthepo‐
tentialimpactonsafety,ashigherspeedhasbeenshowntoleadtomoreandmoresevere
crashesformotorizedvehicles[10].Figure3showsthespeeddistributionfromBikeSAFEin
comparisonwithapreliminarydistributionfrome‐BikeSAFElimitedtoonlyonecyclistand
suggestingelectricalbicyclesbeingalmost10km/hfasteronaveragethantraditionalones.
Figure3.SpeeddistributionsfromBikeSAFE(left)ande‐BikeSAFE(onlyonesubject;
right).
Figure4.ObediencetotrafficrulesfromBikeSAFER.
6
Byfurthercomparingthecomfortboundariesofelectricalbicyclesandtraditionalbicyclese‐
BikeSAFEwilldeterminetheextenttowhichelectricalbicyclesaremoreorlessmanoeuvrable
thantraditionalbicyclesindifferentsituations(forexampleatintersections)andgiveanindi‐
cationofwhetherandhowcyclinglanesshouldbeadaptedtoelectricalbicycles(forinstance
intermsofminimumwidthandcurvature)topreservemanoeuvrabilityandvisibilityofthecy‐
clistathigherspeed.
Naturalisticdatacanalsobeusedasobservationaldata(i.e.datacollectedbyanalystsobserv‐
ingroadusersbehaviourinrealtraffictodeterminetheextenttowhichcyclistsobeytraffic
rulesandrecommendations.Figure4showssomeresultsfromBikeSAFER[11]wheretheuse
ofhelmets,properlightingwhendark,andprevalenceofintersectionspassedwitharedlight
wereestimatedfromvideoclipsrandomlypickedfromthewholedataset.Thesameanalysis
willbecarriedonine‐BikeSAFEtodeterminetheextenttowhichelectricalbicyclesaremore
orlesspronethantraditionalbicyclestopassaredlight.
3.3Accidentcausation
Asaccidentprobabilityisrelatedtospeed[10],comparingspeedacrossgeographicalcoordi‐
natesforelectricalandtraditionalbicycleswillbeanimportantstartpointforthee‐BikeSAFE
safetyanalysis.BysubtractingfromtheBikeSAFEspeedmapinFigure5anewe‐BikeSAFE
speedmap,itwillbepossibletovisualizeandlistthelocationswherethedifferenceinspeed
betweenelectricalandtraditionalbicyclesislarge.Thislist,oncerankedaccordingtospeed
andprevalenceofcyclingaccidentinSTRADA,willgeographicallyfocusoursafetyanalysisand
helpdefininglimitsinthecurrentcyclinginfrastructuresdependingonthegeometricalfea‐
tures(suchascurvature,slope,andproximitytointersections)ofthelistedlocations.
Safetycriticaleventsfrome‐BikeSAFEwillbereviewedandcodedsimilarlyasinBikeSAFE[9].
ByplottingtheGPScoordinatesofeachsafetycriticalevent,itwillbepossibletocreateamap
similartotheonefromBikeSAFEinFigure6.Further,randombaselineeventswillbeextracted
frome‐BikeSAFEsimilarlyasinBikeSAFE.Bycomparingthefourgroupsofevents(baselineand
criticalevents,fromelectricalandtraditionalbicycles)e‐BikeSAFEwilldeterminetheextentto
whichsafetycriticalsituationsaredifferentforelectricalbicyclescomparedtotraditionalones,
intermsofthedifferentcodingvariables.Forinstancecomparisonbetweencriticalsituations
Figure5.–GeographicaldistributionofaveragespeedfromBikeSAFER.
7
mayshowwhetherelectricalbicyclesarelesssafeatintersections.Furthercomparisonof
baselineswillshowwhetherelectricalbicyclesandtraditionalbicyclesareexposedtodifferent
situations.Thislateranalysiswillshow,forinstance,whethertheinteractionwithotherroad
usersismoreprevalentwhenridinganelectricalbicyclethanwhenridingatraditionalbicycle,
whichweexpectasaconsequenceofthehigherspeed.
Theaimoftheanalysisofthesubjectivedataine‐BikeSAFEwillbesimilartotheoneinthe
BikeSAFEproject.Thegoalwillbetoinvestigatecyclists’perceptionofsafety‐criticalsituations
whenridinganelectricalbicycle.Acomparisonofthesafety‐criticaleventsfoundinBikeSAFE
withtheonesine‐BikeSAFEwillmakepossibletoexaminehow1)thefrequencyofsafety‐
criticaleventsand2)thetypeofconflictencounteredwilldistinguishbetweencyclingwithan
electricalbicyclecomparedtoatraditionalone.Furthermore,withtheinterviewprotocolused
ine‐BikeSAFEandusingthevideoscenewhileinterviewingthecyclistitispossibletoprovidea
muchmoredetailedinsightintosafety‐criticalsituationsfromacyclist’sperspectivewhichis
notavailablefromotherinformationsources,suchasaccidentdatabases,cyclingsimulators,
andon‐roadtesting.Completingtheinterviewprotocolattheendofthetwoweeksofdata
collectionwillalsoensureamuchbetterrecollectionfromthecyclist.
4DISCUSSION
Asthee‐BikeSAFEprojectiscurrentlycollectingdata,thissectionfocusesonthelessons
learnedsofarwhileadaptingthenaturalisticmethodologytoelectricalbicycles.
4.1Datacollection
Thee‐BikeSAFEprojectiscurrentlycollectingdata,prioritizingvolunteerswhoalsoparticipat‐
edtotheBikeSAFEcollection.Whilesuchprioritizationseemslegitimateasitwillproducea
datasetenablingwithin‐subjectanalyses,itmayalsocarryonpotentialsubjects’biasesfrom
thepreviouscollection.AnalysisofsubjectivedatafromBikeSAFEdidnotshowthepresence
ofanyoutliersuggestingthattheparticipantsinBikeSAFEwereindeedareasonablerepresen‐
tationofthecyclingcommunityinGothenburg.Nevertheless,atthepresenttime,femalecy‐
Figure6.BaselineandsafetycriticaleventsfromBikeSAFER.
8
clistswhoparticipatedtoBikeSAFEseemmorepronetoparticipatetoe‐BikeSAFEthanmale
cyclists,thuspotentiallygenderbiasingthenewdataset.
Theinstallationofelectricalbicyclespresentedseveralnewchallengeswhichwereunex‐
pected.Acquiringsignalsfromthebicyclecontrolunitrequiredadeepunderstandingofthe
electricalbicyclehardwareandwiring.Further,poweringtheBikeSAFEloggerfromthebicycle
batterywasconvenientbutalsocreatednewpossiblemisuseofthesystemasthecyclistispo‐
tentiallyabletoturntheloggerdownwhileoperational.
Newsignalsfromthepedalsensorsandthecurrentsensorsneededtobeopportunelypre‐
processedbeforetheycouldbeused.Itisworthnotingthatcorrectlyacquiringinformation
aboutpedalspeedisnotpossiblewitha10Hzfrequency,whichistheusualstandardfornatu‐
ralisticdatacollection,whilethe100Hzsamplingfrequencyusedine‐BikeSAFEwasdefinitely
sufficient.Forthecurrentsensorthemainissuewasinsteadcalibration,howeverthismaybea
sensorspecific(Phidgets1122CurrentSensor)issue.InSweden,thehandlebarthrottlefor
electricalbicyclescanonlybringabicycleupto6km/h,withoutpedalinput,becauseoflegal
requirements.However,whenbuyinganelectricalbicycle,thecustomerisinformedthatby
removingasmallconnectorfromthecontrolbox,the6‐km/hlimitationwillbedisabledand
thethrottlewillthenenablethebicycletoreach25km/h.Bycombiningthepedalsensordata
withthecurrentsensordata,itwillbepossibletodeterminewhetheranyofthebicyclistsde‐
cidedtotakeawaytheconnectorandoperatethebicycleillegally.However,byconnectingthe
throttledirectlytoourlogger,itwouldbemucheasiertoestablishhowthethrottleisactually
used.Ife‐BikeSAFEwouldprovethatsomebicyclistsareindeedusingthehandlebarthrottle
beyondregulations,futurecollectionshoulddefinitelylogthethrottlesignal.
4.2Dataanalysis
Themainconcernfordataanalysisinthee‐BikeSAFEprojectisaboutstatisticalsignificanceof
theresults.BikeSAFEprovedthat100hoursofcyclingaresufficienttoshowstatisticalsignifi‐
canceforsomeofthemainfactorscontributingtocriticalsituations.However,suchadataset
wasnotalwaysabletoshowsignificanceforrelativelysporadiccircumstances.Forinstance,
thepresenceofconstructionworksonthebicyclelaneseemedtohaveanimpactofthelikeli‐
hoodtoencounterasafetycriticalsituation,however,theconfidenceintervalsfromthe
BikeSAFEdatasetweretoowidetoclaimsignificanceforsucharesult.Itispresentlynotclear
whethere‐BikeSAFEwillbeabletoproduceadatasetcomparabletoBikeSAFE.Inpartbecause
oftheverylimitedbudgetofe‐BikeSAFE,whichresultedinalower(3vs.5)numberofinstru‐
mentedbicycles.However,iftheelectricalbicycleswoulddrivetheparticipantstoridelonger
durationsthantheydidinBikeSAFE,thee‐BikeSAFEdatasetmaybecomeaslargeasthe
BikeSAFEonedespitealowernumberofcyclists.
Preliminaryresultssuggestthatelectricalbicyclesareindeedfasterthantraditionalbicyclesin
traffic.However,thisresultisstilltobeconfirmedanditssafetyimplicationstobedefined.In
thisrespect,comparisonswithpreviousstudiessuchastheonespresentedinGehlertetal.,
2011[12]andSEEKING[13]willhelpdeterminewhethersafetyconcernsaboutelectricalbicy‐
clesareindeedfundedandtheextenttowhichtheseconcernsmaybedifferentindifferent
Europeancountries.Inanycase,ridingfasterincreasescyclists’attentiondemandasinterac‐
tionwithotherroadusers,suchasovertakingmanoeuvres,becomesmorefrequent.Further,
electricalbicyclesarenotalwaysclearlydistinguishablefromtraditionalones.Asaconse‐
quence,itmaynotbeobviousforotherroadusers,(e.g.adriveratanintersection),toesti‐
matetheirspeed(e.g.whendecidingtocrossthebicyclepath).
5CONCLUSIONS
ElectricalbicyclesareincreasinglypopularinEuropeandpromotedbythelocalauthoritiesin
severalEuropeancountries.Safetyconcernsaboutelectricalbicyclesarelegitimateasthese
newbicyclesareoperateddifferentlyandexhibitdifferentdynamicscomparedtotraditional
bicycles.Atimelywaytodeterminethesafetyimpactofelectricalbicyclesisbyusingnatural‐
9
isticcyclingdataasinthee‐BikeSAFEproject,whereinstrumentedbicyclescontinuouslycol‐
lectdatafromacameraandothersensorsintraffic.
Electricalbicyclesaremorecomplexthantraditionalbicyclescreatingnewchallengesfornatu‐
ralisticdatacollectionandanalysis.Infact,electricalbicyclespresentnewcommandsandpro‐
pulsionpossibilities.Thisinformationneedstobeloggedtodeterminetheextenttowhichcy‐
clistsareactuallytakingadvantagefromtheelectricalmotorandthewaycyclistsinteractwith
theelectricalpropulsionsystem.
e‐BikeSAFEanalysiswillleverageontheexistingBikeSAFEdataset(anaturalisticcyclingda‐
tasetfromtraditionalbicycles)todeterminethenetsafetyimpactofelectricalbicycles.Anal‐
yseswillfocusoncyclingbehaviourandaccidentcausationandwillbeguidedbythecentral
hypothesisthatlocationswherethevelocityofelectricalbicyclesdifferthemostfromtradi‐
tionalbicyclesarethemostrelevantforsafety.
ACKNOWLEDGEMENTS
TrafikverketandtheSwedishStrategicFundsforTransportationarecurrentlysponsoringthe
projectpresentedinthispaper.
REFERENCES
[1] G. S. Trafikkontoret, "Cykelåret 2012," 2012.
[2] J. J. de Hartog, H. Boogaard, H. Nijland, and G. Hoek, "Do the Health Benefits of
Cycling Outweigh the Risks?," Environmental Health Perspectives, vol. 118, pp. 1109-
1116, Aug 2010.
[3] SAFER. (2013). e-BikeSAFE. Available:
http://www.chalmers.se/safer/EN/projects/traffic-safety-analysis/associated-projects/e-
bikesafe
[4] M. Dozza and A. Fernandez, "Understanding Bicycle Dynamics and Cyclist Behavior
from Naturalistic Field Data," IEEE Transactions on Intelligent Transportation
Systems, vol. (accepted for publication), 2013.
[5] K. L. Campbell, "The SHRP 2 Naturalistic Driving Study: Addressing Driver
Performance and Behavior in Traffic Safety. ," in TR News, 282., ed, 2013.
[6] M. Dozza, "preBikeSAFE (TRV 2011 86383) Slutrapport -Forskningsrapporter
(Tillämpad mekanik) ISSN 1652-8549; nr 2012:12," Chalmers University of
Technology 2012.
[7] M. Dozza, T. Andersson, M. Idegren, and A. Fernandez, "A Platform Enabling
Intelligent Safety Applications for Vulnerable Road Users," IET Intelligent Transport
Systems (in press), 2013.
[8] M. Dozza and A. Fernandez, "Understanding Bicycle Dynamics and Cyclist Behavior
from Naturalistic Field Data," IEEE Transactions on Intelligent Transportation
Systems, vol. accepted, 2013.
[9] M. Dozza and J. Werneke, "Introducing naturalistic cycling data: What factors
influence bicyclists’ safety in the real world?," Transportation Research Part F-Traffic
Psychology and Behaviour - accepted pending revisions, 2013.
[10] R. Elvik, "A re-parameterisation of the Power Model of the relationship between the
speed of traffic and the number of accidents and accident victims," Accid Anal Prev,
vol. 50, pp. 854-60, Jan 2013.
[11] M. Dozza, "BikeSAFER: Final Report -Forskningsrapporter (Tillämpad mekanik) ISSN
1652-8549; nr 2013:04," Chalmers University of Technology 2013.
[12] T. Gehlert, M. Kuhn, K. Schleinitz, T. Petzoldt, S. Schwanitz, and R. Gerike, "The
german pedelec naturalistic cycling study - study design and first experiences,"
presented at the Proceedings of the first International Conference on Cycling Safety,
Helmond, 7-9 November 2012, 2012.
10
[13] (2013). SEEKING project. Available:
http://www.bmvit.gv.at/ministerium/staatspreis/mobilitaet2011/nominierte/seeking.html