Risk Riding Behaviors of Urban E-Bikes: A Literature Review

Abstract

In order to clearly understand the risky riding behaviors of electric bicycles (e-bikes) and analyze the riding characteristics, we review the research results of the e-bike risky riding behavior from three aspects: the characteristics and causes of e-bike accidents, the characteristics of users' traffic behavior, and the prevention and intervention of traffic accidents. The analysis results show that the existing research methods on risky riding behavior of e-bikes mainly involve questionnaire survey methods, structural equation models, and binary probability models. The illegal occupation of motor vehicle lanes, over-speed cycling, red-light running, and illegal manned and reverse cycling are the main risky riding behaviors seen with e-bikes. Due to the difference in physiological and psychological characteristics such as gender, age, audiovisual ability, responsiveness, patience when waiting for a red light, congregation, etc., there are differences in risky cycling behaviors of different users. Accident prevention measures, such as uniform registration of licenses, the implementation of quasi-drive systems, improvements of the riding environment, enhancements of safety awareness and training, are considered effective measures for preventing e-bike accidents and protecting the traffic safety of users. Finally, in view of the shortcomings of the current research, the authors point out three research directions that can be further explored in the future. The strong association rules between risky riding behavior and traffic accidents should be explored using big data analysis. The relationships between risk awareness, risky cycling, and traffic accidents should be studied using the scales of risk perception, risk attitude, and risk tolerance. In a variety of complex mixed scenes, the risk degree, coupling characteristics, interventions, and the coupling effects of various combination intervention measures of e-bike riding behaviors should be researched using coupling theory in the future.

Full text



Int.J.Environ.Res.PublicHealth2019,16,2308;doi:10.3390/ijerph16132308www.mdpi.com/journal/ijerph
Review
RiskRidingBehaviorsofUrbanE‐Bikes:
ALiteratureReview
ChangxiMa1,*,DongYang1,JibiaoZhou2,*,ZhongxiangFeng3andQuanYuan4
1SchoolofTrafficandTransportation,LanzhouJiaotongUniversity,Lanzhou730070,China;
dongyang26@126.com
2SchoolofCivilandTransportationEngineering,NingboUniversityofTechnology,Ningbo315211,China
3SchoolofAutomotiveandTrafficEngineering,HefeiUniversityofTechnology,Hefei230009,China;
fzx@hfut.edu.cn
4StateKeyLaboratoryofAutomotiveSafetyandEnergy,TsinghuaUniversity,Beijing100084,China;
yuanq@tsinghua.edu.cn
*Correspondence:machangxi@mail.lzjtu.cn(C.M.);zhoujb2014@nbut.edu.cn(J.Z.);
Tel.:+86‐131‐0942‐9716(C.M.);Tel:+86‐188‐1527‐6878(J.Z.)
Received:12May2019;Accepted:26June2019;Published:28June2019
Abstract:Inordertoclearlyunderstandtheriskyridingbehaviorsofelectricbicycles(e‐bikes)and
analyzetheridingcharacteristics,wereviewtheresearchresultsofthee‐bikeriskyridingbehavior
fromthreeaspects:thecharacteristicsandcausesofe‐bikeaccidents,thecharacteristicsofusers’
trafficbehavior,andthepreventionandinterventionoftrafficaccidents.Theanalysisresultsshow
thattheexistingresearchmethodsonriskyridingbehaviorofe‐bikesmainlyinvolvequestionnaire
surveymethods,structuralequationmodels,andbinaryprobabilitymodels.Theillegaloccupation
ofmotorvehiclelanes,over‐speedcycling,red‐lightrunning,andillegalmannedandreverse
cyclingarethemainriskyridingbehaviorsseenwithe‐bikes.Duetothedifferenceinphysiological
andpsychologicalcharacteristicssuchasgender,age,audiovisualability,responsiveness,patience
whenwaitingforaredlight,congregation,etc.,therearedifferencesinriskycyclingbehaviorsof
differentusers.Accidentpreventionmeasures,suchasuniformregistrationoflicenses,the
implementationofquasi‐drivesystems,improvementsoftheridingenvironment,enhancementsof
safetyawarenessandtraining,areconsideredeffectivemeasuresforpreventinge‐bikeaccidents
andprotectingthetrafficsafetyofusers.Finally,inviewoftheshortcomingsofthecurrent
research,theauthorspointoutthreeresearchdirectionsthatcanbefurtherexploredinthefuture.
Thestrongassociationrulesbetweenriskyridingbehaviorandtrafficaccidentsshouldbe
exploredusingbigdataanalysis.Therelationshipsbetweenriskawareness,riskycycling,and
trafficaccidentsshouldbestudiedusingthescalesofriskperception,riskattitude,andrisk
tolerance.Inavarietyofcomplexmixedscenes,theriskdegree,couplingcharacteristics,
interventions,andthecouplingeffectsofvariouscombinationinterventionmeasuresofe‐bike
ridingbehaviorsshouldberesearchedusingcouplingtheoryinthefuture.
Keywords:trafficengineering;e‐bikes;riskyridingbehavior;trafficaccidents;interventions

1.Introduction
Inrecentyears,electricbicycles(e‐bikes)becamethebestchoicefordailytravelofsome
residentsinlargeandmedium‐sizedcitiesinChinaduetotheirlowprice,convenience,and
flexibility.UnlikeinNorthAmericaandEurope,e‐bikesarethemaintrafficmodeinmanyof
China’smajorcitiesandareusedprimarilyforcommutingratherthansimplyforleisure.According
tothestatisticsoftheChinesecyclingassociation[1],in2017,thetotalnumberofe‐bikesinChina
was250million,theoutputofe‐bikeswas30.97million,andtheexportwas7.301million,withan

Int.J.Environ.Res.PublicHealth2019,16,23082of18
exportvalueof$1.44billion.InNanning,Haikou,Kunming,Guilin,andothercities,e‐bikesfar
outnumberbicycles;Nanninghasmorethan1.8millione‐bikes[2],whichisknownas“thecityof
electricbicycles”,duetoithavingthelargestnumberofe‐bikesinthecountry.Therefore,e‐bikesare
currentlyoneofthemostimportantmeansofcommutertransportation[3].
Despitetheobviousadvantages,therapidgrowthofe‐bikesalsocausesaseriesofsafety
problems.Liketraditionalbicyclesandpedestrians,e‐bikesalsobelongtothecategoryofvulnerable
groupsontheroad.Duetotheirfastspeed,e‐bikeshavemoreseriousaccidentrisks.Accordingto
thestatisticalannualreportofChina’sroadtrafficaccidentsin2015,thenumberofe‐bikeaccidents
was8.2timeslargerthanthatofbicycleaccidentsand5.4timeslargerthanthatofpedestrian
accidents[4].FromJanuarytoJune2016,thenumberofe‐bikeaccidentsaccountedfornearly70%of
thetotalnumberofaccidentsinJiangsuProvince[5].Thehospitaldataarenotoptimisticeither.The
hospitalizationrecordsofe‐bikeusersinHefeifrom2009to2011showthatone‐thirdofe‐bikeusers
wereseriouslyinjured[6].AccordingtothehospitalizationrecordsofSuzhoufromOctober2010to
April2011,thenumberofinjurede‐bikeusersaccountedfor57.2%oftherateofroadtraffic
hospitalization[7].Inadditiontotheseverityofaccidents,thenumberofe‐bikeaccidentsalsoshows
atrendofcontinuousgrowth.Accordingtothestatisticaldata[8],thenumbersofe‐biketraffic
accidentsin2011and2016were10,347and17,747,respectively,andthenumberofdeathsincreased
by71.52%inthefiveyears.Thenumbersofe‐bikeinjurieswere11,381and19,678,respectively,
increasingby72.90%.Basedonthecasesofinjuriesorcasualtiesoftwo‐wheeledvehiclesinfive
citiesinChinafromJuly2011–June2016,electricbicycleaccidentswereacommontypeof
two‐wheeledvehicleaccidentsinChina,accountingfor34.79%ofthetotalnumber.Amongthese
accidentsinvolvingelectricbicycles,thosecausingminorinjuriestotheridersaccountedfor70.0%,
whiletheproportionofseriousinjurieswas12.6%,andtheproportionofdeathswas10.6%[9].Due
tothefrequentoccurrenceandseverityofe‐bikeaccidents,citieslikeGuangzhou,Shenzhen,
Wenzhou,andFuzhoubannedorrestrictedtheuseofe‐bikes[6].Atthesametime,intermsof
nationallawsandregulations,relevantprovisionswerealsoformulatedtorestrainillegalbehaviors.
Forexample,Article70oftheregulationsfortheimplementationoftheroadtrafficsafetylawofthe
People’sRepublicofChinastipulatesthat“whenridingabicycle,anelectricbicycle,oratricycleand
crossingamotorvehiclelaneonaroad,theridershouldgetoffthevehicleandcarryit.Ifthereisno
crosswalkorpedestriancrossingfacilities,orifitisinconvenienttousethem,theridershouldgo
straightthroughafterconfirmingsafety.”

Figure1.Keywordco‐occurrencenetworkofelectricbicycle(e‐bike)safetystudies.

Int.J.Environ.Res.PublicHealth2019,16,23083of18
Numerousstudiesshowedthathumanfactors,especiallythebehaviorofe‐bikeusers(Figure
1),areimportantinmosttrafficaccidents,asisthecasewithe‐bikes.Figure1showsthekeyword
co‐occurrencenetworkofe‐bikesafetystudies,inwhichwefoundthatpreviousstudiesmainly
focusedonbehavior,safety,risk,crashes,choice,andsoon.E‐biketrafficviolationsmainlyinvolvea
violationoftrafficsignals,aviolationofregulationsonmannedvehicles,afailuretodrivein
non‐motorizedlanes,adversedriving,etc.[10].Forexample,Zhangetal.[11]foundthatthe
violationoftrafficsignalsbye‐bikesisoneofthemaincausesofroadtrafficaccidents,accounting
for54.18%ofthetotalaccidentdata.CherryandDuetal.[12,13]analyzedtheviolationsofe‐bike
usersandfoundthatred‐lightrunning,over‐speeding,andoverloadingwerethemaincausesof
roadtrafficaccidents.Demographicvariables,socialcognition,andotherfactorsofe‐bikeusersare
alsocloselyrelatedtotheoccurrenceoftrafficaccidents.Huetal.[6]analyzedtheinfluencingfactors
ofe‐bikeaccidentsandfoundthatage,gender,andvehicletypehadasignificantimpact.Yaoand
Wu[14]establishedtherelationshipbetweensafetyattitude,riskperception,andviolationsofe‐bike
users,andtheresultsshowedthatbothgenderanddrivingexperiencehadasignificantimpacton
e‐bikeaccidents.Papoutsietal.[15]analyzedtheage,gender,accidenttime,andaccidentcauseof
e‐bikeusersusingtheaccidentdataofahospitalinSwitzerland.Guoetal.[16]foundthate‐bikesare
morelikelytobeinvolvedinred‐lightrunningthanordinarybikes.Zhouetal.[17]empirically
analyzedthesignificantfactorsaffectingaccidentsinvolvinge‐bikesandtheuseoflicenseplates,
andfoundthattheuseoflicenseplatesone‐bikesreducedthepossibilityofaccidents.Guoetal.[18]
foundfromamodelcomparisonanalysisthatthecyclist’sgender,cyclingbehavior,typeofe‐bike,
speedlimit,andotherfactorshadasignificantimpactontheseverityofe‐bikecollisions.Basedon
thepreviousstudies,thevisualizationofriskyridingbehaviorsofe‐bikescanbefoundthrough
collaborationnetworkanalysis,asshowninFigure2.InFigure2,eachnoderepresentsanauthoror
anorganization,andthenodesizesindicatethenumberofpublishedpapers.Thelinksbetween
nodesrepresentthecollaborations,wherebythegreaterwidthofalinkrepresentsacloser
collaboration.Specifically,thecitationnetworkamongproductiveauthorsisshowninFigure2a,the
co‐authorshipnetworkamongproductiveauthorsisshowninFigure2b,andthecollaboration
networkamongresearchinstitutionsisshowninFigure2c.

(a)


Int.J.Environ.Res.PublicHealth2019,16,23084of18
(b)

(c)
Figure2.Visualizationofriskyridingbehaviorsine‐bikestudies:(a)thecitationnetworkamong
productiveauthors;(b)theco‐authorshipnetworkamongproductiveauthors;(c)thecollaboration
networkamongresearchinstitutions.
Withtheincreasingattentionpaidtotheproblemsdiscussedabove,itisurgenttobetter
understandtheriskyridingbehaviorofe‐bikesandthephysiologicalandpsychological
characteristicsofe‐bikeusers,soastoreducetheaccidentrateandimprovethesafetyawarenessof
e‐bikeusers.Atthesametime,onthebasisofstudiesontheriskyridingbehaviorofe‐bikes,itisof
greatsignificancetoimproveroadtrafficsafetyandreducetrafficaccidentsbyfurtheranalyzingthe
trafficaccidentcharacteristicsandtrafficaccidentcausesofe‐bikes.
2.AnalysisofRiskyRidingBehaviorCharacteristicsandInfluencingFactorsofE‐Bikes
2.1.DataAcquisitionandAnalysisMethods
1. Intermsofdatacollectionofriskyridingbehaviorofe‐bikes,questionnairesurveymethods
[19–25]andvideocollectionmethods[26–28]aremainlyadopted.Thequestionnaireinthe
questionnairesurveymethodwasdesignedbasedonpreviousstudiesonthedrivingbehavior
ofmopeds,motorcycleusers,andautomobiledrivers.Mostoftheuseditemswereselectedfrom
themopeduserbehaviorquestionnairedesignedbyStegetal.[19],themotorcycledriver
behaviorquestionnairedesignedbyElliottetal.[20],andtheChinesecyclingbehavior
questionnairedesignedbyXieandParker[21].Inthequestionnairedesign,theitemswhich
appliedtoe‐bikeusersorwhichcouldbemodifiedwereretained,whiletherestwerediscarded,
andnewfeaturesofe‐bikeridingwereadded.Respondentswereaskedtoratethefrequencyof
eachridebehavioronafive‐pointLikertscale,rangingfrom“never”(1)to“almostalways”(5).
Severaltypicale‐bikeuserswerepre‐tested,andthequestionnairewasrevisedaccordingto
theirfeedbacktoimprovetheclarityandreadability.
Thequestionnairesurveyapproachiswidelyusedintrafficsafetyresearchtogather
informationsuchasdrivingbehavior,safetyattitude,andriskperception[17,19–24].Forexample,
YaoandWu[14]studiedtheriskfactorsaffectingtheparticipationofe‐bikeusersinaccidentsbased
onthequestionnairesurveymethod,anddeterminedtherelationshipbetweensafetyattitude,risk
perception,andabnormalridingbehavior.Thesurveyincluded603e‐bikeusersinBeijingand
Hangzhou.Theresultsshowedthatgenderanddrivingexperienceweresignificantlyrelatedto
trafficaccidents.Menweremorelikelytohaveaccidentsthanwomen,andcyclistswithadriver’s
licensewerelesslikelytohaveaccidentsthanthosewithoutone.Thestudyalsofoundthattwo
typesofabnormalcyclingbehaviors,errorsandaggressivebehavior,wereimportantfactorsin
predictingtrafficaccidents.Inordertostudythemechanismandcauseofelectricbicyclecollisions,

Int.J.Environ.Res.PublicHealth2019,16,23085of18
Hertachetal.[29]usedlogisticregressionmodeltoanalyzethequestionnairedataof3659Swiss
e‐bikeriders.Itwasfoundthatabout17%ofelectricbikeridershadtrafficaccidents.Thetypesof
accidentsmostlyinvolvedslipping,fallingwhencrossingthethreshold,andslippingwhen
enteringatram/railway.Thecausesofaccidentsincludedtheroadbeingtooslippery,theriding
speedbeingtoofast,alossofbalancewhileriding,etc.Inordertounderstandtheawareness,
cyclingbehavior,andlegislativeattitudeofe‐bikeridersandnone‐bikeroadcyclistsinTianjin,
Wangetal.[30]conductedcomparativeanalysisandresearchonthetwotypesofcycliststhrougha
largenumberofquestionnairesurveysandinterviews.Theresultsshowedthat74.2%ofe‐bike
ridersthoughtitwasnecessarytowearhelmets,and54.7%ofe‐bikeridersthoughtwindshield
installationwaswrong,whichwashigherthanotherroadusers(49.1%and48.4%,respectively).
However,infieldsurveys,e‐bikeriderawarenessofvariousviolationslaggedfarbehindwhatis
right.Guoetal.[31]investigatedandinterviewed884e‐bikeridersinNanjingtostudythe
influencingfactorsofe‐bikeregistrationinChina.Theresultsshowedthatregistrationwas
influencedbygender,age,education,drivinglicense,familycarownership,familyincome,and
electricbicycletravelfrequency.
Reasonetal.[25]proposedalogicalframeworktoevaluateabnormaldrivingbehavior,and
designedthedrivingbehaviorquestionnaire(DBQ)todistinguishthreetypesofbehavior:error
behavior(plannedactionfailedtoachieveexpectedresult),failurebehavior(actionintention
deviation),andirregularitybehavior(deliberatelydeviatingfromthenormalsafetybehaviorora
sociallyacceptedcodeofconductforviolations).AmodifiedversionofDBQwasalsousedtostudy
theunusualbehavioroftwo‐wheeledvehiclesusers,suchasmotorcycleusersandmopedusers.
Elliottetal.[20]developedamotorbikedriverbehaviorquestionnaireandfounddifferencesin
trafficerrors,controlerrors,speedviolations,stunts,andtheuseofmotorcyclesafetyequipmentin
theUnitedKingdom(UK).StegandBrussel[19]developedaquestionnaireonthebehaviorof
mopedusersandverifiedthedifferencebetweenerrors,faults,andviolationsofmopedusersinthe
Netherlands.
Thevideocollectionmethodinvolvestheuseofelectronicmonitoringequipmentontheroadto
observeandcollectstatisticsontheridingbehaviorandusercharacteristicsofe‐bikeusers,resulting
inarelativelyrichamountofdatacollection.Zhouetal.[10]usedthe“globaleye”networkvideo
monitoringtechnologyofChinatelecomtoobtainreal‐timevideodataofe‐bikesinNingbo,
analyzedthemainfactorsaffectingtheendurancetimeofe‐bikes,andconcludedthattheweather,
crossroad,andthepresenceoftrafficpolicehadthelargestinfluence.Du[13]observed18,000e‐bike
usersattheintersectionofSuzhouandsummarizedriskyridingbehaviors.Konstantina[26]
observed90,000e‐bikeusersatsixmonitoringpointsinIowa,andstudiedtheinfluenceofroad
conditions,geographicallocation,weatherconditions,andotherfactorsontheuseofhelmets.
Truing[27]observed26,000usersofmotorcyclesande‐bikes,andobtainedacorrelationbetweenthe
useofmobilephoneswhileridingandfactorssuchasvehicletypeandage.Huanetal.[28]used
videomonitoringdataofintersectionstobuildamodeltoanalyzetheinfluencefactorsofwaiting
timeofe‐bikeusersandred‐lightrunningbehavioratintersections,andfoundthatalowernumber
ofe‐bikeusersorahighernumberofmotorvehiclesatanintersectionresultedinalesserlikelihood
ofusersexhibitingred‐lightrunningbehavior.
2. Structuralequationmodels(SEMs)andbivariateprobit(BP)modelsaremainlyadoptedinthe
analysisofriskyridingbehaviordataofe‐bikes.BeforetheconstructionofanSEMorBPmodel,
thedatafromquestionnairesurveysshouldbetested.
Theprocessofdatainspectionwascomposedofthreeparts.Firstly,exploratoryand
confirmatoryfactoranalysiswasperformedtoexaminethebasicdimensionsandstructuresusedto
detectabnormalcyclingbehaviorinthequestionnairedesign.Severalfittingindexesareusually
used,includingapproximateroot‐mean‐squareerror(RMSEA),goodness‐of‐fitindex(GFI),
adjustedgoodness‐of‐fitindex(AGFI),andcomparativefittingindex(CFI).Whenthefittingdegree
ofthemodelischecked,itisrequiredthattheRMSEAislowerthan0.08andtheGFIisgreaterthan
0.90,whiletheAGFIandCFIshouldindicateagoodmatchbetweenthemodelanddata[32].
Cronbach’salphaisaconsistencycoefficientusedtomeasurethehomogeneityofitemsinasingle

Int.J.Environ.Res.PublicHealth2019,16,23086of18
dimension,thatis,toevaluatethereliabilityandinternalconsistencyofscalesdescribingabnormal
cyclingbehavior,safetyattitude,andriskperception.AccordingtoNunnally’sstandard[33],avalue
ofequaltoorgreaterthan0.7indicatesacceptablereliability.
Secondly,correlationsanddifferenceswereanalyzed.Acyclingbehaviorscalewasestablished
tocheckwhethertherespondentsinvolvedintrafficaccidentsinthereportweresignificantly
differentbasedondemographicvariables,riskperception,safetyattitude,andabnormalcycling
behavior.Forunivariateanalysis,thechi‐squaretestwasusedforclassificationvariables,and
univariateanalysisofvariancewasusedforcontinuousvariables.Formultivariateanalysis,abinary
logisticregressionmodelwasusedtoidentifyfactorssignificantlyassociatedwithparticipantsinvolved
intheaccident.
Thirdly,theSEMmodelwasbuilttoexplorethecausalrelationshipbetweensafetyattitude,
riskperception,andabnormalcyclingbehaviorusingAMOS17.0software(SPSSInc,Chicago,IL,
U.S.A.).Thetwo‐stepprogramfortheSEMmodelwasrecommendedbyAndersonandGerbing
[34].Firstly,aconfirmatoryfactoranalysiswasusedtoevaluatetheSEMmodelandthe
measurementmodelsofsafetyattitude,riskperception,andabnormalridingbehaviorsubscales,as
wellastheirfittingdegreewiththeirrespectivepotentialstructures.Secondly,thestatistical
acceptabilityoftheSEMmodelwastested,andthemaximum‐likelihoodfunctionestimationmodel
wasadopted.Thecommonlyusedfittingindexesforinspection[12]ofRMSEA,GFI,AGFI,andCFI
wereusedforthefittingofthemeasurementmodel.
Severalresultsemergedintheliteraturebasedontheaboveoperationsteps.Forexample,based
ontheSEMmodel,YaoandWu[14]foundthatbothsafetyattitudeandriskperceptionsignificantly
affectedabnormalridingbehaviorsofe‐bikes.Guoetal.[35]designedabivariateprobability(BP)
modeltochecktheimportantfactorsrelatedtothecollisionofe‐bikesandthelicenseplatesof
e‐bikes,andconsideredthecorrelationbetweenthem.Themarginaleffectofcontributionfactors
wascalculatedtoquantifytheireffectontheresults.Theresultsshowedthatgender,age,education
level,drivinglicense,familycar,experienceofusinge‐bikes,compliancewithlaw,andactive
drivingbehaviorofe‐bikeusershadasignificantinfluenceone‐bikeaccidentsandlicenseplateuse.
Inaddition,thetypeofe‐bike,frequencyofe‐bikeuse,impulsivebehavior,degreeofriding
experience,andriskperceptionscalewerefoundtoberelatedtocollisionsinvolvinge‐bikes.This
furtherconfirmsthepreviousresearchconclusionofYaoandWu[14]thattheprobabilityoftraffic
accidentsofe‐bikeuserswithahighriskperceptionisrelativelylow.
2.2.CharacteristicsofRiskyRidingBehavior
E‐bikeridingisaffectedbyvariousriskfactors,whichincludeusers,vehicles,roadsandthe
environment,management,andotheraspects[36],asshowninTable1.
Table1.Influencingfactorsofelectricbicycle(e‐bike)riskyridingbehavior.
InfluencingFactorFactorsSet
RiderfactorsAge,gender,educationlevel,healthstatus,personalitycharacteristics,traffic
safetyawareness,cyclingbehavior,cyclingtechnology,decision‐makingability
VehiclefactorsBrakingperformance,steeringperformance,comfort
Roadand
environmental
factors
Trafficflowvolume,speed,widthofnon‐motorizedlane,formofroadsection,
conditionofroadsurface,conflictinterferencetype,weatherconditions,
artificialenvironment
ManagementfactorsRiskmanagement,organization,riskperception,communication
Otherfactors[37]Alcohol,drugs,socialnorms,confidence
E‐bikeusershavemanyriskyridingbehaviorsintheprocessofriding.Forexample,Brianetal.
[38]studiedtheridingbehaviorofe‐bikesandfoundthattheviolationrateofe‐bikeswasrelatively
high,andtheriskycyclingbehaviorsthatcausedtheviolationmainlyincludednotridinginthe
rightdirection,over‐speeding,trafficconflictswithotherroadparticipants,andwaitingforthe
signallightattheintersection.ZhaoMingetal.[39]conductedafour‐daytrafficsurveyinJinhuato

Int.J.Environ.Res.PublicHealth2019,16,23087of18
studytheriskyridingbehaviorsofe‐bikeusers.Theresultsshowedthatover‐speeding,manned
riding,red‐lightrunning,anddrivinginreversewerethemainriskyridingbehaviorsofe‐bike
users.Duetal.[13]researchedelectricbicycleridingbehaviorandfoundthatthereweresomerisky
ridingbehaviorssuchascarryingpassengersduringtheriding,takinguplanes,drivingthrougha
redlight,reversecycling,andmakingphonecalls.Theyalsofoundthatthenumberofmaleusers
wasproportionaltotheuseofhelmetsandtakingupmotorvehiclelanes,whilethee‐bikehavinga
platewascloselyrelatedtothebehaviorofcarryingapersonorcargo.Miaoetal.[40]studied
non‐motorvehicleusers’riskybehaviorandfoundthatnon‐motorvehicleusersmainlyhad
behaviorssuchasoccupyingthelanes[41],ridingsidebysideinnon‐motorvehiclelanes,being
closetolargevehiclesridingontherightside,forcinginfrontoflargevehicleswhenturningright,
weaving,drunkriding,over‐speeding,andridinginaclosedroad.Inaddition,non‐motorvehicles
alsogenerateaseriesofriskyridingbehaviorsduetothe“exceedingstandard”ofnon‐motor
vehiclesandthewetandslipperyroadsurfacethroughintersectionsorhighwayexitsonrainyand
snowydays.Wang[42],Jia[43],andJiangandLi[44]foundthroughinvestigationthatcyclists’risky
ridingbehaviorsintheprocessofcyclingincludedspeeding,illegalturning,reversedriving,
encroachingonmotorvehiclelanesandsidewalks,runningredlights,forcedovertaking,sudden
parkingorturning,andnon‐compliancewithregulations.
Onthebasisofstatisticalanalysisofaccidentdataandcauses,Ren[37]gave12kindsofrisky
ridingbehaviors,andthescoreofunsafebehaviorsisshowninFigure3.AscanbeseenfromFigure
3,alowerscoreofunsafebehaviorsofcyclistsiscorrelatedwithahigherfrequencyofthose
behaviors,indicatingthatthetrafficaccidentsandsafetyhazardsofsuchbehaviorsaregreater.It
canbefoundfromFigure1thatthetwounsafecyclingbehaviors,mannedridingandillegal
occupationofmotorvehiclelanes,havethehighestprobabilityoftrafficaccidents.

Figure3.Evaluationofunsafebehaviorofe‐bikes.
Thereisaconflictbetweene‐bikesandotherroadparticipantsatintersections.Manyscholars
studiedthewaitingbehaviorofcyclistsatintersections.Guoetal.[15]analyzedthered‐lightriding
behaviorofnon‐motorvehiclesatsignalizedintersectionsandfoundthate‐bikeusersweremore
likelytoruntheredlightthanordinarybikes.Xuetal.[45],instudyingthecharacteristicsofthe
illegalbehaviorsofmopedsatintersections,pointedoutthattheillegalbehaviorsincludedred‐light
running,drivingintheoppositedirection,occupyingthemotorcyclelane,illegalwaiting,andillegal
manningorobjects.Liu[46],Zhao[47],andHuanMeietal.[48,49]concludedfrompreviousstudies
thatrunningaredlightwasthemostseriousriskycyclingbehavioratsignalizedintersections.Yang
etal.[50]studiedthetrafficlightwaitingbehavioroftraditionalbicycleusersandelectricbicycle
usersatintersections,andfoundthattheprobabilityofcyclistsrunningaredlightwasproportional

Int.J.Environ.Res.PublicHealth2019,16,23088of18
tothewaitingtime,thatis,whenthered‐lighttimewaslessthan49seconds,about50%ofcyclists
rantheredlight,while,whenthetimeoftheredlightwaslessthan97seconds,about75%ofusers
rantheredlight.Theyalsofoundthate‐bikeusersweremoresensitivetochangesintheexternal
environment.Brianetal.[38],Huanetal.[49],andLiuandYang[51]foundthatthelongeracyclist
waitedatanintersection,thehighertheviolationratewas,whichwascloselyrelatedtoherd
mentality.Intheirsample,morethan50%ofuserscouldnottolerate49secondsorlonger,while25%
ofcyclistscouldtolerate97secondsorlonger.Zhouetal.[10]studiedtheinfluencingfactorsof
endurancetimeatintersectionsone‐bikesbyobtaining57,213samplesofendurancetimeofe‐bike
crossingsinNingbo.Statisticsshowedthattherewerethreesignificantvariablesaffectingthe
endurancetime:weather,crosswalklength,andthepresenceoftrafficpolicetoenforcethelaw.
Dongetal.[52]analyzedthesafetyinfluencemechanismofgreenflashinglightsonthestopping
decision‐makingbehaviorofe‐bikeusers,andfoundthatgreenflashinglightscouldpromptusersto
stop,whileitwaseasytogenerateaggressivepassingbehavior.Ifthegreenflashinglightisset
properly,itcaneffectivelyimprovethesafeoperationofe‐bikesatintersections.
Comparedwithpreviousstudies,wealsofoundthatsomefactorssuchastheweather,
temperature,androadinfrastructurewerealsocloselyrelatedtoe‐bikeridingbehavior.For
example,Konstantina[26]studiedthefactorsaffectinghelmetswornbye‐bikeusers,andfoundthat
theproportionofhelmetswornbycyclistsonmainroadsandsecondaryroadswaslower.Duetal.
[13]foundthat,comparedwithsunnyandcloudydays,cyclistsworemorehelmetsinrainydays.In
recentyears,duetothehotsummerweather,theultravioletintensityisgreater,andthe
phenomenonofred‐lightrunningismoreserious.Inviewofthecurrentsituationofweatherand
temperature,themanagementdepartmentsetupsunshadesaturbanintersectionsandachieved
goodresults.Zhangetal.[53]studiedthecyclingbehaviorsofcyclistsatsunshadeintersections,and
theresultsshowedthatthebehaviorofrunningredlightsatsunshadeintersectionsonsunnydays
wassignificantlylowerthanthatoncloudydays.Cyclistsatintersectionswithoutawningswere
significantlymorelikelytocrossredlights.Therefore,thesafetyofcyclistsatintersectionscanbe
improvedbysettingshadeawningsatintersections.Forexample,Zhangetal.[54]exploredthe
violationbehaviorofnon‐motorvehiclesoccupyingthemotorway,andabinarylogisticregression
modelwasadoptedtofindtheinternalreasonofsuchviolationbehavior.Theresultsshowedthat
thetrafficviolationrateofmalecyclistswashigherthanthatoffemalecyclists,andtheviolationrate
ofrainydayswashigherthanthatofsunnyandcloudydays.Secondly,theviolationrateinthe
morningpeakwashigherthanthatintheeveningpeakandoff‐peakhours.Thetrafficdensityof
motorvehiclesandnon‐motorvehicleshadastronginfluenceontheillegaltrafficbehaviorof
non‐motorvehicles.Thirdly,thedataanalysisshowedthattheaverageillegaltrafficrateof
non‐motorvehicleswas36.1%,indicatingthatmorethanone‐thirdofnon‐motorvehicleshadtraffic
violations.Wangetal.[55]usedtheaccidentdataofelectricbicycleandmotorvehiclecollisionsas
theresearchobjecttostudythefactorsaffectingthedegreeofinjuryofelectricbicycleriders.Itwas
foundthatthefactorsassociatedwithelectricbikeridercollisionswereviolationofsignalcontrols,
notaccordingtostipulationstogiveway,agegreaterthan60yearsold,andmale.Amongthese
factors,theratiosofaccidentsthatinvolvedaviolationofsignalcontrolsandnotaccordingto
stipulationstogivewaywere2.201and1.495,respectively,indicatingthattheriskofdeathor
seriousinjuryofelectricbicycleriderswasgreaterduetoaccidentsbetweenelectricbicycleriders
andmotorvehicles.Theaccidentoccurrenceratiosofriderswithagesgreaterthan60yearsand
maleriderswere1.383and1.317,indicatingthat,iftheageofanelectricbikeriderwasgreaterthan
60years,theriskofdeathorseriousinjurywashigherthanthatofyoungerdriversintheeventof
anaccident.Thiswasrelatedtotheweakerabilityoftheolderridertorespondtotheoutsideworld,
inadditiontotheirpoorphysicalconditionandtheexistenceofluck.Tangetal.[56]proposedto
useacellularautomatonmodeltostudythestraight,lane‐change,andretrogradebehaviorof
e‐bikeridersatsignalizedintersections.Theresultsshowedthatthelanechangeandretrograde
behaviorshadthemostsignificantinfluenceontheridingtrajectoryofe‐bikeriders.Thesimulation
resultscouldbetterexplainthecomplicatedtrafficphenomenoncausedbye‐bicyclesatsignalized
intersections.Thebasisfortheintroductionofe‐bikecontrolmeasureswasprovided.Yuetal.[57]

Int.J.Environ.Res.PublicHealth2019,16,23089of18
studiedthereactionofe‐bikeriderstopedestriancountdownsignaldevices(PCSDs),andfound
thatPCSDscouldeffectivelyreducethenumberofred‐lightviolationsbycyclists,effectively
preventingtherunningofredlights,butincreasingthepossibilityofearlybehavior.Fishmanand
Cherry[58]gavealiteraturereviewofthemajortrendsinthedevelopmentofglobalelectricbicycle
trafficfrom2006to2015.Theresultsshowedthat,intermsofroadsafetyissues,theviolationsof
Chineseelectricbicycleridersatintersectionsweremorecommon,whereastheriskofelectric
bicycleaccidentswashigherthanthatoftraditionalbicycles.
Inaddition,therearesomeotherridingbehaviorswhichcanalsobecomesafetyhazardsin
trafficaccidents.Forexample,Stellingetal.[59]studiedtheauditoryperceptionduringcollisions
andaccidentsofcyclistswhentheylistenedtomusicormadephonecalls[41].Theresultsshowed
thatlisteningtomusicandtalkingonthephonehadnegativeeffectsoncyclists’hearingand
attention,andtherewasagreaterriskoftrafficaccidents.Theyalsofoundthattherewasno
relationshipbetweenthefrequencyoflisteningtomusicandmakingphonecallsandthefrequency
oftrafficaccidents.Truongetal.[27]investigatedandanalyzedthemobilephoneuseofe‐bike
users,andfoundthat8.4%ofthemusedmobilephonesduringriding,andtheuseofmobilephones
duringridingwasrelatedtoridingweather,thenumberoflanes,differentlanes,thedurationofred
lights,andthepresenceofpolice.Inaddition,wearingsafetyhelmetsandsomeriskyriding
behaviorsattheexitofrain/snowexpresswaysorhighwayswerecloselyrelatedtotheoccurrence
andseverityofe‐bikeaccidents.
3.AnalysisofCharacteristicsofE‐BikeUsers
3.1.VisionandHearing
Visionandhearingarefundamentaltoridingane‐bike,andtheyaredirectlyrelatedtothe
perceptionoftheexternalenvironment.Thelevelofperceptionhasasignificantimpactonthesafety
oftheroad.Drunkcycling[25]reducesthevisualandothersensoryfunctionsofcyclists,andthe
acquisitionandjudgmentofexternalinformationarepronetoerrors.Inaddition,themoodofthe
drunkuserisveryunstable,andriskyridingbehaviorssuchasover‐speedingorrunningaredlight
areeasilygenerated.
E‐bikeusershavedifferinglevelsofdynamicandstaticvision[41],basedonfactorssuchasage
andphysicalfeature.Theuser’sdynamicobservationvariesaccordingtothespeedofcycling,
wherebyafasterspeedresultsinanarrowerdynamicfieldofvision,whileillusionscanalsooccur
duetodifferencesbetweensenses,whichareimportantfactorsincausingtrafficaccidents.Duetoits
smallsizeandlowdrivingnoiseontheroad,e‐bikesarenoteasilyheardorperceivedbyotherroad
participants[12];thus,therearemanyunsafeanduncertainfactorswhenconsideringnon‐motor
vehiclelanesandotherparticipants.Zhao[60]studiedthevisualbehaviorofe‐bikeusers.Their
resultsshowedthat,indifferentcyclingenvironments,thedistributionofeyemovementtimewas
uneven.Alargerproportionofscanningtimeledtoashorterdurationofeachfixationpoint,whilea
morecomplexenvironmentnarrowedtheuser’sgaze.Intheprocessofcycling,thefollowinga
vehiclecanbeconsideredahazardoussituationaselectricbikesovertakemorefrequently.
3.2.DifferentAgeGroups
Theprobabilityoftrafficaccidentscausedbycyclistsofdifferentagesvariesduetotheir
wide‐rangingphysicalfunctions.Wuetal.[61]studieduserridingbehaviorandtherelationship
withageusingsurveydata,andfoundthatyoungandmiddle‐agedpeopleweremorelikelytoride
througharedlightthanolderpeople.Theyalsofoundthatsmallgroupsofridersorindividual
ridersweremorelikelytorunredlights,whereasmiddle‐agedandolderridersweremorefearfulof
trafficaccidentswhileriding,leadingtomorecautiousbehavior[62].Accordingtothesurveydata
ofTruongetal.[27],theaverageageofe‐bikeusersis23yearsold,whilethatofmotorcycledrivers
is30yearsold.

Int.J.Environ.Res.PublicHealth2019,16,230810of18
3.3.GenderDifference
Usersofdifferentgendershavedifferentphysicalskillsandpsychologicalstates;thus,gender
differencealsohasanimpactonthebehaviorofcyclists.Wuetal.[61]studiedtheridingbehaviors
ofmaleandfemalecycliststhroughsurveydata,andfoundthattheprobabilityofmaleusers
runningaredlightwashigherthanthatoffemaleusers,especiallywhenthemopedsofmaleusers
hadgreaterdynamicperformance[41,62,63].AccordingtothequestionnairesurveyresultsofYao
andWu[13],itwasfoundthatgenderwassignificantlyrelatedtotrafficaccidents,wherebymen
weremorelikelytohavetrafficaccidentsthanwomen.Truongetal.[27]alsopointedoutthatthe
incidenceofmobilephoneuseincyclingwaslowerinwomenthaninmen.
3.4.ReactionAbility
Aslowresponsetimeandtheaccuracyoftheresponsearedirectlyrelatedtotheoccurrenceof
trafficaccidents.Wang[42]summarizedthereactionprocessesofusersasfollows:external
stimulus–consciousacknowledgement–response.Inotherwords,theinformationobtainedthrough
thesensorysystemisfedbacktothee‐bike,followedbyacertainbehavioraloperationafterthe
centralnervoussystemmakesadecision.Theresponsivenessisrelatedtothecyclingspeedofthe
vehicleandthecomplexityofthedrivingenvironment.Forexample,thedrivingspeedofanelectric
bicycleisfasterthanthatofanordinarybicycle;thus,therequirementforitsresponsivenessis
higherthanthatofabicycleuser.Ifthecomplexityofthedrivingenvironmentexceedsthe
processingabilityofthecyclistswithintheallowedrange,theirreactionabilitywillnotkeepupwith
thechangesoftheenvironment,andunsaferidingscenesareeasilygenerated.Fu[64]foundthat,
duetofatigue,drugs,pathology,andotherphysiologicalcharacteristics,cyclists’consciousnesslevel
woulddecline,theirreactionwouldberelativelyslow,andtheywouldbepronetodrowsinessand
othersymptoms.Theabovesymptomsweremainlymanifestedincyclingbehaviorwithwrong
ridingandpoorjudgment.
Insummary,differentuserridingbehaviorsoccurduetoseveralfactors,suchasgender,age,
andtheabilitytohearandrespondtooutsidestimuli.Incasesofdrinking,fatigue,illness,anddrug
use,thebrain’snervoussystemisinastateofdottiness,whichslowsreactiontimeandincreases
erraticbehaviorsuchasrunningredlights,seriouslyaffectingtheuserandotherparticipantsinterms
ofroadsafety.
3.5.PsychologicalFactors
Duetothedifferencesincyclistpsychology,thecharacteristicsofeachuserareunique.
Therefore,onthebasisofunderstandingthegeneralpsychologyofcyclists,specialpromotionand
educationshouldbecarriedouttoachievethepurposeofsafety.Wang[42]pointedoutthata
normalpsychologicalenvironmentisrequiredforsafecycling.Intheprocessofcycling,peopleare
pronetofear,transcendence,dispersion,conformity,habit,frustration,competitiveness,and
distraction,whicharehiddendangersleadingtotrafficaccidents.Fu[64]pointedoutthat,because
thespeedofe‐bikesisfasterthanthatofbicycles,cycliststendtoshowexhibitbehaviorssuchas
competitiveness,transcendence,conformity,independence,anddispersion,leavingthemproneto
unsaferidingbehaviors(suchasspeeding,following,runningredlights,etc.).
3.6.RelationshipbetweenDifferentFactorsandRidingBehaviors
Itcanbeseenfromtheaboveresearchthate‐bikeridershavedifferentpersonalattributes,and
theroadtrafficinfrastructureselectedbytheriderswhenridingisdifferent,whichwillgenerate
differentriskyridingbehaviors.Thedifferentvariablesusedandthecorrelationsdiscussedin
previousstudiesareshowninTable2.

Int.J.Environ.Res.PublicHealth2019,16,230811of18
Table2.Independentanddependentvariablesusedinpreviousstudies.
IndependentVariablesDependentVariables
VisionMannedbybike
HearingChattingwhileriding
DifferentagegroupsListeningtomusicwhileriding
GenderdifferenceCallingwhileriding
ReactionabilityRidingsidebyside
Psychologicalfactors
FearOverthespeedlimit
TranscendenceRunningaredlight
DispersionReversedriving
ConformityJaywalking
HabitExcessiveturningspeed
FrustrationDrinkinganddriving
CompetitivenessMotorwayoccupancy
DistractionNotadheringtostipulationstogiveway
Aircraftnon‐isolationbeltForcedovertaking
RedlightdurationSuddenstoppingorturning
TrafficsignmarkingFatigueriding
Accordingtothelistofindependentvariablesanddependentvariablesusedinprevious
studies[10,12,13,25,27,36–59,60–64],wefoundthatthecharacteristicsofe‐bikeusers,suchasvision,
hearing,agegroup,reactionability,andpsychologicalfactors,arethemainreasonsaffectingthe
riskyridingbehaviorofcyclists,aswellasredlightdurationandtrafficsignmarking.Forexample,
Duetal.[13],Wang[42],andZhangetal.[53]foundthatindependentvariables,suchasgroup
psychologyandwaitingtoolongfortheredlight,werethemainriskyridingbehaviorsrelatedto
runningaredlight.Fu[64]alsofoundthatpsychologicalfactors,suchastranscendence,habit,and
distraction,werealsothemainriskyridingbehaviorsrelatedtoover‐speed,aswellasgenderand
age.Whatismore,whencyclistsrideonroadswithinorganicnon‐isolationzones,theytendto
illegallyoccupymotorvehiclelanes[55,56].
Itcanbeseenfromtheabovereviewthattheriskyridingbehaviorsofe‐bikesarerelatedtothe
psychologicalcharacteristicsofusers.Thepsychologicalcharacteristicsrelatedtosafetyrisksareall
causedbytheweaksubjectivesafetyawarenessofcyclists.Therefore,itisexpectedthat,toprevent
theoccurrenceofsuchriskycyclingbehaviors,itisnecessarytocarryouttargetedpsychological
intervention.Tosumup,cyclists’riskyridingbehaviorsarethemainfactorscausingtraffic
accidents.Thesebehaviorsincludeillegallyoccupyingmotorvehiclelanes,runningredlights,and
illegallycarryingpeopleorobjects.Ingeneral,theseriskyridingbehaviorsaretheresultofalackof
knowledgeaboutthesafetyfeaturesofe‐bikesandtheweakawarenessofroadtrafficsafety.
Therefore,itisnecessarytostrengthensafetyawarenessandeducationtoimprovepsychological
preparedness.Furthermore,itisnecessarytoincreasecyclingskilltrainingtoimprovecoping
ability,andtoformulaterelevantlawsandregulationstoregulatecyclingbehavior.
4.PreventionandInterventionofE‐BikeTrafficAccidents
Inviewofthepreventionandinterventionofe‐biketrafficaccidents,previousresearchput
forwardrelevantmeasuresbasedmainlyonfouraspects:strengtheningtrafficmanagement,
improvinglawsandregulations,improvingthecyclingenvironmentofe‐bikes,andstrengthening
thesafetyeducationandtrainingofusers.
1. Intermsofstrengtheningtrafficmanagement,Ma[65]proposedgivingprioritytothe
developmentofpublictransportationtolimittheincreaseofthenumberofe‐bikes,followinga
studyofcollisionexperimentsbetweene‐bikesandmotorvehiclesandananalysisofthesafety
degreeofe‐bikes.Othersuggestedmeasuresincludelimitingthemaximumdesignspeedof
e‐bikestorealizesourcecontrol,carryingoutcyclingskilltrainingandassessment,
implementingannualinspection,andensuringthesafetyofvehicles.Dong[41]proposedthat,

Int.J.Environ.Res.PublicHealth2019,16,230812of18
inordertomoreclearlyidentifye‐bikeridingontheroad,reverselasertechnologycanbe
adoptedonthebodyandlicenseplateofthee‐biketopreventtheoccurrenceofcollision
accidents.JiangandLi[44]proposedtoregulatethee‐bikeindustry,strengthenthe
managementofusersande‐bikes,strengthentrafficmanagement,improveroadconditions,set
upe‐bikecitymanagementdepartments,andintroducerelevantlawsandregulationsonthe
basisofanalyzingthepotentialsafetyrisksofe‐bikes.LiuandYang[51]putforward
correspondingpermitmanagementandcompulsoryspeedlimitsystemsfore‐bikeusers’
red‐lightrunningbehavior,including(i)training,assessment,andlicensingbeforee‐bikeriding,
and(ii)penaltiesforillegalridingofe‐bikes.Similarly,Shi[66,67]andChen[68]proposed
unifiedregistrationandlistingfore‐bikes,astandardizedproductionofe‐bikes,the
implementationofapermitsystem,ande‐bikeuserspayingforinsurance.Xuetal.[45]
emphasizedstrengtheningproductionmanagement,regulatingmopeds,andpreventing
vehiclesfrom“exceedingthestandard”.Wuetal.[69]investigatedthestatusquooflicenseplate
registrationofe‐bikes,andfoundthat30.58%ofrespondentsregisteredtheirlicenseplates,but
therewerestill69.42%ofuserswhohadunregisteredlicenseplates.Carole[70],onthebasisofa
comprehensivedefinitionofriskyridingbehavior,suggestedthattherapidgrowthinvolumeof
electricbicycleswasconsistentwiththehighelectricbicycleaccidentrateinChina,andpointed
outthatrelevantdepartmentsshouldfirstestablishguidingprinciplesforthepreventionof
trafficaccidentsthroughaforward‐lookingguidancestrategy.
2. Intermsofimprovingrelevantlawsandregulations,Ma[65]advocatedthatlawsshouldbe
passedtoexplicitlyprohibitminorsfromridinge‐bikes.Chen[68]andWuetal.[69]proposed
thatusersmustwearhardhatstoridesoastonotbefined.Shietal.[66]andJangandLi[44]
proposedrevisingandincreasingthetrafficlawsandregulationsfore‐bikes,forcingcycliststo
changetheirriskyridingbehaviorsthroughlegalmeans.Truongetal.[27],throughan
investigationandanalysisofmobilephoneuseofe‐bikeusers,proposedcombiningexisting
legislationwithextensiveeducationandpublicitytoreducepotentialdeathandinjurycaused
bymobilephoneuseduringcycling.
3. Intermsofimprovingthecyclingenvironmentofe‐bikes,Dong[71]andDong[41]putforward
thattheseparationbetweenmachinesandnon‐machinesshouldbeensured.Gu[43]foundthat
speedisthekeytocontrollingusercyclingsafetyfactors;thus,measuressuchasvertical
migration(speedhumps,decelerationmachines,pavementtextures,raisedcrossings,vibrating
belts),horizontaldeflections(turns),roadnarrowing,andothers(closedroadwidth,central
islands,coatingsurfaces,controlofvehiclespeed,anddecelerationgraphics)weresuggested
aimedatslowingdowntraffictoreducehazardsfacedbyelectricbicycleusersandotherroad
members.Xuetal.[45]proposedimprovingthetransportationinfrastructurefore‐bikes.Zhao
[47]usedthetime–spaceisolationmethodtoisolatemachinesdrivingdirectlyandturningleftat
intersections,andsetano‐drivingareaofnon‐motorvehiclesatintersections.Theysuggested
settingleftandrightturninglanesatintersections,alongwithspecialsignallightsat
intersectionswithalargenumberofe‐bikes.Atsmallintersections,non‐motorvehicleswould
berequiredtocrosssidewalks,wherebyanon‐motorvehiclebypassareawouldbesetupatthe
intersection.
4. Intermsofstrengtheningusersafetyeducationandtraining,Wu[72],Xuetal.[45],Zhao[47],
andChen[68]proposedstrengtheningsafetyawarenessandeducation,emphasizingthe
importanceofroadsafetytoconsciouslyimprovesafetyawareness,therebyavoidingrisky
ridingbehavior.Inaddition,Chen[68]pointedoutthatthequasi‐drivingmodeofmotor
vehiclesshouldbeimitated,andtheskillsofcyclistsshouldbetrainedtoimprovetheirability
andexperience.
Inaddition,Wangetal.[73]foundthroughinvestigationthatthecurrentphenomenonof
childrenbeingcarriedone‐bikesisquitecommon,andpointedoutthatthesafetyguaranteeof
existingelectronicbicycleseatsislimited.Thus,thedesignstrengthandstructureoftheseatsshould
bereasonablystrengthenedtoensurethepersonalsafetyofchildrenwhenriding.Yuanetal.[74]

Int.J.Environ.Res.PublicHealth2019,16,230813of18
analyzed150trafficaccidentsofe‐bikesinBeijing,andfoundanincreaseatintersectionsandin
motorvehiclelanesinthesuburbswheretheoccurrenceofe‐bikeswashigh.Safetyeducationand
trainingshouldbecarriedoutforusersinthesuburbs,andthedesignoftrafficsafetyfacilitiesin
theseareasshouldbestrengthenedatthesametime.Yuanetal.[75],usingstatisticalanalysisof
vulnerableroadusersinBeijingcollisions,comparedpedestrian,bicycle,andelectricbicycle
accidentfrequency,severity,andinfluencingfactors.Itwasconcludedthatelectricbicycleaccidents
weredependentoncollisionspeedandpositionofthevictim,withsimilaritiesanddifferences
comparedtotheothergroups.Suggestionsweregivenforaccidentintervention.
Allinall,scholarsputforwardcomprehensivepreventionmeasuresforelectricbicycletraffic
accidentsonthebasisoftrafficmanagement,lawsandregulations,ridingconditions,andsafety
educationandtraining,includinguniformregistrationlicenses,aquasi‐drivingsystem,
improvementoftheroadandintersectiontrafficsafetyfacilities,strengtheningsafetyawarenessand
education,drivingskilltraining,andmore.Theseapproacheswouldhelpimprovethetrafficsafety
consciousnessoftheuserwhilesafeguardingthem.
5.ResearchProspect
Asafutureresearchdirectionofurbanroadtrafficsafety,itisofgreatsignificancetostudy
cyclists’riskyridingbehaviorstoreducetheoccurrenceoftrafficaccidentsinvolvinge‐bikesandto
ensurethesafetyofthelifeandpropertyoftrafficparticipants[76–82].Althoughthecurrenttraffic
safetyproblemsofe‐bikesattractedtheattentionofscholars,duetothedifficultiesinobtaining
trafficaccident‐relateddataandthelimitationsofresearchideas,therearestillmanydeficienciesin
thestudyofe‐bikeriskyridingbehaviors,whichneedtobefurtherstudied.
1. Incurrentelectricbicycleriskyridingbehaviorresearch,mosthistoricaldataareobtainedfrom
trafficaccidentsorsamplingsurveyresultsforstatisticalanalysisandresearch;however,dueto
incompletetrafficaccidentdata,theanalysisresultshavecertaindeviation.Instead,studiescan
makeuseofadvancedmonitoringsystemsandvideorecognitiontechnologytoobtainmore
comprehensiveelectricbicycletrafficaccidentdata,withtheanalysisofbigdatatominerisk
cyclingassociationrulesbetweenbehaviorandtrafficaccidents,tofindastrongrelationship
betweenvariousinfluencingfactors.Usingtheseresults,stronglycorrelatedriskyriding
behaviorscanbeusedtoputforwardeffectiveinterventionmeasures,inordertoeffectively
copewithfrequentelectricbicycletrafficaccidents[83–86].E‐bikeusersalsorangeinagefrom
18to80.Duetothedifferentagesofcyclists,theirphysicalfunctionandcyclingbehaviorare
different.Currently,intermsofstudiesoncyclists’behaviorsatdifferentages,theonlyrelevant
researchachievementsarebasedonthebehaviorofrunningredlights,whilenoresearchwas
carriedoutonotherriskycyclingbehaviors,whichareinurgentneedofstudy,suchasthe
implementationofanagethresholdtoensurethesafetyofe‐bikeriding.
2. Thehighaccidentratesandcasualtyratesofe‐bikesaremainlycausedbyriskyridingbehavior
cuetotheweaksubjectiveriskawarenessofusers,andthecurrentresearchresultsinthisregard
arenotsubstantialenough.Inviewofthis,itispossibletostrengthentheresearchoncyclists’
subjectiveriskawarenessinfuturestudies.Forexample,usingscalesofriskperception,risk
attitude,andthedegreeofrisktolerance,questionnairesurveyscanbeconductedonuserrisk
consciousness.Then,correlationanalysisandregressionanalysiscanbeusedtostudythe
relationshipbetweenelectricbicycleridingbehaviorandriskawareness,accordingtodifferent
personalitiesanddemographicvariables.Thisstudywouldbehelpfultoexplorethedifferences
inriskawarenessamongdifferentcyclists.Accordingtothesedifferences,safetyeducationand
trainingcanbecarriedoutinatargetedway,soastoultimatelyachievethegoalofpreventing
riskyridingbehavior.
3. Asacommutingtoolforsomeurbanresidents,e‐bikesareoftenusedinrainy,foggy,orsnowy
days.Atpresent,therearenotmanyresearchresultsonriskyridingbehaviorofe‐bikesunder
specialweatherconditions.Inviewofthis,quantitativeresearchshouldbeundertaken,
studyinge‐bikeridingrulesinspecialweatherconditionssuchasrain,fog,oriceandsnow.A

Int.J.Environ.Res.PublicHealth2019,16,230814of18
usersaferidingenvironmentthresholdshouldbedeterminedtoprioritizeelectricbicyclesafety,
thusavoidingriskybehaviorandreducingtrafficaccidents.Inaddition,althoughmanystudies
lookedattrafficsafetyinvolvinge‐bikes,therearefewstudiesonthepsychologicaland
physiologicalcharacteristicsofcyclists.Ontheonehand,researchrelatedtouserpersonality,
cognition,self‐assessment,andself‐adjustmentislacking.Ontheotherhand,researchonuser
attention,analysisability,spatialability,emergencyability,andnervoussystemand
physiologicalcharacteristicssuchascognitiveabilityisalsomissing.
4. Inthefuture,itwillbeinterestingtocapturethedataofthecurrentresearchone‐bikebehavior,
aswellastherelatedtechnologiestobeapplied.Moreover,byobtainingaccuratecyclingdata,
researcherscanexplorethepotentialimpactofcyclists’personalcharacteristicsoncycling
behavior,andaddresstheseissuesthroughsafetyeducationanddrivingtraining.
6.Discussions
E‐bikeridersindifferentregionshavedifferentriskyridingbehaviorsduetodifferencesin
regionalcultureandtrafficregulations.Forexample,China’sdataacquisitionfore‐bikeriders
generallyincludes(a)ridingconflictsandillegalridingbehaviors,usuallyobservedthroughvideo
surveillanceequipmentatintersections[10,13,27,28,58],and(b)accidentdataofhospitals,violation
records,andpersonalinjuryrecords[58].Researchdataonbehaviorsofe‐bikeriderswerealso
obtainedusingquestionnairesmethods[17,19–24,29–31].
Overall,theriskyridingbehaviorofe‐bikeridersinChinaandEuropeissimilar.InChina,
previousstudies[58]foundthattheriskyridingbehaviorsincludeddoubleriding(orcarpooling),
chattingwhileriding,listeningtomusicwhileriding,callingwhileriding,ridinginparallel,
over‐speeding,jaywalking,reverseriding,suddenlycrossingtheroad,speedingtoofastwhen
turning,illegaloccupation,failuretocomplywithregulations,forciblyovertaking,suddenly
stoppingorturning,andfatigueriding.InEurope,previousstudies[26,38]foundthattherisky
ridingbehaviorsincludedridinginthewrongdirection,over‐speeding,trafficconflictswithother
e‐bikeriders,jumpingthelights,notwearinghelmets,listeningtomusicwhileriding,making
phonecallswhileriding,etc.Fromtheabove,theregionalissuesregardinge‐bikebehaviorshowed
somedifferencesbetweenChinaandEurope,sincethelaws,regulations,andcharacteristicsof
e‐bikeusersaredifferent[58].
7.Conclusions
1. Currently,e‐biketrafficaccidentsoccurfrequently,andmorethan90%ofe‐biketrafficaccidents
arecausedbycyclists’riskyridingbehaviors,includingillegaloccupationofmotorvehicle
lanes,over‐speeding,runningredlights,andgoingagainstthetrafficflow.Itisofgreat
significancetostudytheriskyridingbehaviorofe‐bikes,researchingaccidentcharacteristicsso
astopreventthemandensurethesafetyofpeople’slifeandproperty.
2. Substantialpreviousresearchwascarriedoutontrafficaccidentcharacteristicsandcausesof
e‐bikeriskybehavior,aswellastheirpreventionandintervention.However,afterreviewingthe
relevantliteratureathomeandabroad,theauthorsfoundthattheexistingrelevantresearch
resultshavethreedeficiencies,andpointedoutresearchdirectionsthatcanbefurtherexplored
inthefuture.
3. Realcyclingscenesareoftenacombinationofavarietyofsituations.Forexample,usersof
differentagesrideelectricbikesequippedwithumbrellasinrainyorsnowyweather.Inviewof
this,itisnecessarytostudythedegreeofdangerofcombinationscenariosfeaturingelectric
bicycleridingbehavior,wherebythedegreeofdangerisnotsimplyafunctionoftheriskyriding
behaviorinasinglescene.Thisresearchrequiresquantitativeanalysisusingthetheoryof
couplingcharacteristicsofvariousscenariosseeninriskyridingbehavior.Onthebasisof
effectiveinterventionsandthecouplingeffectofvariouscombinationsofinterventionsto
achieveaneliminationofaccidentrisk,onecanensuretrafficsafety.

Int.J.Environ.Res.PublicHealth2019,16,230815of18
AuthorContributions:Conceptualization,C.M.;datacuration,C.M.andD.Y.;methodology,C.M.andJ.Z.;
resources,Z.F.andQ.Y.;supervision,C.M.andJ.Z.;writing—reviewandediting,C.M.andJ.Z.;funding
acquisition,C.M.andJ.Z.
ConflictsofInterest:Theauthorsdeclarenoconflictsofinterestregardingthepublicationofthispaper.
Funding:ThisresearchwasfundedbytheNationalNaturalScienceFoundationofChina(ProjectNo.
71861023),theProgramofHumanitiesandSocialScienceoftheEducationMinistryofChina(ProjectNo.
18YJC630118),theZhejiangProvincialPhilosophyandSocialScienceFundofChina(No.18NDJC107YB;
19NDJC119YB),theEducationalSciencePlanningoftheZhejiangProvincialEducationDepartment(No.
2018SCG103),theNaturalScienceFoundationofZhejiangProvince(No.LQ19E080003),theNaturalScience
FoundationofNingboMunicipality(No.2018A610127),andtheFoundationoftheHundredYouthTalents
TrainingProgramofLanzhouJiaotongUniversity.
References
1. ChinaNewsNetwork.China’sElectricBicyclesHaveaSocialHoldingCapacityof250MillionVehicles.
Availableonline:http://news.ddc.net.cn/newsview_71477.html(accessedon2September2018).
2. Sohu.com.TheNumberofElectricVehiclesinNanningHasExceeded1.8Million.Availableonline:
http://www.sohu.com/a/229498412_394146(accessedon2September2018).
3. TrafficAdministrationBureauoftheMinistryofPublicSecurity.People’sRepublicofChinaannualreportof
roadtrafficaccidents(2016);TrafficManagementBureauoftheMinistryofPublicSecurity:Beijing,China,
2017.
4. TrafficManagementBureauofMinistryofPublicSecurity.StatisticsofRoadTrafficAccidentsinPRC(2015);
Beijing:TrafficManagementBureauofMinistryofPublicSecurity,2016.
5. JSTV.com.Nearly70%ofTrafficAccidentsareRelatedtoElectricBicycles.Availableonline:
http://news.jstv.com/a/20160805/123632.shtml(accessedon2September2018).
6. Hu,F.;Lv,D.;Zhu,J.;Fang,J.Relatedriskfactorsforinjuryseverityofe‐bikeandbicyclecrashesinHefei.
TrafficInj.Prev.2014,15,319–323.
7. Du,W.;Yang,J.;Powis,B.;Zheng,X.;Ozanne‐Smith,J.;Bilston,L.;He,J.;Ma,T.;Wang,X.andWu,M.
Epidemiologicalprofileofhospitalisedinjuriesamongelectricbicycleridersadmittedtoaruralhospitalin
Suzhou:Across‐sectionalstudy.Inj.Prev.J.Int.Soc.ChildAdolesc.Inj.Prev.2014,20,128–33.
8. CompiledbyNationalBureauofStatisticsofChina.ChinaStatisticalYearbook[M];ChinaStatisticsPress:
Beijing,China,2017.
9. Chen,Q.;Dai,B.Characteristicsandcasualtyanalysisoftwo‐wheeleraccidentsinchina,datasource:The
chinain‐depthaccidentstudy(CIDAS).SAETech.Pap.2018,2018,e01487191.
10. Zhang,Z.Y.;Hao,X.Y.;Wang,D.;Wang,D.Researchonpedestrian’sendurancetimeatsignalized
intersectioninBeijing.J.Transp.Syst.Eng.Inf.Technol.2015,15,213–219.
11. Cherry,C.;Cerver,R.UsecharacteristicsandmodechoicebehaviorofelectricbikeusersinChina.Transp.
Policy2007,14,247–257.
12. Du,W.;Yang,J.;Powis,B.;Zheng,X.;Ozanne‐Smith,J.;Bilston,L.;Wu,M.Understandingon‐road
practicesofelectricbikeriders:anobservationalstudyinadevelopedcityofChina.Accid.Anal.Prev.2013,
59,319–326.
13. Yao,L.;Wu,C.TrafficsafetyforelectricbikeridersinChina:Attitudes,riskperception,andaberrant
ridingbehaviors.Transp.Res.Rec.J.Transp.Res.Board2012,2314,49–56.
14. Papoutsi,S.;Martinolli,L.;Braun,C.T.;Exadaktylos,A.K.E‐bikeinjuries:Experiencefromanurban
emergencydepartment‐aretrospectivestudyfromSwitzerland.Emerg.Med.Int.2014,2014,1–5.
15. Guo,Y.;Liu,P.;Bai,L.;Xu,C.;Chen,J.Redlightrunningbehaviorofelectricbicyclesatsignalized
intersectionsinChina.Transp.Res.Rec.J.Transp.Res.Board2014,2468,28–37.
16. Zhou,J.B.;Guo,Y.Y.;Wu,Y.;Dong,S.Assessingfactorsrelatedtoe‐bikecrashande‐bikelicenseplateuse.
J.Transp.Syst.Eng.Inf.Technol.2017,17,229–234.
17. Zhou,J.B.;Wang,Q.Y.;Zhang,M.J.;Dong,S.;Zhang,S.C.Anempiricalstudyonsevenfactorsinfluencing
waitingendurancetimeofe‐bike.J.Transp.Syst.Eng.Inf.Technol.2017,17,242–249.
18. Guo,Y.;Wu,Y.;Lu,J.;Zhou,J.ModelingtheUnobservedHeterogeneityinE‐bikeCollisionSeverityUsing
FullBayesianRandomParametersMultinomialLogitRegression.Sustainability2019,11,2071.

Int.J.Environ.Res.PublicHealth2019,16,230816of18
19. Steg,L.;Annemarie,V.B.Accidents,aberrantbehaviours,andspeedingofyoungmopedriders.Transp.
Res.PartFTrafficPsychol.Behav.2009,12,503–511.
20. Elliott,M.A.;Baughan,C.J.;Sexton,B.F.Errorsandviolationsinrelationtomotorcyclists’crashrisk.Accid.
Anal.Prev.2007,29,491–499.
21. Xie,C.Q.;Dianne,P.AsocialpsychologicalapproachtodrivingviolationsintwoChinesecities.Transp.
Res.PartFTrafficPsychol.Behav.2001,5,293–308.
22. Ulleberg,P.;Torbjorn,R.Personality,attitudesandriskperceptionaspredictorsofriskydriving
behaviouramongyoungdrivers.Saf.Sci.2003,41,427–443.
23. Chen,C.F.Personality,safetyattitudesandriskydrivingbehaviors—EvidencefromyoungTaiwanese
motorcyclists.Accid.Anal.Prev.2009,41,963–968.
24. Scheper,J.P.Thesafetyofelectricallyassistedbicyclescomparedtoclassicbicycles.Accid.Anal.Prev.2014,
73,174–180.
25. Reason,J.;Manstead,A.;Stradling,S.;Baxter,J.;Campbell,K.Errorsandviolationsontheroads:Areal
distinction.Ergonomics1990,33,1315–1332.
26. Konstantina,G.ModelingmotorcyclehelmetuseinIowa:Evidencefromsixroadsideobservational
surveys.Accid.Anal.Prev.2009,41,479–484.
27. Truong,L.T.;Nguyen,H.T.;De,G.C.Mobilephoneuseamongmotorcyclistsandelectricbikeriders:A
casestudyofHanoi,Vietnam.Accid.Anal.Prev.2016,91,208–215.
28. Huan,M.;Yang,X.;Jia,B.Crossingreliabilityofelectricbikeridersaturbanintersections.Math.Probl.Eng.
2013,2,831–842.
29. Hertach,P.;Uhr,A.;Niemann,S.;Cavegn,M.Characteristicsofsingle‐vehiclecrasheswithe‐bikesin
Switzerland.Accid.Anal.Prev.2018,117,232–238.
30. Wang,Z.;Neitzel,R.L.;Xue,X.;Zheng,W.;Jiang,G.Awareness,ridingbehaviors,andlegislativeattitudes
towardelectricbikesamongtwotypesofroadusers:AninvestigationinTianjin,amunicipalityinChina.
TrafficInjuryPrev.2019,20,1–7.
31. Guo,Y.;Li,Z.;Wu,Y.;Xu,C.Evaluatingfactorsaffectingelectricbikeusers’registrationoflicenseplatein
ChinausingBayesianapproach.Transp.Res.PartFTrafficPsychol.Behav.2018,59,212–221.