搜索
EstimatidnofRoadSituationsandDriver'sIntentioninAutomo~veElectronicControlSystemQinGuiheLeiYulongNiuillingkuiGehlinDaiYisongSilinUniversityofTechnology,ReruninStreet142,ChangchunCity,ChinaAbstract:Thispaperdescribesthesignification,rhefeasibility,amethodbasedonpatternrecognitionandanapplicationexampleofestimationofroadsituationsanddriver'soperatingintentions(RSDOI).TheinformationofRSDOlisnotsameinformasthestateparameters.AccordingtothespecialtyofRSODI,a.fuzzypatternrecognitionmethodisraisedtoestimatetheinformation.Somerelationsbetweenthedriver,road,automobileandtheeleckoniccontrolsystemsoftheautomobileareillustrated.Kqwon-kAutomotiveElectronic,PatternRecognation,FuzzyLogic,SignalProcessingMoreandmoreattentionisbeingpaidtotheideaof'consideringhiman,machineandworkingenviroanmentsituationssyntheticallywhenamechaicalsystemisdesignedEl].TodesignahighidltFd@pdmechcdl--electroilic--h€onnatatiomlurijoiisystem,meofthekeytechnologiesisRowtheelectroniccontrol~mitFECL?,whichis+emostimportantandintelligentpartofthemadme,sensestheuser'sYnterrtionsandthcworkingenviromieotfeanrres121.R%enbeirgdesignedanautomotiveECI:shoriidbemadeoblaiiairmeinformtionaboutRSDOJwith&,ebehmiorofautomobileanddriver.Theinfamatb~~cmbeestimatedwithcombinationofthesenseofdriverandmachine,thecombinationofoperationandsensefunctionsandthefeedbackinformaliorioftheactionsofdria'ex.andairton;obile.InmmedevelopmentworksfautomobileelectrmiccontrolsystemsresearchofestimationofRSDCZiIhasbecninvolved[3][4].'%eideaistendedtoobtainthei&?mnationjustomthestateparmetersofautomobileandtheoperatingparametersofdriver,whichareusedbyECUofanautomotivecontrolsystem.Becausecostofautomobilesiscritical,itisveryimportanttoestimatetheinformationwitheasyineasurcdparaneter[SI.Thepatternrecognitionmethodisos~allyusedintheissue[6].Computingconsuming,lowerpr.eeisionaidYWOWactqxabifityarethemostprincipaldefectsoffieniehod[7].Inthepaper.anideabasedonfimypatternrecognitionisraised.Themethodneedslesscornputkgandhaswideradaptability;andinsomecaseitcanbeuseddirectlyinthecontrolalgorithm.I1.EssentialInformationofAutomotiveElectronicControlSystemsTocarryoutpropercontrolautomotiveelectroniccontrolsystemsneedsuchinformationasautomotivestateparameters,automotivenlnsningroadenvironmentfeatures,driver'sinteritionsanddriver'sopemtingcommands.Intheinfornzation,whatcanbeeasilyobtainedbyECUwithcommonsensorsarestateparametersandoperatingCoInmaricis.Usually,onlythetwokindsofinformationareusedintradi?ionalautomotiveeiecbroiiiccontrolsystem.AlongwiththeincreasingrequirementofautomotwepePforaiancesinroadadaptingproperty,hunm-machmefriendandintelligenceofautomobiles,ECWneedsmoreandmoreinformationaboutroadenvirorunegtanddriver'sintentions.ThoughthesekindsofinformationusQdliycannotbemeasureddirectly,theycanbeestimated.vvvithdid:stateparametersanddwver'soperations.Therelationsbetweenstateparmicters,wpemtmgparameters,thecontrolofECU,theroadcwiromentadthed~dvei%ioperatingintentionscanbeillustratedasFig.1.111.IdentificationofRSLOIRoadfeaturesaffectingzutomobiievorkmgsMeaiidoperatingperfomtan~eincludecnngest~o:~,cmatue,slopeandroughnessDnvcr'sintentionshciemeanaccclcratmn,decelerationarids~~rpassmgSomewadfeaturesareinptsofautomotivertdnrvngcourse.'Iogetaerwithotherillputstheydeternunetherunmngstateofautomobile.'Ilexfeaturescat1beestimatedwithLheautonotivedyniimcaimodels,themeasuredmputsandthestateparametersItistheiarnrungcourseoftheautombilethatlirushc.;?hetrandomofthemfornnatron.'I'hatISso-calledabc,sensorfi:nciionsofthemachmes.BFig.1.RelationofDriver,AutomobileandRoadActuatorsSomeroadfeaturesnotonlycan'tbemeasuredwithordrnarysensorseasilybutalsocan'tbereflectedbytheautomotivestateparameters,buttheycaneasilybesensedbythedriver.Undertherestraintofhisintentions,drivingknowledgeandoperatingrulesthedrivershouldoperateaccordingtothesensedinformation.Withanalyzingthefeaturesoftheoperatedparameterstheroadmformationcanbeestimated.Therefore,thedriverestablishesalinkagebetweenroadenvironmentandECU.Thatisso-calledthesensorfunctionsoftheoperators,@ig.2.).Whenadriverholdssome&tentionsduringdrivinghewillmanagetorealizethembyoperations.Therefore,theintentionsarerelatedwiththeoperatingparameters.ThedrivingintentionscanbeclassifiedintosevertypicalRoadenvironment--+models.Thedrivingintentionmodelsaredefinedwitheasymeasuredsignals.Thepatternrecognitionmethodscanbeusedtoestimatethedrivingintentions.TheinterestedRSDOIvarywithautomotivecontrolsystems.TheRSDOIwhichaffectstheperformanceofautomatedmechanicaltransmission(AMT)mainlyincludeslope,turnroad,congestion,acceleration,decelerationandsurpassing.Inthetypicalsuchsituations,automobilespeedv,throttleopeningt,accelerationa,breakingfiequencyb,breakingtimeaccumulationb,,inaintervalaresampledatto,t,,tz,......t,andnormalizedtoconstructthefollowingfuzzyformtimeseries:FInferenceandunconsciousreflection0Operatingas.0+&+a-bparameterssg.a9Fi:v)a3e.0Fig.2.SensorhnctionsofdriverKnowledgeofautomobiledynamics,safetyandroadfeaturesNv.={viy......v:,1ItgivesoutadescriptionofRSDOIatthetime.Nt,={t: