DOC

Study_0

By Julia Roberts,2014-09-22 12:48
11 views 0
Study_0study,Study

    Study

HIGHTECHNOLOGYLETIERSfVo1.16No.1IMar.2010fPP.9093

    ;doi:10.3772/j.issn.1006-6748.2010.01.017

    ;StudyonsimulationandschedulingalgorithmofCANcontrol ;forindependentlydrivingelectricvehicle?

    ;MengXiang(孟祥),CaoWanke~,LinCheng,SunTielei

    ;(NationalEngineeringLaboratoryforElectricVehicle,BeijingInstituteofTechnology,Beijing100081,

    P.R.China)

    ;Abstract

    ;Safety-criticalapplicationssuchastheindependentlydrivingsystemsofelectricvehicle(EV)re

    ;quireahighdegreeofreliability.Thecontrollerareanetwork(CAN)isusedextensivelyinthecontrol

    ;sectors.Anewrealtimeandreliableschedulingalgorithmbasedontime-triggeredschedulerwithafoc

    us

    ;offtheCAN.baseddistributedcontrolsystemsforindependentlydrivingEVisexploited.Adistributed

    ;controlnetworkmodelforadua1.whee

    independendydrivingEVisestablished.Thetimingandreliabili. ;tyanalysisintheworstcasewiththealgorittunisusedtoevaluatethepredictabilityanddependabilityand

    ;thesimulationbasedonthealgorithmwithCANoesoftwareisdesigned.Theresultsindicatethealgorith

    m

    ;ismolepredicableanddependable.

    ;Keywords:independentlydriving,schedulingalgorithm,real-timeanalysis,time-triggeredsched

    ;uler,controllerareanetwork(CAN)

    ;Thecontrollerareanetwork(CAN)hasgainedwide.

    ;spreadacceptancefirstinautomotivevehicle(e.g.auto

    ;motivecorllmunicationsystem)andtheninindustrialau. ;tomationandnetworkeontro1systembasedonfield.busas ;well?,23.CANisou~tandingbecauseofitslowcostand ;usefulfeatureslikereliabilityandflexibilityinrea1.time ;controlsystemsofautomotivevehicle(e.g.en~neman

    ;agementsystem).In-vehicleCANalsosupportscomplex ;distributedapplicationslikeantiloekbrakingsystem, ;power-trainsystemandautomaticcruisecontrolsystem. ;Followingarationallinetodosomeworkwel1.CANis

    ;valuableforthedistributedeontrolsystemsofelectricve. ;hicle.However.X.by.wiresystemandindependently ;drivingsystemhaveplacedarequirementondesignengi- ;neerstoproducemoreefficientandrobustdistributedcon

    ;trolsystems.Suchsystemsrequireboundedrea1.timebe. ;haviorunderalloperatingconditions1),.

    ;Inthispaperthedependableandflexiblescheduling

    ;algorithmofCANf0rindependentlydrivingsystemofEV ;isthefoealissue.Thepaperisstructuredasfollows:in ;Section1adistributedCAN.basedcontrolmode1foradu. ;al-whindependendydrivingEVisdesigned.andthen ;Section2explainsthetheoryofthetime?-triggeredsched-- ;uler,rea1.timeanalysisandsimulationareimplemented ;respectivelyinSections3and4.Finally,Conclusionand ;futureworkarepresented.

    ;Inourpmject,whichissupportedbytheNational ;HighTechnologyResearchandDevelopmentProgramof ;China.anelectricvehicleth2independently-driving ;motorsisdesigned.Thedistributedcontrolsystemofthe ;EVconsistsofleftmotorandleftmotorcontroller,fight ;motorandrightmotorcontroller,vehiclecontroller,bat

    ;terymanagementsystem(BMS)andinformationlnanage- ;mentsystem(IMS)asshowninFig.1.Thosecontrollers ;andsystemsareinterlinkedbyCAN.

    ;Fig.1ThecontrolnetworkbasedonCANofEVwith2 ;independentlydrivingmotors

    ;?Supmrt~bythe(.(),.

    ;NationalHighTechnologyRe~chandDevelopmentProgrammeofChinaNo2008AA11A146ChinaPo

    stdoctoralScienceFound

    ;afton20a9O450298j.

    ;?Towhomcorrespondenceshouldbeaddressed.E-mail:war~ecaol0000@gnmil.ton

    ;ReceivedonSept.1,2009

    ;H1GHTECHN0L0GYIETrERSIV01.16No.1IMar.2O1O ;TheCANisusedtoscheduleasetofsignalsofreal? ;timestatedataandcontrolcon~nandbetweenfivedifferent ;subsystems.Inthiscase,thenetworkneedstognarantee ;astrictlydeterministicbehaviorandamuchhigherdegree ;ofsafety.However,nativeeventtriggeredCANseemsnot

    ;tobesuitableforsupportingtherequirements.Sowe

    ;exploitanewrealtimeschedulingalgorithmbasedon

    ;timetriggeredschedulertoimprovethepredictabilityand ;dependabilityofCAN.

    ;2Aschedulingalgorithmbasedontime-trig-

    ;geredscheduler

    ;Inarealtimeautomotivesystem,realtimesignals

    ;areusuallysentperiodicallyorafteroccurrenceofan ;event.Controllerareanetwork,whosemediumaccess ;mechanismisacarriersensemultipleaccessschemewith ;nondestructivebitwisearbitration(CSMA.NDA),ises

    ;sentiallyevent-triggered.CANisflexibleenough,butnn

    ;predictablecollisionsmayaffectdeterminismseverelyand

;even”babblingidiot”problemofafaultynode.whichre—

    ;peatedlytransmitsaveryhighpfiori~messageonthe ;bus,canblockthewholenetwork.PreviouslymanFreal

    ;timeapplicationsbasedonunderlyingCANweredesigned ;withthefixed.priorityschedulingpolicysuchasrate ;monotonicscheduling(RMS)anddeadlinemonotonic ;scheduling(DMS).0ntheotherhand,Timetriggered

    ;communication(e.g.rITI,CANbasedon11)MA)iswel1 ;adaptedtob~picaltransmissionofrealtimedatawithhi

    ;determinability,orlowiitter.However,inprevious ;work.someliteraturesindicatedthat,rrCANdidn’tsup-

    ;portflexibletime.triggeredtrafficandsoitcontributedto ;alowflexibility.

    ;Aplan.basedschedulingalgorithmbasedontime

    ;triggeredschedulerisexploited.SimilartoTTCANproto

    ;co1.thetime-triggeredschedulerisimplementedaccord

    ;ingtoconsecutiveelementarycycles(EC).A11nodesare ;synchronizedatthestartofeachECbyreceptionofapar

    ;ticularmessageknownasECtriggermessage(TM), ;whichissentbyaparticularnodecalledtemporalmaster ;asshowninFig.2.rI1hetransmissionofthismessage,in

    ;cludingstuffbits,takesTMconstanttimeunits.How

    ;ever.unlikeinTTCANprotocol,theplanbasedalgo-

    ;rithmbasedontimetriggeredschedulerimprovesthetrig. ;geredmessage,whichisusedtocarryatriggermessage ;planindatafieldofitselfasshowninFig.3.Inatrig

    ;geredmessageplan.thebitwith1indicatesthecorre

    ;spondingscheduledmessage.

    ;Triggeredmessageplansaremanagedbyasyn

    ;chmnousrequirementtable(SRT)residinginthetempo- ;ralmaster.AccordingtotheSRT,therespectivetriggered ;P/CTableP,cTableP/CTable

    ;jECTfigge~dMessage(TM).......................................................

    ;SRTofmaster

    ;CAN

    ;Fig.2Timetriggeredscheduler

    ;ECTriggefmessageSynchronousmessages

    ;Atriggermessageplan

    ;BytelByteO

    ;圈匝?圆匪皿姗

    ;/,\

    ;Bitl3Bit4Bit2Bitl

    ;Fig.3Atriggermessageplaninelementarycycle ;messageplaninECisbroadcasted.A11controlnodes(in.

    ;cludingthetemporalnodes)havetodecodetheplanand ;checkwhethertheyareproducersofanyofthespecified ;messages.Thecheckingisimplementedbyscanningalo- ;calproduced/consumedtable(P/CTable)residingina ;localnodeofitself.Specifiedsynchronousmessageswi11 ;starttotransmit,andeventualcollisionsonbusaccessare ;resolvedbythenativedistributedMACprotocolofCAN. ;3Real?timeanalysisoftheschedulingalgO

    ;rithm

    ;Beforeintroducingtheanalysissometermsarede. ;fined.Theperiodofavenmessagemisdenotedas ;.denotestheworstcasetimetakentotransmitthe

    ;messagephysicallyonthebus.Thetemabdefinesthe ;numberofbyteinthemessagem.Tjdenotesthetime ;forqueuingjitterofthemessagem.Theworst-easere. ;sponsetimeofavenmessagemisthelongesttimebe- ;tweenthequeuingofamessageandthemessagearriving ;atdestinationstations,andisdenotedas. ;Thedead-

    ;lineofthemessageisdenotedas.Theworstcasere

    ;sponsetimeisdefinedtobecomposedofthequeuingde- ;layandthetransmissiondelay.Thequeuingdelayiscom

    ;posedoftwotimeparameters:denotesthelongest ;timethatanylowerprioritymessagecanspendonthe ;bus;denotesthelongesttimethatisspentbyall ;higher-prioritymessagesonthebusbeforethemessage,n ;istransmitted.r6isthetimetakentotransmitabiton ;CAN[7].

    ;Fromtheabovedescriptionitcanbeseenthat: ;Tom:(8b+g+10+L-_j)rb,

    ;HIGHI1ECHN0LoGYIEmRSIVo1.16No.1IMar.2010 ;J’34,standardmessage,ng

    ;54,expandedruessagem

    ;TSm

    ;v

    ;m

    ;?

    ;a

    ;(

    ;x

    ;)

    ;{TCm+36}

    ;=

    ;r

;I

    ;Ci一个一

    ;V’?如(m)

    ;(1)

    ;(2)

    ;(3)

    ;Tnm=++zTcm(4)

    ;Areal-timemessageisschedulableifandonlyif ;?Tom(5)

    ;Thenetworkutilizationfactor,denotedasUN”is

    ;definedas:

    ;

    ;UNet(6)

    ;Therelativeresponsetimejitterofamessageisthe ;maximumtimevariationofresponsetimesofanytwocon

    ;secutivetransmissions,andisdenotedasRat.RRJisde

    ;finedas[8]

    ;Redm=maxIrL(7)

    ;4AnalysisresultsandsimulationinCANoe

    ;TherealtimemessagesscheduledbyCANforinde

    ;pendentlv.drivingsystemofEVar-edefinedasshownin ;Table1.Allmessagesareexchangedbyexpandedframe ;formatandthenetworkbaudrateissetas25Okbits/s.The ;plan-basedscheduling(PBS)andratemonotonic ;scheduling(RMS)arerespectivelyusedtoschedulethe ;real-timemessages.Theanalysisresultsareshownin ;TaJ)le2andinFig.4.

    ;Table1Therealtimemessagesofstatedataandcontrolcorrmlandsignals ;Table2TheanalysismsNNofrealtimeperforrmance

    ;basedOiltWOschedulingalgorithms

    ;Theanalysisresultsofthenetworkutilizationfactor ;athatUNecofthenetworkscheduledbyRMSis ;33.912%andUNetofthenetworkscheduledbyPBSis ;40.192%.ThestatisticsresultsofsimulationwithCANoe ;systemarethatUNetofthenetworkscheduledbyRMSis ;26.78%asshowninFig.5andUN.tofthenetwork ;scheduledbyPBSis31.74%asshowninFig.6.Be

    ;causeofcalculatingtheworst-cBsestuffbits,theresults ;ofanalysisarelargerthanthatofsimulationresults. ;Therequirementsfordependableandflexibleser. ;vicesoftherea1.timenetworksystemarebecomingin. ;creasinglyimportantinelectricvehicle,inparticular,for ;rea1.timeindependentlydrivingsystemofelectricvehicle ;basedonCAN.Inthepaper.itiSshownhowtofinda

    ;dependableandflexibleplanbasedschedulingalgorithm. ;AvaluablenetworkperformanceanalysisiSdevelopedfor ;thealgorithm.Theresultsshowedthattheplan-based ;schedulingalgorithmWasmolepredicableanddepend

    ;able,andtherelativeresponsetimeiitterofthenetwork ;systemWasimproved.Anditismoreefficienttoimprove ;thenetworkutilizationofnativeCAN.eteastheplan. ;basedschedulingalgorithmisusedtoimproveflexibilityof ;timer-triggeredscheme(e.g.TFCAN).

    ;HIGHrECHN0L0GYIEIIRSIVo1.16No.1IMar.201093 ;

    ;

    ;

    ;

    ;8

    ;?

    ;

    ;numberofmessage(Mx)

    ;(a)TheWOL~b-caseresponsetime

    ;Fig.4

    ;numberofmessag(Mx)

    ;(b)Therelativeresponsetimejitter

    ;Analysisresultsofreal--timeperformance

    ;没立蕊0

    ;cAH1l

    ;Bulload【蒜]2678l

    ;P~].oad[嚣】2678

    ;{std.Datafr/s0

    ;staMS

    ;Fig.6StatisticsresullsscheduledbyPBS

    ;Finally.thteanalysisandsimulationresultsofanin. ;dependentlydrivingelectricvehicleverifiedthatthepm- ;posedPBSwaspracticalforrea1.timeandsafety..critic,al ;applications.ThenextworktoimprovePBSwillmainly ;includetheplaJ1optimizationandeITortoleranceandso ;nn.

    ;References

    ;[1]CenaG,Valenzano,VitturiS.Advancesinautomotivedigital ;ComputerStandards&Interfaces,2005,27: ;665-67

    ;l2JKhawarM,Zuberi,KangGShin.Schedulingmessageson ;controllerareanetworkforreal-timeCIrMalications.1EEE ;TransactioOFtRoboticsandAutomation,1997,13(2):3l0- ;314

;[3]LeenG,Heffeman.rrCAN:anewtimetriggeredcontroller

    ;areanetwork.MicroprocessorsandMicrosystems,2002,26:T’L-

    ;94

    ;[4]LuiS,AbdehaherT,ARZENK,eta1.Realtimescheduling ;theory:ahistori~alperspective.Real-Time~tem+,2004,28: ;i01155

    ;[5]HanssonHA,NoheT,Norstr6mC,eta1.Intergratingrelia-- ;bilitvandtiminganalysisofCANbasedsystem.1EEETrans.

    ;ndnoFtIndustrialElectronics,20o2,49(6):l2401250

    ;[6]JoaquimF,LIJfsA,Jos6AF,eta1.Combiningoperational ;flexibililWanddependabilil.tynmCAN.IEEETransactions

    ;onIndustriallnformatics,21006,2(2):95102

    ;[7]TindellK,BumsA,WellingsAJ.Calculatingcontrollerarea ;network(CAN)messageresponsetimes.ControlEngPrac

    ;tice,1995,3(8):1163.1169

    ;[8]CaoW,LinC,ZhouW,eta1.Areal-timeplanning-based ;schedulingpolicywithCANautomotivecommunicationsys

    ;terns.In:ProceedingofCISE,LosAngeles/Anaheim,Amefi

    ;ca.2009.186-.191

    ;nengXiang,assistantresearcher,bornin1979. ;HereceivedhisM.S.degreefromBeijingInstituteof ;Technologyin2005.NowheisaPh.DcandidateinNa-- ;tionalEngineeringLaboratoryforElectricVehicles.His ;currentresearchinterestsineludenetworkcontrolofdec

    ;trievehicle

    ;

Report this document

For any questions or suggestions please email
cust-service@docsford.com