[28] 2011 Estimation of Multivehicle Dynamics by Considering Contextual Information

《[28] 2011 Estimation of Multivehicle Dynamics by Considering Contextual Information》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、IEEETRANSACTIONSONROBOTICS,VOL.28,NO.4,AUGUST2012855EstimationofMultivehicleDynamicsbyConsideringContextualInformationGabrielAgamennoni,Member,IEEE,JuanI.Nieto,Member,IEEE,andEduardoM.Nebot,SeniorMember,IEEEAbstractHumandriversareendowedwithaninborna

2、bil-dictionandanticipatorydriving.Itisindispensableforforward-itytoputthemselvesinthepositionofotherdriversandreasonlookingdriving-assistancesystemsandautonomousvehiclestoabouttheirbehaviorandintendedactions.State-of-the-artdriving-incorporatesemanti

3、csandhigh-levelcontextualinformationasassistancesystems,ontheotherhand,aregenerallylimitedtophys-well.icalmodelsandadhocsafetyrules.Inordertodrivesafelyamongsthumans,autonomousvehiclesneedtodevelopanunderstandingThereasoningcarriedoutbyhumandriversca

4、nbemodeledofthesituationintheformofahigh-leveldescriptionofthestateusingprobabilitytheory.Probabilisticmodelscanbridgethegapoftrafficparticipants.Thispaperpresentsaprobabilisticmodeltobetweenlow-levelsensorreadingsandhigh-levelcontextinfor-estimatethe

5、stateofvehiclesbyconsideringinteractionsbetweenmation,whileatthesametimeprovidingasuitableframeworkdriversimmersedintraffic.Themodelisdefinedwithinaproba-forhandlinguncertainty.Thecombinationofdifferentlayersofbilisticfilteringframework;estimationandpre

6、dictionarecarriedoutwithstatisticalinferencetechniques.Memoryrequirementsin-abstractionyieldsaholisticapproachcapableofsimultaneouslycreaselinearlywiththenumberofvehicles,andthus,itispossibleinferringtheposeofthevehicles,aswellastheircontext.Thistosc

7、alethemodeltocomplexscenariosinvolvingmanypartici-allowsforanunderstandingofcomplexsituationsandtherebypants.Theapproachisvalidatedusingreal-worlddatacollectedresultsinamorerobustestimationofthecompletestateofbyagroupofinteractinggroundvehicles.traffi

8、cparticipants,whichisthebasisforprobabilisticdecisionIndexTermsAgentinteraction,anticipatorydriving,driverbe-making.havior,intelligenttransportationsystems,roadsafety,situationalawareness.II.RELATEDWORKEstimatingthestateoftrafficparticipants,oragents,