ImprovingTrafficPredictionbyIncludingRainfallData

ImprovingTrafficPredictionbyIncludingRainfallData.pdf
Weatherconditionstendtohavemeasurableimpactontrafficconditionsoftheroads.Thisrelationshipiscommonlystudiedatthenetworklevelwithoutexplicitexplanationofthelinkperformances.Furthermore,existingstudiestypicallyusehighresolutiontrafficdatawhichmaynotbeavailableacrosstheentirenetworkandespeciallyduringtheadverseweatherconditions.Inthisstudyweexploretheimpactofrainfallintensityonlow-resolutionspeedbanddata.Wealsotestwhetherthisadditionalinformationaboutrainfallmayimprovethepredictionaccuracyofdata-drivenmodelsforindividualroads.Todoso,weincorporatetheinformationabouttherainfallintensityintosupportvectormachine(SVM)predictionalgorithm.Asabenchmarkweonlyconsidertemporalfeaturestopredictnearfuturetrafficconditionsduringrainyweather.Numericalresultsfor616roadsegmentsinSingaporeconfirmthatrainfallimpactstrafficconditionsintermsofdecreasingthedrivingspeed.Thisreductionincreaseswiththerainintensity.Furthermore,theresultsshowthatadditionalrainfalldataenhancesthepredictionaccuracyforcertainnumberoflinks;whilefortheotherstherainfallinformationisnotthatuseful.
作者：LeongWaiLeong KelvinLee KumarSwapnil XiaoLi HoYaoTongVictor NikolaMitrovic MuhammadTayyabAsif JustinDauwels PatrickJaillet
作者单位：SchoolofElectricalandElectronicEngineering,CollegeofEngineering,NanyangTechnologicalUniversity,SingaporeCity639798,SingaporeDepartmentofElectricalEngineeringandComputerScience,SchoolofEngineering,Cambridge,MA02139,USA;OperationsResearchCenter,MassachusettsInstituteofTechnology,Cambridge,MA02139,USA;CenterforFutureUrbanMobility,Singapore-MITAllianceforResearchandTechnology,SingaporeCity138602,Singapore
母体文献：2015年南京第十四届亚太智能交通论坛论文集
会议名称：2015年南京第十四届亚太智能交通论坛  
会议时间：2015年4月27日
会议地点：南京
主办单位：交通运输部
语种：chi
分类号：
关键词：ITS  intelligenttransportationsystems  trafficprediction  SVM  rainfallcorrelation
在线出版日期：2019年1月18日
基金项目：
相似文献
相关博文

ImprovingTrafficPredictionbyIncludingRainfallData.pdf

文件大小:

1.75 MB

高速下载