摘要
Detectinganddescribingmovementofvehiclesinestablishedtransportationinfrastructuresisanimportanttask.Ithelpstopredictperiodicaltrafficpatternsforoptimizingtrafficregulationsandextendingthefunctionsofestablishedtransportationinfrastructures.Thedetectionoftrafficpatternsconsistsnotonlyofanalysesofarrangementpatternsofmultiplevehicletrajectories,butalsooftheinspectionoftheembeddedgeographicalcontext.Inthispaper,weintroduceamethodforintersectingvehicletrajectoriesandextractingtheirintersectionpointsforselectedrushhoursinurbanenvironments.Thosevehicletrajectoryintersectionpoints(TIP)arefrequentlyvisitedlocationswithinurbanroadnetworksandaresubsequentlyformedintodensity-connectedclusters,whicharethenrepresentedaspolygons.Forrepresentingtemporalvariationsofthecreatedpolygons,weenrichthesewithvehicletrajectoriesofothertimesofthedayandadditionalroadnetworkinformation.Inacasestudy,wetestourapproachonmassivetaxiFloatingCarData(FCD)fromShanghaiandroadnetworkdatafromtheOpenStreetMap(OSM)project.ThefirsttestresultsshowstrongcorrelationswithperiodicaltrafficeventsinShanghai.Basedontheseresults,wereasonouttheusefulnessofpolygonsrepresentingfrequentlyvisitedlocationsforanalysesinurbanplanningandtrafficengineering.
出版日期
2017年04月14日(中国Betway体育网页登陆平台首次上网日期,不代表论文的发表时间)