バスプローブデータを用いた豊田市の道路渋滞分析に関する研究
43/48

37 9Data Introduction and Preparation•WeusedprobedataoftheBLSequippedinOidenbusesofthisroute.•ThetimeperiodofbusprobedataisfromJanuary1sttoJune30th,2016for6months,andthecontentsofprobedataincludetimestamp,longitude,latitudeandpositionwithacollectionintervalof30s.•Wedividedtheweekdaysingeneralintotwotypes,workingandnon-workingdaysofTMC.Moreover,theimpactofcommutingbehaviorisalsoconsideredreferringtopeak(6a.m.to9a.m.)andoff-peak(others)periods,respectively.()(2)((3,)/(3,4))*((4)(2))TSCTSPDPCDPPTSPTSPProcedure to estimate the travel time:1.Generate the fixed bus route digital data based on the Digital Road Map.2.Match the bus probe data into the nearest road link to modify the position.3.Calculate the distance of adjacent bus GPS data, and estimate the timestamp of influx and outflux nodefor each road link between the matched GPS. Then, we can estimate the travel time between nodes in DRM straightforward.10Methodology•ThisstudyappliestheGaussianmixturedistributiontofitthetraveltimedistributionoftheobservedroadbottleneck.•TheGaussianmixturedistributionisconstitutedoftwoGaussiandistributions,whichrepresentstraveltimeinthecongestedstateandthatintheuncongestedstate.0.000.040.080.120.160.200.240501001502002503003504004505005506006507007508008509009501000Probability DensityTravel time (Unit: second)histogramGaussianmixture

元のページ  ../index.html#43

このブックを見る