您好,欢迎访问三七文档
当前位置:首页 > 商业/管理/HR > 销售管理 > 短时交通预测的动态出行信息服务协同工作平台
115200910JOURNALOFGEO2INFORMATIONSCIENCEVol111,No15Oct1,2009:2009202212;:2009208230.:863(2006AA12Z2092007AA12Z241)(40871184)(088RA400SA):(1970-),,,LBSGISE2mail:luf@lreis1ac1cn,,(,100101):,,,,;GIS;,,,:;;;;1GIS(ITS)(LBS)GIS,,,,,[1],,,,,,,RDSGPRSDAB,,,2,,,;,,1,1Fig11Architectureofthecollaborativeworkingplatformforreal2timetravelinginformationservice,,,,GIS,GIS,,,,,,GIS,GIS,,,(),,,,GISGIS,GIS,Socket,GIS,,,,,,GIS,GIS,,3,,,,,,,,,,,[2],ARIMAKalman-8162009[327],BPCorSim,BP,,,,,,,BP,,(),,,,()[8],,,,,[9],BP,,,,,GIS[10,11],TSIS/CorSim,,,,,,();,/,CorSim,(),,CorSIM/;;,,BPCorSimGIS;,/;,4411GIS,GIS,,,,GIS,GIS,,,,Dijkstra[12],,:9165:(1),,T;T(i),T(i)=iT,i0;L(k)k;V(i)(k)ik;t(k)(0)k;t(k)(1)t(k)(0);(2),,Dijkstra;(3),t(k)(0):i=[t(k)(0)-T(0)]T+1,kt(k)(0),t(k)(1),:t=t(k)(1)-t(k)(0)tDijkstra,,,,,,412;20077-9,20077-95,,,,Oracle9iVisualC++610CorSim,CorSim,BP,,151115%15Tab11Averageforecastingerrorsfordrivingspeedona5minuteperiod010950109401217010940108201119011020111522(a),2(b)3,2(c)8,2(d)1,,,,,,,,,,,,5,,,,,,,,,,,,,02620092Fig12Dynamicoptimalroutesbasedonforecastedtrafficsituation,Socket,GIS,,,:[1],,..,2000(1):20-23.[2],..,2004,4(4):68-71.[3]YinH,WongS,XuJ.UrbanTrafficFlowPredictionUsingaFuzzy2neuralApproach.TransportationResearchPartC,2002(10):85-98.[4],,..,2003,23(12):7-8.[5],,..,2006,6(2):70-74.[6],..,2006,32(17):48-51.[7]..:,2004.[8],,..,2003,15(8):1179-1183.[9]LuF,DuanYYandZhengNB.APracticalRouteGuid2anceApproachBasedonHistoricalandReal2timeTrafficEffects.ProceedingsofGeoInformaticsFairfax,U.S.2009,12-14.[10],,.7.,2007,25(1):66-70.[11],,,..,2008,33(2):38-41.[12],,.Dijkstra.,1999,4(12):1044-1049.1265:ACollaborativeWorkPlatformforDynamicTravelInformationServicewithShort2termTrafficPredictionLUFeng,DUANYingying,ZANGZhigang(LREIS,InstituteofGeographicScienceandNaturalResourcesResearch,CAS,Beijing100101,China)Abstract:Implementingconvenienttravelinginformationserviceisacrucialtaskforcarryingoutthestrategyofin2telligenttransportationsystemanddevelopinglocation2basedservices.Atpresent,mostofthedomestictravelinginformationservicesystemsonlyproviderelativelystaticinformationwhichcanptreflectthepossibleshort2termchangesoftraffic,andresultinverylimitedpracticaluse.Althoughtherehaveemergedsomecarnavigationprod2uctsandotherapplicationsinvolvingreal2timetrafficinformation,consideringtherapidchangeofcitytrafficsitua2tion,theseproductsandapplicationsstillfacepracticaldifficultiesforalltheinformationreceivedreal2timelywillgetoutdatedwithinafewminutes,whichmakesthesocalleddynamicapplicationsbasicallytime2slicelimitedstat2icones.Thispaperpresentsapracticalshort2termtrafficpredictionapproachinreal2timeconditionsbyintegratinghistoricaltrafficbasedstatisticalreasoningwithbackpropagationneuralnetworkbasedanalyticalmodel,andcom2mercialmicroscopictrafficsimulationsoftware.Thehistoricaltrafficbasedstatisticalreasoningutilizestheinherenttrafficrulesbyidentifyingthegeneralspatial2temporaldistributionpattern,andignoresanyinputsthatdonptcon2tributetotheoutputduringthetrainingprocess,includingthegrosserrorinthecollectedtrafficdata.Thecom2mercialmicroscopictrafficsimulationsoftwareprocessthetrafficabnormitiesthatalwaysexistinbigcities,andhasnorequirementforthecollectedreal2timetrafficinformationtocoverthewholeroadnetwork,henceprovidesanef2fectivesupplementforshort2termtrafficforecast.ThenanapproachisdevelopedcombiningGISserver,trafficpre2dictionserveranddatabasemanagementsystemtoimplementdynamicrouteguidance.Thetrafficpredictionserverreceivesreal2timetrafficinformationobtainedfromfloatingvehiclesandachievesshort2termforecastingresultsforthewholeroadnetworks,thenfedtheresultsbackintothedatabasemanagementsystemandGISserver,sothatatime2dependantoptimalroutingcanbeconductedthroughadynamicleasttravelingtimealgorithmdevelopedinthisstudy.Aprototypefulfillingtheaboveaspectshasbeendevelopedandvalidatedwithacityroadnetworkandreal2timelycollectedtrafficinformation.Theapproachpresentedinthispaperisarguedtoprovideapracticalsolutionforreal2timepublictravelinginformationserviceanddynamicwebmaps.Keywords:citytraffic;short2termtrafficprediction;historicaldatareasoning;microscopicsimulation;dynamicrouteguidance2262009
本文标题:短时交通预测的动态出行信息服务协同工作平台
链接地址:https://www.777doc.com/doc-1598927 .html