基于多源异构数据的高速路网交通拥堵状态识别

打开文本图片集

中图分类号：U491.265 文献标志码：A文章编号：1003-3114（2025）04-0807-08

doi：10.3969/j. issn.1003-3114.2025.04.018

Abstract：Withthecontinuous developmentofcoretechnologiesintheInternetof Things，suchasradiofrequencyidentification and5Gcommunication，highwaytraffccongestionstaterecognitionhasintroducedmoreheterogeneousreferencesources．Toavoid complexityandusabilitychallengesoftraditionalvisualandGPSmethods，ahighwaytraficcongestionstaterecognitionalgorit frameworkbasedonulti-sourceheterogeneousdataisproposed.Consideringthecharacteristicsof5Gandradiofrequencyidentificationwirelesscommunicationetworkdataaneficientdatapreprocessngalgoitisdsignedtoimproedataqualityandreliability. ForthepreprocessedmobilesignalingandETCgantrydata，afusiondistance-weightedalgorithmandanS-shaped3-parameter（S3） trafiflowodelareutildtovaluatetetraficongestionstatefoadsegmentsMoreoveadecisio-evelfusinmetodisde signedtoacuatelydescribetetraficfowspdandcongestionconditiosatteghwalevel.inally，hroughcasesudynd verificationusingactualdatafromtheentirehighwaynetworkinZhejangprovince，theprecisionandrecallrateoftraffccongestion state recognition reached 93.6% and 95.26% ，respectively， showing a significant improvement compared to single-source methods. Theproposedalgoritdemonstratesitsefectivenessandpracticalityinthefeldofhghwaycongestiostaterecogitionandcanpro vide a basis for corresponding road network traffic management decisions.

Keywords：congestion state recognition; data fusion； IoT；mobile signaling

0引言

受车流量持续增长、节假日免费通行等因素影响，交通拥堵问题日益突出，如何实时、动态地掌握拥堵路段的空间分布、拥堵程度、持续时长等道路实际通行状况，是国内外研究的热点。（剩余9966字）