贝叶斯线性回归的桥梁移动车辆荷载识别算法

打开文本图片集

Abstract：Atpresent，vehicleloadsaremainlyobtainedtroughbridgeweigh-inmotion（BWM）system.TheMosesalgorit，wichisthe coreofBWIMsystm，hasoverfitingproblmsenidentifingaxleswithrelativelyclose welbses.Itleadstolwrecogitiocuracy. Toadesthisproblem，thispaperpropossanovelBWalgoritbasedonBayesianliarregresso（BL）.Tealgorithuste initialdistributionofaxleweightsandmeasurementerror，andutilizesBLRtoiterativelyupdatetheaxleweightsofvehicles.Itcan theoreticallyeducetheinfluenceofoveftingonterecognitionresultsandimproveteauacyofaxleweights.Tefeasibilityand robustnesoftheBLRlgoritareverifidbycomparingthcogtionresultsofMosesalgorith，Tikhoovgularzatioalgoritmand BLRalgorittoughnumericalsiulatiosandin-situtestResultssowtheproposedalgorithasigherrecognitionacuracygood stability，wider applicability，and can be better applied to the intelligent management of bridges.

Keywords：bridge weigh-in-motion;BLRalgorithm;least squares;overfiting;measurement error

车辆荷载作为桥梁结构主要的活荷载，在结构设计中有着举足轻重的地位。（剩余17039字）