Research On Multi Lane Vehicle Lane Change Of Expressway

In this paper,a new multi-Lane vehicle lane changing model is proposed to solve the problem of vehicle lane changing on expressway.This model combines the decision-making model of lane changing based on support vector machine(SVM)and the conserved high order traffic flow model(CHO)based on Lagrange coordinates.It is driven by data and supported by traffic flow theory to study the problem of multi Lane vehicle lane changing.In this paper,we first transform the CHO model in Euler coordinates and Lagrange coordinates,and propose a CHO full discrete car following model.Secondly,we analyze the advantages of SVM in lane changing decision-making,and construct the original data set through the rule-based multi lane traffic flow model,using Z-score data standardization and Synthetic Minority Over Sampling Technique(SMOTE),Considering the cost difference of misclassification of vehicle lane changing data,a two index evaluation system of classification accuracy and "Receiver Operating Characteristic"curve is proposed to adjust the parameters of SVM,and an effective multi Lane vehicle lane changing model is established.In addition,through the simulation of two kinds of common scenes of high speed and low density,low speed and high density,this paper analyzes the change of dynamic simulation process and lane change rate,and the simulation results show that the vehicle lane change model based on SVM and CHO model can make lane change decision accurately according to the current driving environment,and effectively simulate high speed and high density The real multi lane traffic situation on high-speed highway has a certain guiding significance for the research of traffic intelligence.