Research On Urban Vehicle Travel Trajectory Extraction And Route Choice Behavior Based On Automatic Number Plate Recognition Data

With the continuous development of modern transportation network and the continuous improvement of road structure,more and more paths can be selected for a journey,and the result of path selection is bound to have more or less impact on the road traffic situation.Therefore,the study of path selection behavior has become an essential link in the field of transportation.Generally,path selection behavior is constrained by many conditions at the same time.Accordingly,the results of different path selection behavior can also reflect various factors affecting decision-making.Research on route choice behavior can not only help traffic managers improve road utilization,but also help drivers make more informed travel decisions.Urban vehicle travel trajectory contains a variety of traffic parameter information.By deconstructing the trajectory information,we can deeply depict and systematically restore the urban traffic travel scene,and provide data-driven decision support for the development and planning of urban road network.As one of the main sources of traffic collection data,automatic number plate recognition(ANPR)data has significant data characteristics and potential value.However,due to the inherent defects of ANPR data,it faces great difficulties in accurately extracting the trajectory.Focusing on the influencing factors of route selection behavior and the data characteristics of ANPR,aiming at the problems of small sample size,low visualization of analysis process and large granularity of ANPR data in traditional route selection behavior research,this paper aims to objectively analyze urban vehicle route selection behavior and obtain large volume and high reduction vehicle travel trajectory data by using ANPR data,The research on urban vehicle travel trajectory extraction and path selection behavior based on license plate recognition data is carried out.Firstly,clean the original data,determine the research area,establish a structured road network,deal with the abnormal values of travel time and travel speed,and design an algorithm to separate the travel chain;Secondly,the k-shortest path(KSP)algorithm is used to construct the reduction solution set,and four travel characteristic decision-making models are designed,including the consistency of time and distance,the number of signal lights,the current road congestion and travel inertia,An entropy weight technique for order preference by similarity to ideal solution and gray relational analysis(EW-GRA+TOPSIS)algorithm is proposed to optimize the decision model and realize the reconstruction and completion of travel trajectory.Finally,a measurement criterion based on matching degree is proposed,and the ANPR data of a region in Mianyang City are selected for experimental verification.The matching degree can reach 70.15% in the worst case.The comparative experiment is carried out by adding TOPSIS and GRA methods to verify the stability and excellent performance of the model algorithm proposed in this paper,and the visualization and travel characteristic analysis of each completion strategy are carried out,It is found that travelers’ travel decision-making tendency will change significantly with the peak and peak,which provides a new idea for the model analysis and improvement of follow-up research.