Study On Dynamic OD Acquisition And Traffic Assignment Of Urban Road Network Based On License Plate Recognition Data

Due to the accelerating process of urbanization and motorization,the travel demand of residents is increasing day by day,and the urban traffic congestion caused by this problem continues to affect the normal production and life of residents.In recent years,with the vigorous development of intelligent traffic system,traffic planning and traffic control by using intelligent traffic data to alleviate urban traffic congestion has been paid more and more attention by traffic managers.Dynamic OD can reflect the change of travel demand in a short time,and dynamic traffic assignment can obtain the traffic flow in each section of the road network in the future under the condition that the future dynamic OD has been obtained,which provides the data basis for the formulation of time-varying traffic control scheme,therefore,the study of dynamic OD and dynamic traffic assignment is of great significance to traffic planning and traffic control.At present,there are few studies on dynamic OD acquisition,and most of them do not consider the actual operation of vehicles,at the same time,the dynamic traffic assignment model is still difficult to be applied to large road networks,therefore,based on the license plate recognition data collected by intelligent traffic equipments,this paper will carry out the study on dynamic OD acquisition and dynamic traffic assignment methods,in order to put forward the method of dynamic OD acquisition of general urban road network and the dynamic traffic assignment model with practical application prospect.Firstly,this paper expounds the problems of the original license plate recognition data,puts forward a generally applicable data preprocessing process combined with the actual traffic phenomenon,and then analyzes the data quality of the basic data.Then,considering the actual operation of vehicles,based on Dijkstra algorithm,road section travel time sets and intersection signal cycle data,this paper puts forward a process of using dynamic time threshold to identify the travel stop point and divide the number of vehicle travels,and then obtains the travel trajectory of the single travel of the vehicle,which provides a data source for the subsequent dynamic OD acquisition and dynamic traffic assignment.Next,this paper acquires the dynamic OD by using the complete and the incomplete travel trajectory,and the results of the dynamic OD can reflect the change of traffic travel demand of the road network in a short time;in order to solve the problem that the incomplete trajectory will cause data noise to the study on dynamic traffic assignment based on the historical dynamic OD and the historical traffic flow,a method of combining the passing record of the missing trajectory and the topological structure of road network is proposed to complete the travel trajectory with single point missing and two consecutive points missing,and then the historical data sets for the subsequent dynamic traffic assignment are obtained.Finally,this paper proposes to use CapsNet to capture the mapping relationship between the historical dynamic OD and the corresponding historical traffic flow,and then establishes a dynamic traffic assignment model based on neural network,through test set verification,model comparison and applicability analysis,it is shown that the model proposed in this paper has the prospect of practical application.