Dynamic OD Estimation And Application Research Based On RFID Data

With the continuous development of urbanization in China,the rapid increase of urban population and the increasing number of cars,the carrying capacity of urban transportation system is facing more and more challenges.Traffic problems are increasingly serious.But,with the rapid development of modern technology,many experts,the company in the industry use technology to develop urban traffic intelligence,and favor the support application model of the formation of intelligent network of vehicles,unmanned,intelligent City brain and other technical.OD estimation based on vehicle travel of the RFID data is one of its important research areas.This paper uses the city vehicle travel RFID data to carry out the road network prior OD matrix estimation and dynamic OD matrix prediction.The main work is as follows:Firstly,on the basis of outliers in raw data processing,this paper establishes a prior OD matrix calculated method based on vehicle travel RFID data through the steps of traffic plot division,RFID spot recognition among the cells and OD flow screening method.Secondly,for studying the method of obtaining dynamic traffic assignment matrix,this paper focuses on analyzing two main methods--simulation method and analytic method,and analyzing its applicable conditions.Then through analyzing the dynamic mapping relation between time-varying traffic flow and time-varying od flow,and using RFID data to get the travel time of road section,constructs the dynamic traffic assignment matrix estimation model based on analytic method.The system state vector of OD estimation is established based on the deviation idea,sets up the state equation and the equation of the OD estimation,and proposes the dynamic OD estimation Kalman filter model based on dynamic traffic assignment.Finally,taking the expressway interchange in Chongqing City West Ring interchange to South Ring Road interchange as the research object,and using the processed abnormal data for error data and redundant data received from RFID data of a certain month's vehicle travel,the paper uses the prior OD matrix calculation method based on the vehicle travel RFID data proposed below,with 20 min as the time interval,to obtain the vehicle prior OD matrix of the expressway.Based on this,the paper counts the travel time of each section in different periods,solves dynamic allocation matrix with the dynamic OD allocation model on traffic flow.And the dynamic OD of vehicle travel in this section is estimated.The results of dynamic OD estimation are evaluated by three kinds of error of MAE,RMSE and MAPE.The final results show that the total OD amount obtained by dynamic OD estimation model based on Kalman is fine.