Study Of Dynamic OD Matrix Estimaiton Based On RFID Data

With the continuous development of city and the rapid growth of vehicle ownership, congestion of the urban road is also increasing. Advanced traffic control and traffic management system has become one of the most important means to ease the urban traffic congestion. Dynamic origin-destination information, the link volume between origin and destination of the network during a particular time interval, is an important data input of upon traffic control and traffic management system, the acquisiton of it is very necessary. However, due to the restriction of traffic data detecting methods, experts always take use of estimating algorithm based on link volume to acquire dynamic OD information, and this study has been widespread concerned in and abroad. Aiming at this problem, this paper applies a new traffic data detection based on RFID technology, to research the realtime OD matrix estimation, the research content and results are as following.Firstly, this paper expounds the principle of RFID traffic acquisition technology, and summarise the current dynamic OD matrix estimating algorithm, points out that the priori OD matrix is important information of dynamic OD matrix estimation, and RFID technology can provide accurate priori dynamic OD matrix.Secondly, according to the principle of RFID acquisition technology, this paper choose a sample road network, collects real-time data, acquires the initial OD matrix based on RFID technology, and puts this initial OD matrix as priori OD matrix, carries out an all-round analysis of the results, which shows that the initial OD matrix is consistent with the actual time and space distribution characteristics, but since part of the vehicle does not load RFID tags, the initial results are slightly less on the numerical size.Thirdly, this pape uses the above priori OD matrix obtained by RFID technology, collects link volume at the same time, acquires the optimized dynamic OD matrix by the maximum entropy estimating method introducing Lagrange multiplier and genetic algorithms, analyzes the results as well, which shows that the optimized OD matrix is according to the priori OD matrix and corrected by link volume of a certain amount.Lastly, this paper prospects a research direction in the future.