Dynamic OD estimation using Automatic Vehicle Location information

OD (Origin-Destination) matrix describes the number of vehicles from origin zones to destination zones. Dynamic OD is an essential input for traffic guidance and traffic control strategies of intelligent transport system (ITS). As the development of Automatic Vehicle Location (AVL) technologies, advanced devices such as GPS, RFID etc. have been used in locating vehicles and we can obtain real-time information from vehicles equipped with AVL devices. With the information provided by AVL vehicles, we propose a model named OD estimation with automatic vehicle location (O DAVL) to estimate dynamic OD matrix. The ODAVL model uses AVL data and abandons history data which ensures that all data used are real-time and avoids the uncertainty of history data. ODAVL model also considers the link flow error between real flow and the flow we estimated which caused by the deviation of real OD and estimated OD that could improve the precision of OD estimation. Experiments show that our method obtains a good solution of OD estimation.