Abstract: Extracting the driver's experience from taxi trajectories to guide the public travel is an important applica- tion of trajectory data.In order to avoid the inaccurate results caused by ignoring the origin and destination in the experience extracting process of traditional methods,this paper proposes an origin destination constraint experience model (ODCEM). Then,the traditional experience model,the ODCEM and the shortest path model are com- pared by the experiments in Shenzhen city.The results shows that (1) the ODCEM can successfully simulate the taxi driver's travel experience during the experience extracting and the path planning; (2) in comparison with the other two methods,the ODCEM increases the travel time and the speed respectively by more than 93. 3% and more than 76. 7%; and (3) the travel distance of the ODCEM is between those recommended by the other two methods.Finally,the influence of the constraint scale is discussed.It is found that the desirable constraint size is 2km ×2km.

Key words: origin destination constraint, taxi trajectory, experience model, path planning