Data Mining Method of Sequential Patterns for Vehicle Trajectory Prediction in VANET.

In order to provide the future direction of vehicle flow for travelers, it is necessary to predict the driving trajectory of vehicles running on the road in a certain segments. According to the characteristics of vehicle sequential pattern, the topology of vehicular ad hoc network (VANET) is linked to the real road vehicle movement trajectory. Some new definitions related to sequential patterns in VANET environment are proposed. Based on RSU and V2V schemes, a vehicle movement database is established, and sequential pattern data mining is carried out. Afterward, their communication overhead is evaluated. The support and confidence of the movement rules generated by vehicle routing patterns are calculated to extract the probability of frequent driving trajectories. [ABSTRACT FROM AUTHOR]