In recent years, the number of private cars has grown rapidly, making road congestion, environmental pollution, and energy consumption increasingly serious. Carpooling is considered to be one of the effective ways to solve the problem of excessive motor vehicles on the road. Several travellers with similar travel routes can save money and reduce fuel consumption by carpooling, which improves vehicle utilization and reduces road utilization. Due to the widespread use of smartphones and mobile application development platforms, carpooling has broad prospects for development. In the carpooling problem, whether there is a high temporal and spatial correlation between the trajectorys is the key to determine whether or not to match. The purpose of this paper is to analyze the potential of carpooling, use the trajectory data, simplify the trajectory through the trajectory area gridization process, and propose several carpool models, and use this as a basis to analyze the trajectory data from time and space. The final solution model is a maximum weight non-bipartite graph match. The experimental results show that under the ideal carpool model, the carpool matching rate reaches 22-41%. With the continuous improvement of the carpooling model, the carpool matching rate increases by 7-9%. The car travel distance and average waiting time saved by carpooling are both It increases with the increase of the time threshold, which is in line with expectations. The results of this paper also confirm that there is a large possibility of carpooling in the city center residents.

CICTP