Carpooling is considered an important solution to alleviate traffic congestion and reduce air pollution in cities. In this study, trajectory cleaning is performed on the data set, trajectory reconstruction is performed on adjacent trajectory breakpoints which larger than the entire sub-region, and time-lag trajectory data is eliminated, so that points of the trajectory are divided into sub-regions by coordinates, finally,the matching of the trajectory points is converted into the matching of the nodes of each sub-region. Then, based on the perfect match and the starting point match (O-D match), a trajectory adjustment similarity matching mode that does not change the total travel distance is explored, and all matching matches are screened by the space-time similarity recognition algorithm based on the longest common subsequence (LCS). The trajectory matching pairs are calculated using the blossom algorithm to achieve the maximum trajectory matching under the influence of different waiting times on the train. Furthermore, the reduction of the total trajectory and the carpool matching rate corresponding to different periods can be calculated by the trajectory reduction, Improve carpooling efficiency within a reasonable delay of pick-up time thresholds.

CICTP