Abstract: In order to improve the understanding and analysis of motion patterns of vehicles, a hierarchical trajectory clustering algorithm is developed according to spatial and temporal characteristics of the vehicle trajectories in the entrances and exits of freeway. In view of the vehicle trajectories are different in length, but in the same direction, the improved Hausdorff distance was proposed and applied to measure the similarity of trajectories. The improved fuzzy C-means hierarchical clustering algorithm of trajectories was further established, in which trajectories were first clustered into different paths according to the spatial geometric position of the trajectories, and then trajectories belonging to the same path were further clustered according to the vehicle speed to obtain the final results with spatial and temporal degree. Experiments in the entrances and exits of freeway were carried out. The results confirm that the proposed trajectory clustering algorithm not only has strong adaptability to the inherent motion pattern of the scene, but also ensures the accuracy and reliability of the clustering results.

Key words: engineering of communication and transportation system, freeway entrances and exits, trajectories analysis, improved Hausdorff distance, clustering algorithm