ABSTRACT
Studying the trajectory tracking of vehicles in tunnels can help enrich microscopic traffic information and improve driving safety. In this paper, a method for perceiving vehicle trajectories in tunnels based on LiDAR data is proposed. The point cloud data is collected by building a LiDAR on the roadside of tunnel. The background point cloud filtering is completed by selecting the region of interest and filtering the ground lines. The driving trajectory of the vehicle in the tunnel is obtained, and the vehicle tracking is completed through target clustering based on DBSCAN algorithm and the data association based on GNN algorithm. The innovation of this paper is to realize the application of vehicle trajectory tracking technology in tunnel scene. The feasibility of this method has been verified through field experiments.
