Abstract:

For insufficient consideration of trajectory safety in existing spatio-temporal joint trajectory planning methods for unstructured environment， a spatio-temporal joint trajectory planning method based on spatio-temporal corridors is proposed in this paper to enable the direct generation of feasible and optimal trajectories within spatio-temporal constraint spaces. Firstly， a motion primitive tree is constructed based on the vehicle kinematic model， and an initial trajectory satisfying the vehicle’s kinematic constraints is obtained through a search of the motion primitive tree. Subsequently， a 3-D spatio-temporal corridor is generated based on the initial trajectory， providing a safe solution space in the spatio-temporal domain for trajectory optimization. Finally， within the spatio-temporal corridor， constraints such as dynamics， safety， and continuity are comprehensively considered， and the trajectory is optimized using a segmented Bézier curve method. The simulation and real-world test results show that the proposed trajectory planning method demonstrates significant advantages over the TEB algorithm in terms of safety and flexibility， while maintaining superior computational real-time performance.

Key words: intelligent vehicle, unstructured environment, trajectory planning, spatio-temporal corridor