A Hierarchical Trajectory Planning Framework Based on 3D Spatiotemporal Coupling

Trajectory planning is critical in autonomous driving systems, as it generates feasible driving paths in dynamic 3D spatiotemporal environments. This paper presents a hierarchical traj ectory planning framework based on 3D spatiotemporal coupling to improve algorithm efficiency and adaptability across various environments. The framework first employs a kinematic model-based sampling method and heuristic search to compute a reference traj ectory that ensures safety and considers vehicle motion coordination, providing strong adaptability to dynamic obstacles. Key trajectory points are then selected for optimization through variable-step sampling, reducing the problem's dimensionality. Kinematic, obstacle, and control constraints are incorporated to enhance trajectory safety and smoothness. Extensive simulations across different scenarios demonstrate that the proposed method outperforms traditional approaches, offering superior robustness, real-time performance, and adaptability.