Flexible Trajectory Planning for Autonomous Vehicles Via Environmental Assessment in Extreme Scenarios
Xiang Li, Ke Lin, Xiaoqing Yang, Kejian Yan, Yanjie Li, Yunjiang Lou
AI summary
Problem
Existing trajectory planners struggle to balance computational efficiency, safety, and scene generalization in diverse extreme scenarios with unstructured obstacles and tight maneuvering spaces.
Approach
The method uses an environmental assessment-guided Hybrid A* algorithm to quickly generate coarse initial trajectories, then refines them through an optimal control problem that adapts sampling time and constructs safety corridors for robust spatio-temporal optimization.
Key results
- EAHybrid A* algorithm reduces computational cost by assessing environmental complexity to guide search direction
- Optimal control formulation with safety corridors and adaptive sampling time ensures kinematic feasibility and obstacle avoidance
- Higher success rates and planning speeds compared to state-of-the-art planners in extreme scenarios
- Validated in CARLA simulator and real-world autonomous vehicles, demonstrating robustness across open and urban environments
Why it matters
Provides a practical, generalizable planning solution for autonomous vehicles operating in complex, unstructured environments like mining sites and dense parking lots.
Abstract
Trajectory planning is a core task in autonomous driving. However, in diverse extreme scenarios characterized by unstructured obstacles, there is a lack of solutions that provide efficient computation, safety, and scene generalization capabilities. To address this issue, we propose a two-stage spatio-temporal joint trajectory planning method based on environmental assessment. In the first stage, we introduce the EAHybrid A* algorithm, which generates high-quality initial trajectories by evaluating environmental complexity, thereby significantly improving computational efficiency. The second stage formulates the trajectory planning problem as an optimal control problem, utilizing environmental assessment for joint spatio-temporal optimization, ensuring kinematic feasibility and obstacle avoidance. Experiments demonstrate that our method achieves higher success rates and planning speeds in extreme scenarios compared to state-of-the-art planning methods. More- over, we have deployed and validated this approach in the CARLA simulator and real vehicles, proving its effectiveness and robustness in handling extreme environments.