Hierarchical Path Planning and Motion Control Framework Using Adaptive Scale Based Bidirectional Search and Heuristic Learning Based Predictive Control
Guodong Du (Beijing Institute of Technology, ETH Zürich)
Yuan Zou (Beijing Institute of Technology)
Xudong Zhang (Beijing Institute of Technology)
Zirui Li (TU Delft - Transport and Planning)
Qi Liu (Beijing Institute of Technology)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
Autonomous vehicles have been used for a variety of driving tasks, in which path planning and motion control are important research parts to realize the autonomous driving. A hierarchical framework consisting of path planning and motion control of the vehicle for non-specific scenarios is proposed in this paper. Firstly, the description and the formulations of the problem are given, and the corresponding models are constructed. Then, the logical construction of proposed framework is expounded with several logical associations and algorithmic improvements. The bidirectional heuristic planning with adaptive scale search is designed and incorporated with robust weighted regression algorithm to plan the optimal global path, while the multi-step predictive control method based on heuristic reinforcement learning algorithm is proposed to improve the effect of the motion control. The results show that the proposed framework for autonomous driving achieves better performance in both path planning and motion control than several existing algorithms and methods. The adaptability of hierarchical framework is demonstrated. Furthermore, the effectiveness of the hierarchical framework in real world scenario application is also validated.