Segment Planning of Multiple Escape Paths in Different Narrow Passage Scenarios
J.S. Wang, Tao He, Chao Xu · 2024
The escape problem in autonomous driving means that the robot will enter the escape mode when it detects that there is an obstacle ahead and it cannot pass safely according to the planned route. At this point, the robot needs to re-plan its route, so the escape path planning problem aims to find a path that can get out of the current predicament and return to the expected trajectory. In this paper, we propose a planning method for solving escape paths in complex scenarios. First of all, this method greatly improves the search efficiency of hybrid A* and avoids some redundant node expansion. Secondly, this method improves the availability of paths on narrow roads while ensuring search efficiency, and does not cause paths to be inaccessible due to space limitations of narrow roads. Finally, the splicing method between each path segment is proposed. Compared with traditional escape path planning methods, the cost of our planning solution is reduced by more than 85%.