Narrow Passage Watcher for Safe Motion Planning by Using Motion Trend Analysis of C-Obstacles

Abstract

Finding safe paths for robots in changing environments is a significant issue for motion planning. However, it could be fairly difficult when there are narrow passages in the configuration space. Solutions to this problem can be applicable to not only mobile robots but also other domains such as computer animation and computational biology. This paper presents a novel method called Narrow Passage Watcher (NPW) to cope with narrow passage issues in changing environments. It approximately predicts the variation trend of narrow passages and analyzes their security and thus guides to safe path planning. Meanwhile, a supporting hybrid boost strategy is presented to increase the sampling density inside narrow passages with different variation trend. Compared with existing work, the predictive mechanism provided by NPW gives the planner foresight so that it can find safer paths in changing environments with a higher success rate. Experiments conducted with a dual-manipulator system with 12 DOFs show that NPW can reduce the number of replanning times and total planning time remarkably as well as improving the success rate of path planning.