Simulated capacitor method for difficult region dynamic boosting in changing environments

This paper presents a novel method to identify and boost difficult regions of in the configuration space (C-space) in changing environments. Difficult regions, especially narrow passages, change their shapes frequently in changing environments, which result in challenging problems to find valid and safe paths. Although a lot of research has been done to identify difficult regions, seldom methods provide robust paths. Moreover, they depend on frequent replanning which wastes a lot of resources. In this paper, a bridge test method based on workspace to configuration space (W-C) nodes mapping is invoked to identify difficult regions dynamically. Consequently, not only difficult regions are identified efficiently, but safe regions which are less likely to be occupied temporarily, are flagged for boosting. Furthermore, the calculated simulated capacitance, which represents the local difficulty level is used to lead boosting procedure. Specifically, boosting nodes would be activated or re-closed according to the changing capacitance, so that calculation resources are concentrate on current difficult regions. As a result, both the total planning time and replanning times are descended a lot. Finally, the simulation experiment with two manipulators shows the proposed method is efficient even in difficult changing environments.

I INTRODUCTION

Sampling-based methods, such as Probabilistic Roadmap Method (PRM)[1] and Rapidly exploring Random Trees (RRT)[2], contributes a lot in robotics path planning. How-ever, sampling-based methods show their weakness in some special regions, such as narrow passages and obstacles' boundaries, due to their small volumes. In recent twenty years, plenty of variants of PRM and RRT [3], [4], [5], [6] have been proposed to deal with these difficult region problems and have achieved great success even in high-dimension configuration space (C-Space). Nevertheless, there are still challenging problems in changing environments, because difficult regions change their shapes when obstacles move.