고밀도 로봇군을 위한 다수 로봇 시스템의 경로 계획

Abstract

This paper addresses path planning problem of multi-robot system for high density robotic workcell.We propose an algorithm that uses decoupled path coordination method with new velocity tuning approach and centralized path coordination approach together to solve a specific problem with multiple and sequential goals. Given the multiple points as algorithm input the planner computes collision-free paths of robots for the given task. We basically use Rapidly-exploring Random Trees(RRT) to get a path for the each step. We proposed tree reuse method to find paths for the multiple goals and also propose new velocity tuning and bias sampling method in decoupled multi-robot coordination method. The velocity tuning method uses the notion of coordination space. It tries to make efficient coordination path using maximum slope in coordination space. The bias sampling method samples more nodes in specific region which is near to the goal region in coordination space for each robot. It is used for specific two cases, same goal problem and the problem with large gap of two robot’s work. A simulator is developed to verify the proposed algorithms. The simulation results shows that the propose method can make a coordination path more quickly than the basic RRT algorithm.