A DECOUPLED APPROACH FOR MULTI-AGENT PATH PLANNING ON GRID MAP

Abstract

In the recent years, there is a lot of interest have been have been shown by the researchers in multi-agent path planning (MAPP). It applies in many real world applications such as logistics, video games, warehouse management, search and explorations, etc. Multi agent path planning approaches can be divided into coupled and decoupled approaches. In this paper, we present a decoupled approach for multi-agent path planning. It works in two phases: individual path planning and path coordination. Path coordination is done by a collision avoidance algorithm (CAA) proposed in this paper. We have implemented and compare the performance of the proposed algorithm with some existing approaches. The results show the effectiveness of our algorithm.