Path Planning for Automatic Guided Vehicle with Multiple Target Points in Known Environment

Abstract

Today, Automatic Guided Vehicles (AGVs) with a path planning algorithm are being used in many industrial fields. There are A*, D*, and D* lite algorithms in the path planning algorithm. In this paper, propose a modified D* lite algorithm using the most efficient D* lite among these algorithms. The modified D* lite path planning algorithm is proposed to improve these D* lite path planning algorithm’s weaknesses such as traversing across obstacles sharp corners, or traversing between two obstacles. The modified D* lite path planning algorithm has function to set target points differently from the existing D* lite path planning algorithm. To do this task, the followings are done. First, a work space is divided into square cells. Second, cost of each edge connecting current node to neighbor nodes is calculated. Third, the shortest paths from the initial point to all multiple target points are computed and the shortest paths from any target point to remaining target points including the goal point are computed by using Hamilton path. Fourth, a cost-minimal path is re-calculated as soon as the laser sensor detects an obstacle and make an updated list of target points. Finally, the validity of the proposed modified D* lite path planning algorithm is verified through simulation and experimental results in known environment.