Trajectory planning for AGV based on the improved artificial potential field- A* algorithm

Abstract There are many redundant nodes and inflection points in the path planned by the traditional A* algorithm, leading to the inefficient trajectory planning of the automatic guided vehicle (AGV) in the multi-static obstacles environment. The artificial potential field (APF) algorithm suffers from the problem of unreachable objectives and falling into optimal local value. This article studies the trajectory optimization of AGVs to improve the trajectory planning algorithm’s iteration efficiency and shorten the trajectory’s total length. This article establishes the forward kinematic and unified robot description format model of the AGV and proposes the APF-A* algorithm for trajectory planning. The search cost and the number of turns are effectively optimized. The article simulates the APF-A* algorithm, the results are compared with the trajectory before optimization, and the optimized time is 60% less than that before optimization. The experimental platform of AGV trajectory planning is built, and the algorithm verification experiment of AGV trajectory planning is carried out. The experimental results show that the algorithm studied in this article achieves path smoothing and trajectory length optimization.