Improved A* algorithm incorporating RRT* thought: A path planning algorithm for AGV in digitalised workshops

Abstract

The implementation of efficient path planning algorithms can facilitate enhanced transport efficiency in AGVs (Automated Guided Vehicles). However, the current path planning algorithms used in the workshop can determine the shortest path, but there are problems such as low search efficiency, long search time, and many path inflection points. Such issues have the potential to negatively impact the transport efficiency of the AGV, rendering them unsuitable for actual workshop conditions. This paper proposes a new algorithm, named IA-RRT* (Improved A* Algorithm Integrating RRT* Thought), for path planning of AGV in digitalised workshops. The algorithm modifies the cost evaluation function of the A* algorithm to provide a less strong direction for the search, narrowing the search scope and preventing the algorithm from converging too quickly and getting stuck in a local optimal path. Meanwhile, the IA-RRT* algorithm combines the randomness concept of the RRT* algorithm with an inflection point penalty term. The objective is to find a path with fewer inflection points, making it more practical for production. Simulation experiments have shown that the IA-RRT* algorithm outperforms several other path planning algorithms in terms of path length cost, algorithm calculation time, and number of inflection points. The paths generated by the IA-RRT* algorithm are applicable to the working situation of AGV and have practical significance.