Improved A-star Method for Collision Avoidance and Path Smoothing

Abstract

The A-Star algorithm is one type of typical and useful method in the field of path planning and has been used in many fields. However, the conventional A-Star algorithm just takes the shortest path as the optimal path generation condition and does not consider the actual morphology of the robot. At the same time, the path search efficiency can be further improved. In this study, the problems with the A-Star algorithm above are improved. Firstly, changing the evaluation function of the traditional A-Star algorithm make it more effective. Then, the search directions are changed from 8 to 5. The orthodox A-Star algorithm has eight directions for exploration, which is not always efficient, so the search directions are adjusted to five to improve its efficiency. Finally, avoid oblique penetrations of the obstacle vertex to avoid the collision. The size of objects is usually neglected in the traditional A-Star algorithm. The route produced by the new method avoids moving too close to obstacles. The result of the experiment proves the effectiveness of the improved method. It improves the traditional A-Star algorithm so that the generated paths avoid sticking too close to the obstacle or even colliding with it.