A 3D UAV trajectory planning method based on Improved Convex Corner A-star algorithm

In recent years, with the increasingly complex application environment of the unmanned aerial vehicle (UAV), how to solve the problem of fast and efficient trajectory planning for the UAV has attracted much attention. This article proposes a 3D UAV trajectory planning method based on the improved convex corner A-star algorithm to address the problems of low computational efficiency, multiple turning nodes, and the possibility of collision with obstacles in traditional A-star algorithm. Firstly, the concepts of convex corners and adjacency relation in 3D space are proposed, and the corresponding algorithms for searching and judging are established. Secondly, a 3D convex corner A-star algorithm based on the convex corners and adjacency relation is proposed. After generating the flight trajectory of the UAV, a Cubic B-spline trajectory smoothing method integrating multiple constraint conditions is applied. Finally, several experiments on a self built map containing dense cylindrical obstacles are conducted. The result shows that the convex corner A-star algorithm has significantly improved the computational efficiency, substantially reduced the number of turning points, and successfully solved the problem of traditional A-star algorithm unable to obtain effective paths in large maps. Moreover, the smoothed trajectory can meet the constraints of the shortest path for UAV flight, the safe distance for obstacles, and the maximum turning radius of UAV.