Three dimensional path planning of cross-domain robot based on improved A algorithm

When the robot moves autonomously in the environment, it needs to plan a collision free three-dimensional path from the starting point to the target point. The node search strategy of traditional A* algorithm has the problems of many feasible path turning points, large turning angle, redundant path points and so on, and it is mainly used in two-dimensional path planning. The improved A* algorithm integrates the environment information into the environment modelling by constructing a 2.5-dimensional environment grid map. It regards obstacles and water grids as passable grids and gives different costs. It designs a multi-objective heuristic function based on motion energy consumption and time. In order to avoid the redundant planning path points and the situation that is not conducive to the robot following, it adds part of the space in the backtracking process. At the same time, the redundant path points are reduced. The simulation results show that the improved A* algorithm can realize the three-dimensional path planning of the cross-domain robot. The length of the planned path is shorter and smoother, and the total number of path points is less.