A FUSION ALGORITHM FOR PATH PLANNING OF MOBILE ROBOTS IN ENVIRONMENTS WITH DYNAMIC OBSTACLES

To find a smooth, safe global path that avoids the local dynamic obstacle, this article proposes a method of integrating the improved A* algorithm and artificial potential field method, namely, MAAPF. Firstly, the multi-objective functions are introduced into the heuristic function of the A* algorithm to reduce the redundant points in the global path. When the robot detects dynamic obstacles, it searches the global path node as the local goal according to the robot's position and detecting range, meanwhile combining the dynamic obstacle trajectory predicted by the autoregressive model and static obstacles in the detection range to construct the local map, then through the artificial potential field method that is improved by adding the goal guidance factor and gravitational distance threshold to complete local dynamic obstacle avoidance, avoid the goal is unattainable and locally optimal. The simulation demonstrates that improving the A* algorithm within a 3D environment and the artificial potential field algorithm has better results than other algorithms. Besides, the MAAPF can obtain a safe optimal path in circumstances with dynamic obstacles.