Path Planning for Soccer Robots Based on Improved Informed-RRT* Algorithm

In the actual soccer robot game, a good path planning result is the premise for the robot to carry out strategic movement. To solve the problems of the traditional Informed-RRT
^*
algorithm, such as long initial path generation time and poor path smoothness, a path planning method based on the improved Informed-RRT
^*
algorithm is proposed to make it suitable for soccer robot walking. First, we propose an initial path sampling based on rectangular sampling range, and then we use multi-threaded double tree growth instead of traditional alternate double tree growth. Secondly, a node expansion method based on simulating multiple resultant forces is proposed, and an adaptive step strategy is introduced. Finally, a method based on path point stretching and sharp angle bluntness is proposed to process path points, and a path with continuous curvature, good smoothness and short path length is obtained. And a series of experiments show that the algorithm in this paper has good efficiency and adaptability for soccer robot.