Obstacle Avoidance Path Planning by Improved RRT Algorithm for Mobile Robot

This paper proposes an improved algorithm to address the efficiency issues of the RRT method in a 2D virtual environment in terms of spatiotemporal complexity. Firstly, an adaptive step size strategy is used to determine the appropriate global initial step size for a given map. Then, a target bias strategy is employed to assign a probability value to the sampling point as the target point, reducing the randomness of the sampling point. Then, incorporating the concept of target gravity improves the goal orientation of the algorithm, making the gravity coefficient adjustable. Finally, the pruning optimization strategy is used to remove redundant points in the initial path, and the pruned optimized path is smoothed. It has been verified that the proposed algorithm has advantages in planning time, path length, and number of path nodes, demonstrating the effectiveness of this method.