Mobile Robot Path planning based on RRT-Reward Algorithm
Jun Yan, Zhen Huang, Ziqi An · 2024
Aiming at the shortcomings of high randomness and low efficiency of the RRT algorithm, this paper proposes an improved RRT algorithm (RRT -Reward Algorithm). Firstly, the map is initialized, filled with concave obstacles, and inflated for processing. Then, the sampling strategy is improved by sampling eight nodes in a circular area, and choosing the best new nodes by calculating the reward value. It can direct the random tree to quickly grow in the direction of the objective. When the robot falls into a local optimal solution, the obstacle expansion boundary sampling strategy is adopted to quickly bypass the obstacles. Finally, the path is pruned and smoothed. This paper's algorithm has good performance through simulation experiments and data analysis comparison.