An Improved Dynamic Step Size RRT Algorithm in Complex Environments
Yuwei Zhang, Ruirong Wang, Chunlei Song, Jianhua Xu · 2021
Rapidly exploring Random Tree(RRT) is an efficient path planning algorithm based on random sampling, which plays an important role in the robot field and autonomous driving field. However, due to the randomness of sampling, its results are usually not optimal. This paper proposes a dynamic step size RRT algorithm, which mainly improves the traditional RRT as follows. First, combined with the Artificial Potential Field(APF), the target makes heuristic guidance for the sampling process. And then, the step size is adaptively changed according to the density of obstacles. After that, a one-shot heuristic strategy is used to speed up the search process. Finally, a bi-directional pruning strategy is adopted to reduce the path length by merging points. The simulation results show that the improved RRT algorithm can find the target faster and better.