Improved Double-tree RRT* Algorithm for Efficient Path Planning of Mobile Robots
Liquan Jiang, Shuting Wang, Jie Meng, Xiaolong Zhang, Yuanlong Xie · 2020
With a modified form of the rapidly- exploring random tree (RRT), RRT* algorithm is an important and effective tool for sampling-based path planning. However, the partial extension and low efficiency of the traditional RRT* make it very difficult to satisfy specific constraints or real-time requirements of mobile robotic scenarios. Based on the double tree structure expansion, a double-tree RRT* (D-RRT*) algorithm is proposed in this paper with the ability to improve space collision detection and search the feasible connection area using constrained nodes. The proposed method can effectively utilize the fast preprocessing ability of a double-tree structure and reduce the implementation of iterations. Meanwhile, node filtering and regression are designed to reduce the node numbers and prevent over space searching. Through the validation examples of mobile robots, it is shown that the proposed D-RRT* method can search for a global safe path with enhanced efficiency and convergence as compared with conventional methods.