Kinematic Constrained Bi-directional RRT with Efficient Branch Pruning for robot path planning

• Published in: Expert systems with applications 2021-05, Vol.170, p.114541, Article 114541
• Main Authors: Wang, Jiankun, Li, Baopu, Meng, Max Q.-H.
• Format: Article
• Language: English
• Subjects: Algorithms; Bi-directional search; Constraints; Kinematics; Path planning; Reagents; Robot; Robots; RRT
• ISSN: 0957-4174; 1873-6793
• DOI: 10.1016/j.eswa.2020.114541

•A novel Bi-directional RRT path planning scheme.•A local path planner taking into account the kinematic constraints.•An efficient branch pruning strategy.•The proposed KB-RRT* can achieve better performance in path planning. In this paper, we present a novel path planning algorithm called Kinematic Constrained Bi-directional Rapidly-exploring Random Tree with Efficient Branch Pruning (KB-RRT*), which is designed on the basis of Bi-directional Rapidly-exploring Random Tree (Bi-RRT) method. The KB-RRT* is suitable for path planning of a differential drive mobile robot. By incorporating the kinematic constraints, the KB-RRT* can avoid unnecessary growth of the tree and quickly find a feasible path for the agent. In addition, the proposed efficient branch pruning strategy helps the generated state find a better parent state and delete the edges with high cost around it. Therefore, the KB-RRT* achieves fast and high-quality path planning. The results of numerical simulations reveal that the proposed KB-RRT* can achieve better performance compared with the conventional path planning algorithms.