Abstract: Aiming at the problems of low planning efficiency, long exploration time and zigzag planning path of traditional RRT* and bidirectional RRT* planning algorithms in complex environment, a path planning algorithm of IBRRT* manipulator based on redundant node filtering mechanism is proposed. Firstly, based on the IBRRT* programming algorithm, a local node substitution mechanism is introduced to avoid the redundant expansion of nodes. When the initial path from the initial point to the target point is explored for the first time, the elliptic state subset sampling algorithm is introduced into the subsequent iterations to exert constraints on the sampling area and avoid the generation of redundant nodes. Finally, the redundant edge nodes which donot contribute to the path optimization are eliminated round by round. Two-dimensional simulation results show that the proposed algorithm has fast convergence speed and high efficiency. Compared with IBRRT* algorithm, the proposed algorithm can reduce the time by about 50% and the number of nodes by more than 75%. Moreover, the effectiveness and practicability of the proposed algorithm are proved by the manipulator path planning experiment.

Key words: path planning, improved bidirectional RRT*（IBRRT*）, initial path, local node substitution, elliptic state subset sampling algorithm

摘要： 针对传统的RRT*和双向RRT*规划算法在复杂环境存在下规划效率低、探索时间长、规划路径曲折等问题，提出了一种基于冗余节点过滤机制的IBRRT*机械臂路径规划算法。在IBRRT*规划算法基础上，引入局部节点替代机制避免节点的冗余拓展，当首次探索到起始点到目标点的初始路径，将椭圆状态子集采样算法引入后续的迭代中，对采样区域施加约束，避免冗余节点生成。最后针对迭代过程中对路径优化没有贡献的边缘冗余节点进行逐轮剔除。二维仿真实验结果表明，提出的算法收敛速度快效率高，相比IBRRT*算法在时间上减少50%左右，节点个数减少75%以上，并且通过机械臂路径规划实验证明了算法的有效性和实用性。

关键词: 路径规划, IBRRT*, 初始路径, 局部节点替代, 椭圆状态子集采样算法