基于改进RRT*算法的七自由度机械臂避障路径规划

Obstacle Avoidance Path Planning of 7-DOF Manipulator Based on Improved RRT* Algorithm

针对传统的快速扩展随机树(rapidly-exploring random tree,RRT*)算法应用于高维空间冗余机械臂路径规划时存在过多无用节点和无效节点而导致规划失败等问题,提出一种改进的节点控制快速扩展随机树(nodes-controlled RRT*,N-RRT*)算法。首先,提出随机树节点扩展控制策略,改善了随机树节点扩展的盲目性,同时减少了无用节点的生成;然后,针对随机树在扩展时可能产生无效节点的问题,提出随机树节点扩展能力检测策略,减少了无效节点的生成。最后,采用MATLAB软件进行了多障碍静态避障路径规划仿真实验,结果表明,与基本RRT*算法相比,N-RRT*算法可减少规划时间和障碍节点,并能提高规划成功率,为机械臂路径规划提供了一个新的研究思路。

In view of the problem that the traditional rapidly exploring random tree(RRT*)algorithm applied to path planning of the redundant manipulator leads to producing various useless nodes due to its inherent property and low success rate because of invalid nodes,the paper proposes a novel nodes controlled RRT*(N-RRT*)algorithm.Firstly,when it comes to the random exploring of tree nodes,a method of nodes constrained extension is proposed,which reduces useless nodes.For the issue that the algorithm generates invalid nodes,the extension ability detection mechanism is brought up.Finally,the simulation experiment of path planning in multiple obstacles is carried out with MATLAB software.Compared with traditional RRT*algorithm,the N-RRT*algorithm can shorten the time consumption and the useless nodes,and increased the experiment success rate,which provides a novel idea for the path planning of manipulator.