基于改进双树RRT*算法的冗余机械臂末端路径规划

针对冗余机械臂的冗余特性与相关RRT*算法在规划机械臂末端路径的应用中存在的搜索效率较低、收敛性不稳定以及没有充分考虑到机械臂末端几何构型与自身运动特性对路径规划影响的问题，提出一种改进策略。首先，引入一种基于根尾节点连线夹角的采样点选择方式，并设置目标逼近区域。根据连续采样成功次数动态选择改进采样与随机采样。接着，将双树扩展策略与上述方法相结合。最后，将初始可行路径进行二次重连得到最终的优化路径。通过验证，改进双树RRT*方法能够有效地提升搜索效率、收敛性以及路径的优越性。虚拟碰撞体与胶囊碰撞体的引入也能较好地应对机械臂末端结构与运动特性带来的影响。使用Mujoco物理仿真引擎进行机械臂运动验证，证明该策略可以为冗余机械臂末端规划出一条较优的可行路径。

Redundant manipulator end-effector path planning based on improved bidirectional RRT* algorithm

Aiming at the redundancy characteristics of the redundant manipulator and the relevant RRT* algorithm had issues with low search efficiency, unstable convergence and not fully considering the influence of the geometric configuration and self-motion characteristics of the redundant manipulator''s end effector on path planning in applications for planning the end-effector path, this paper proposed an improved strategy. Firstly, this paper introduced a sampling point selection method based on the angle between the root and end node connection, along with the establishment of a target proximity region. This method dynamically switched between improved sampling methods and random sampling methods based on the number of successful continuous samples. Subsequently, this paper combined the bidirectional tree expansion strategy with the aforementioned method. Finally, it reconnected the initial feasible path twice to obtain the optimized path. Through verification, it improved that the Bidirectional Tree RRT* method could effectively enhance search efficiency, convergence, and path optimality. The introduction of virtual collision bodies and capsule collision bodies could also effectively respond to the impact of the structure and motion characteristics of the end effector of the redundant manipulator. This paper used the Mujoco physics simulation engine to validate the movement of the redundant manipulator. This paper demonstrates that this strategy can plan a more optimal feasible end-effector path for the redundant manipulator.