| 摘 要: 针对应用在机械臂的传统快速扩展随机树(RRT)算法搜索目的性不强、收敛慢和路径代价大等缺点,提出了一种自适应步长和终点导向偏置的改进RRT算法。首先,改进RRT算法依据扩展点一定范围内的环境复杂度来决定扩展步长;然后,以终点采样时,将终点以离开障碍物的最近距离方向进行偏置,经过两次赘余点移除,得到更短路径;最后,用三次B样条插值法对路径平滑处理。实验表明在三维空间中改进后算法与传统算法相比,路径代价减少了18.4%,运行时间减少了26.6%,迭代次数降低了92.1%。改进算法也成功运用到机械臂上。 |
| 关键词: 路径规划 RRT算法 自适应步长 终点导向偏置 |
|
中图分类号:
文献标识码: A
|
| 基金项目: 国家自然科学基金项目(61663005) |
|
| Path Planning for Robotic Arm Based on Improved RRT Algorithm |
|
ZHOU Gang, LI Handong
|
(College of Electrical Engineering, Guizhou University, Guiyang 550025, China)
1101808591@qq.com; 470394668@qq.com
|
| Abstract: To address the shortcomings of the traditional Rapidly-exploring Random Tree (RRT) algorithm applied to robotic arms—such as weak search purposefulness, slow convergence, and high path cost—an improved RRT algorithm incorporating adaptive step size and goa-l oriented bias is proposed. First, the step size for expansion is determined based on the environmental complexity within a certain range around the expansion point. Second, when sampling toward the goal, the goal position is biased along the direction of the closest distance away from obstacles.After removing redundant points twice, a shorter path is obtained. Finally, the path is smoothed using cubic B-spline interpolation. Experimental results demonstrate that, compared to the traditional algorithm in 3D space, the improved algorithm reduces path cost by 18.4% , runtime by 26.6% , and iteration count by 92.1% . The improved algorithm has also been successfully applied to a robotic arm. |
| Keywords: path planning RRT algorithm adaptive step size goa-l oriented bias |
