摘要
针对双向快速扩展随机树(RRT-Connect)算法在焊接机器人避障路径规划过程中易于陷入工件所形成的凹形障碍区域,造成算法性能下降的问题,提出了一种改进的RRT-Connect算法。改进的算法对随机树上每一个新产生的树节点进行检测,对位于凹形障碍区域的树节点及其附近区域进行标记,并将标记的区域从自由状态空间中剔除,避免随机树再次陷入标记的凹形障碍区域。仿真实验表明,改进的RRT-Connect算法在搜索次数上降低了47.30%,路径长度下降了9.96%,路径平滑度指标也有明显改善。
The RRT-Connect algorithm is easy to fall into the concave obstacle area formed by the workpiece during the path planning process of the welding robot,which causes declination of its performance.Therefore,we propose an improved RRT-Connect algorithm.The improved algorithm detects each new vertex and marks the vertex and its vicinity located in the concave obstacle area.To prevent the random tree from falling into the marked concave obstacle area again,the improved algorithm removes the marked area from the obstacle-free state space.Simulation results show that the improved RRT-Connect algorithm reduces the number of iterations by 47.30%,the path length decreases by 9.96%,and the path smoothness is also significantly improved.
作者
孙灵硕
SUN Ling-shuo(School of Logistics Engineering,Wuhan University of Technology,Wuhan 430000,China)
出处
《自动化与仪表》
2021年第5期47-52,共6页
Automation & Instrumentation
基金
国家自然科学基金项目(51975444)。
