摘要
文中针对传统RRT(Rapidly Exploring Random Tree)路径规划算法所存在的搜索随机性强、盲目性、路径冗余性及轨迹不连续性的问题,提出了一种改进RRT的运动规划算法,并对轨迹进行优化。首先,设计了去除搜索区域的动态采样区域的方法,通过去除已搜索区域,控制随机点的产生位置,降低搜索的随机性。其次,采用人工势场导向节点扩展策略,将障碍物与目标点的影响考虑到节点的扩展中,降低搜索的盲目性。然后,对初步路径进行优化,去除冗余部分、缩短路径距离。最后,使用最小snap的方法结合飞行走廊和时间重新分配,在保证不与障碍物碰撞的前提下使轨迹光滑连续,并将物理量控制在合理范围内。仿真试验结果表明:改进后的算法在搜索到路径时相较于未改进算法所扩展的节点平均减少了34.3%,对初步路径进行优化后路径长度平均缩短了25.8%。
In this paper,for the problems of high search randomness,blindness,path redundancy,and trajectory discontinuity that exist in the traditional RRT(Rapidly Exploring Random Tree)path planning algorithm,a motion planning algorithm to improve RRT is proposed and the trajectory is optimized.First,a method of removing the dynamic sampling region of the search region is designed to reduce the randomness of the search by removing the searched region and controlling the location of the random point generation.Then,the artificial potential field-oriented node expansion strategy is adopted to take the influence of obstacles and target points into account in the expansion of nodes,so that the expansion of nodes has the characteristics of being far away from obstacles and converging to the endpoint,and the blindness of search is reduced.Secondly,the preliminary path is optimized to remove redundant parts and shorten the path distance to ensure that the path maintains a safe distance from obstacles.Finally,the minimum-snap method combines the flight corridor and time redistribution to make the trajectory smooth and continuous and control the physical quantities within a reasonable range to ensure no collision with obstacles.The results of simulation experiments show that the improved algorithm searches for the path with an average reduction of 34.3%compared to the nodes expanded by the unimproved algorithm,and the path length is shortened by an average of 25.8%after the optimization of the preliminary path.
作者
王路平
滕超凡
WANG Luping;TENG Chaofan(Engineering Training Center,Shenyang Aerospace University,Shenyang 110136;School of Aeronautics and Astronautics,Shenyang Aerospace University,Shenyang 110136)
出处
《机械设计》
CSCD
北大核心
2024年第S02期47-53,共7页
Journal of Machine Design
基金
中国大学生创新创业项目(202310143002)
作者简介
王路平(1989—),女,高级实验师,硕士,主要研究方向:智能控制系统。E-mail:293529957@qq.com;通信作者:滕超凡(2003—),男,大学本科,主要研究方向:机器人路径规划。E-mail:2727626344@qq.com
