融合RRT*与APF算法的机器人路径规划研究
Robot path planning based on the fusion of RRT*and APF
杨勇 1成英1
作者信息
- 1. 天津职业技术师范大学汽车与交通学院,天津 300222
- 折叠
摘要
针对传统快速扩展随机树(rapidly-exploring random tree star,RRT*)算法在全局路径规划过程中存在收敛速度慢、搜索路径不平滑、内存占用多等问题,提出了一种RRT*与人工势场法(artificial potential field,APF)的融合搜索算法.为了加快RRT*算法在搜索过程中的收敛速度,在算法中利用人工势场法的思想引导扩展随机树快速向目标点生长;对融合算法在空间中的采样范围做出改进,使算法在APF产生的合力特定范围内进行采样,提高算法在空间中的搜索效率,减少无用节点的扩展.仿真结果表明:相比传统的RRT和RRT*算法以及APF-RRT融合算法,APF-RRT*融合算法能够规划出更短、更平滑的路径,路径距离缩短了 1.5%～10.83%;算法的搜索时间也显著缩短了 1.97%～49.78%;与其他算法相比,APF-RRT*融合算法的路径节点数量减少了 4.66%～41.95%,路径平滑性也得到了提高.
Abstract
To address the problems of slow convergence speed,unsmooth search paths and high memory usage in the global path planning process of the traditional rapidly-exploring random tree star algorithm(RRT*),a fusion search algorithm that integrates RRT*with the Artificial Potential Field(APF)method was proposed.To speed up the convergence of the RRT*algorithm in the search process,the principle of APF was used to guide the extended random tree to grow rapidly toward the target point.The sampling range of the fusion algorithm in space was improved,so that the algorithm can sample within a specific range of the net force generated by APF,improve the search efficiency of the algorithm in space,and reduce the expansion of useless nodes.The simulation results show that the APF-RRT *fusion algorithm outperforms traditional RRT and RRT*algorithms,as well as the APF-RRT fusion algorithm,by generating shorter and smoother paths,with path dis-tances reduced by 1.5%to 10.83%.Moreover,the search time of the algorithm is also significantly shortened by 1.97%to 49.78%.In addition,compared with the other algorithms,the number of path nodes of the APF-RRT *fusion algorithm is reduced by 4.66%to 41.95%,and the path smoothness is also improved.
关键词
融合算法/快速搜索随机树算法/人工势场法/全局路径规划/移动机器人Key words
fusion algorithms/rapidly-exploring random tree algorithm/artificial potential field method/global path plan-ning/mobile robots引用本文复制引用
基金项目
天津市教委科研计划项目(2020KJ122)
出版年
2024
NETL
NSTL
万方数据