• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2025, Vol. 61 ›› Issue (5): 165-177.doi: 10.3901/JME.2025.05.165

• 机器人及机构学 • 上一篇    

扫码分享

基于Voronoi骨架的移动机器人融合路径规划

武星1, 李杨志1, 臧铁钢1, 孟昭旭2, 陈俊哲1, 王晨涛1   

  1. 1. 南京航空航天大学机电学院 南京 210016;
    2. 中国兵器工业第二〇八研究所 北京 102202
  • 收稿日期:2024-03-28 修回日期:2024-07-05 发布日期:2025-04-15
  • 作者简介:武星(通信作者),男,1982年出生,博士,研究员,硕士研究生导师。主要研究方向为移动机器人感知定位与导航控制。E-mail:wustar5353@nuaa.edu.cn;李杨志,男,2000年出生,硕士研究生。主要研究方向为移动机器人路径规划与控制。E-mail:Liyz13193439079@163.com
  • 基金资助:
    国防基础科研计划重点项目(JCKY2022209B001)和江苏高校“青蓝工程”优秀青年骨干教师项目(2022)资助项目。

Combined Path Planning Based on Voronoi Skeleton for Mobile Robots

WU Xing1, LI Yangzhi1, ZANG Tiegang1, MENG Zhaoxu2, CHEN Junzhe1, WANG Chentao1   

  1. 1. College of Mechanical & Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016;
    2. No. 208 Research Institute of China Ordnance Industries, Beijing 102202
  • Received:2024-03-28 Revised:2024-07-05 Published:2025-04-15

摘要: 针对栅格地图下移动机器人路径规划存在路径拐点多、运行不平滑、安全性不高等问题,提出一种基于Voronoi骨架的融合路径规划方法。首先采用Delaunay三角剖分、对偶Voronoi图构造及无向图邻接矩阵变换等方法将栅格地图变换为Voronoi骨架图。其次基于骨架顶点使用A*算法进行全局路径规划,并根据空间避障需求进行路径点优化,提高全局路径的安全性与精炼性。再次优化DWA算法速度采样空间,通过引入全局优化引导改进评价函数,构建全局引导型DWA算法,并采用动态目标点将局部路径规划与全局规划路径进行融合。实验结果表明,该融合路径规划方法具有导航安全性、路径平滑性和运行高效性,可使移动机器人快速规避随机动-静态障碍物。

关键词: Voronoi骨架图, A*算法, DWA算法, 路径规划, 移动机器人

Abstract: In response to these problems of excessive turning points, non-smooth motion, and insufficient safety for path planning of mobile robots on grid maps, a combined path planning method based on the Voronoi skeleton is proposed. Firstly, the grid map is transformed into a Voronoi skeleton graph by means of three methods, including Delaunay triangulation, dual Voronoi diagram construction, and adjacency matrix transformation of undirected graph. Secondly, global path planning is conducted using the A* algorithm based on skeleton vertices, followed by optimization of path points according to spatial obstacle avoidance requirements, thus enhancing the safety and refinement of the global path. Furthermore, the velocity sampling space of the DWA algorithm is optimized. Global optimization guidance is introduced through an improved evaluation function to construct a globally guided DWA algorithm. Then local path planning is combined with global path planning by using dynamic target points. Experimental results demonstrate that this combined path planning method possesses navigation safety, path smoothness, and operational efficiency, enabling mobile robot to avoid random dynamic-static obstacles rapidly.

Key words: Voronoi skeleton, A-star algorithm, DWA algorithm, path planning, mobile robot

中图分类号: