面向受限工业环境的移动检测机器人全局路径规划方法  

Global Path Planning Method for Mobile Detection Robots in Constrained Industrial Environments

作　　者：吴应 武越 张凯 孙晓界 陈思远 WU Ying;WU Yue;ZHANG Kai;SUN Xiaojie;CHEN Siyuan(Southwest Jiaotong University,Chengdu 610031,China;Sichuan University,Chengdu 610065,China;Robotic Satllite Laboratory,Chengdu 610065,China;Henan University of Science and Technology,Luoyang 471023,China)

机构地区：[1]西南交通大学电气工程学院,成都610031 [2]四川大学空天科学与工程学院,成都610065 [3]机器人卫星实验室,成都610065 [4]河南科技大学信息工程学院,河南洛阳471023

出　　处：《沈阳理工大学学报》2025年第3期47-52,59,共7页Journal of Shenyang Ligong University

基　　金：国家自然科学基金项目(62003069);机器人卫星实验室开放项目;中国博士后科学基金第74批面上资助项目(2023M741038)。

摘　　要：针对当前全局路径规划方法在受限工业环境中所规划的全局路径存在安全性不足、曲率不连续等问题,提出一种融合改进Dijkstra算法和二次规划的移动检测机器人全局路径规划方法。首先改进Dijkstra算法,在原始评价函数的基础上增加障碍物代价项,使规划路径远离障碍物,提出自适应评价函数,平衡路径长度与路径安全性,使用改进的Dijkstra算法生成从起始位置到目标位置且远离障碍物的引导路径;然后改进二次规划方法,对引导路径进行平滑处理,改进约束条件以保证平滑后路径的安全。仿真实验结果验证了本文算法的有效性:与ROS内置的Dijkstra算法及A*算法相比,本文算法能够提升全局路径的平滑性与安全性;与其他融合算法相比,本文算法能够在不降低全局路径平滑性的同时提升其安全性。To address the issues of insecurity and curvature discontinuity of the current global path planning in restricted industrial environments,a global path planning method for mobile detection robots on the basis of improved Dijkstra algorithm and quadratic programming is proposed.Firstly,an improved Dijkstra algorithm is proposed,which adds an obstacle cost term to the original evaluation function to make the planned path avoid obstacles,and introduces an adaptive evaluation function to keep a balance between the path length and safety.This improved Dijkstra algorithm results in a guiding path from the starting position to the target position while keeping a safe distance from obstacles.Then,an improved quadratic programming method is used to smooth the guiding path,with enhanced constraints to ensure the safety of the smoothed path.Simulation results demonstrate the effectiveness of the proposed method.Compared with the built-in Dijkstra and A*algorithms in ROS,the proposed algorithm improves both the smoothness and safety of the global path.Additionally,compared with other fusion algorithms,the proposed method enhances the safety of the global path without compromising its smoothness.

关 键 词：受限工业环境 移动检测机器人 全局路径规划 DIJKSTRA算法 二次规划 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]