未知环境下基于激光雷达的启发式导航算法  

Heuristic navigation based on LiDAR in unknown environment

作　　者：常路 单梁[1] 张伟龙 王维西 戴跃伟[1,2] CHANG Lu;SHAN Liang;ZHANG Wei-long;WANG Wei-xi;DAI Yue-wei(School of Automation,Nanjing University of Science and Technology,Nanjing 210094,China;Nanjing Center For Applied Mathematics,Nanjing 211135,China)

机构地区：[1]南京理工大学自动化学院,南京210094 [2]南京应用数学中心,南京211135

出　　处：《控制与决策》2024年第9期2903-2912,共10页Control and Decision

基　　金：中国博士后科学基金项目(2021M701724)。

摘　　要：针对未知环境下机器人难以准确理解障碍分布并有效搜索目标的问题,提出一种基于激光雷达的启发式导航算法.在雷达数据中引入上升/下降沿进行障碍边沿界定及检测,提出边沿匹配算法识别狭窄间隙和连续障碍,经间隙去除与安全性拓展,得到表征周围环境中的连续障碍集合;面对遮挡目标的障碍,提出启发式临时目标选取方法并设计其切换条件,使机器人以较短路径提前避让各类障碍并渐近抵达全局目标.基于Matlab进行一系列仿真和对比,结果表明,所提出的算法可显著提高机器人在复杂未知环境下的通行效率、轨迹平滑度和全局搜索能力.最后基于ROS的实验验证所提出算法在实际环境中的有效性.In unknown environments,it is difficult for mobile robots to precisely understand the obstacle distribution and effectively search the target.This paper proposes a heuristic navigation method based on LiDAR.First,the rising/falling edges are introduced in LiDAR-data to define and detect obstacles’edges,and the edge matching algorithm is designed to recognize the narrow gaps and continuous obstacles.After the gap removal and obstacle expansion,the continuous obstacle set is obtained to describe the surrounding environment.Second,facing the obstacle that blocks the target,a heuristic selection method for temporary targets is proposed and the switching condition is designed,which make the robot avoid each kind of obstacle in advance with a shorter path,and reach the target progressively.Simulations and comparisons based on Matlab are conducted,whose results show that the proposed method greatly improves the traffic efficiency,path smoothness and global searching ability of mobile robots in unknown environments.The proposed method is also implemented in experiments based on ROS to verify its effectiveness in real environments.

关 键 词：移动机器人 未知环境导航 激光雷达 启发式搜索 环境理解 障碍分割与提取 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]