基于改进Lego Loam算法的移动机器人路径规划及导航研究  

Path Planning and Navigation for Mobile Robots Based on Improved Lego Loam Algorithm

作　　者：凌新宇 张立臣 黄洪斌 LING Xin-yu;ZHANG Li-chen;HUANG Hong-bin(School of Electrical and Electronic Engineering,Anhui Institute of Information Technology,Wuhu 241000,China)

机构地区：[1]安徽信息工程学院电气与电子工程学院,安徽芜湖241000

出　　处：《榆林学院学报》2025年第2期101-105,共5页Journal of Yulin University

基　　金：安徽省高等学校省级自然科学研究计划项目(2023AH052918)。

摘　　要：为了解决移动机器人在三维环境中容易碰到障碍物以及Lego Loam算法在提取地面时的杂点问题。本文提出将最近邻插值法与体素滤波相结合的SLAM系统,在静态环境下进行仿真实验,在不丢失点云特征的情况下对地面杂点进行滤波。与原Lego Loam算法相比,地面点云的发布时间缩短了16.89%,经过omp并行处理后总点云提取及匹配时间降低2.9%。将所设计的算法应用于实际机器人平台进行验证,根据获得的点云图完成重定位,并同步映射二维栅格图进行导航。在动态环境下,本文将传统的二维定位算法与三维定位算法进行了对比,该平台下三维点云定位的可行性优于传统的二维定位算法。在完成定位和导航配置后,机器人最终可以有效地到达目标点。In order to solve the problem of mobile robots easily colliding with obstacles in 3D environments,as well as the stray point problem in the Lego Loam a lgorithm when extracting the ground,the SLAM system,which integrates the nearest neighbor interpolation method and voxel filtering,is proposed to conduct simulation experiments in a static environment,so as to filter ground clutter without losing point cloud features.Compared with the original Lego Loam algorithm,the release time of ground point cloud is shortened by 16.89%.After omp parallel processing,the total point cloud ex traction and matching time is reduced by 2.9%.The designed algorithm is applied to the actual robot platform for verification,and the relocation is completed according to the obtained point cloud image,and a two-dimensional raster map is synchronously mapped for navigation.In the dynamic environment,the traditional two-dimensional positioning algorithm is compared with the three-dimensional positioning algorithm in this paper,and the feasibility of three-dimensional point cloud positioning under this platform is superior to the traditional two-dimensional positioning algorithm.After completing the positioning and navigation configuration,the robot can finally reach the target point effectively.

关 键 词：机器人 Lego Loam算法 点云 导航 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]