Poles Extraction Based LiDAR Localization for Autonomous Robot Systems  

作　　者：HONG Xin DONG Yi 

机构地区：[1]College of Electronic and Information Engineering,Tongji University,Shanghai 201800,China

出　　处：《Journal of Systems Science & Complexity》2024年第4期1413-1424,共12页系统科学与复杂性学报（英文版）

基　　金：supported by Shanghai Rising-Star Program under Grant No.22QA1409400;the National Natural Science Foundation of China under Grant Nos.62073241 and 62173250;Shanghai Municipal Science and Technology Major Project under Grant No.2021SHZDZX0100。

摘　　要：With the development of intelligent technology,the problem of Li DAR-based localization has played an increasingly important role in the localization of robots in unmanned systems.Since the dense point clouds from the environment requires a large amount of memory,specific and stable structural features,such as pole-like objects in the environment,can serve as the ideal landmarks for localization in unmanned systems,which can effectively reduce the memory usage and mitigate the effects of dynamic environment changes.In this paper,the authors propose a pole-like objects extraction approach and then,it is applied into the localization system of mobile robots.Under the same experimental conditions,the proposed method can extract more pole-like objects and achieve better long-term localization accuracy than some existing methods in different urban environmental datasets.

关 键 词：Pole-like landmark extraction robot localization 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置] TN958.98[自动化与计算机技术—控制科学与工程]