Scan matching using line segment relationships for map-based localization of mobile robot  

作　　者：葛坤峰 陈卫东 王景川 冉宇瑶 

机构地区：[1]Dept. of Automation, Shanghai Jiaotong University [2]Dept. of Computer Science and Technology, Zhejiang Sci-Tech University

出　　处：《Journal of Harbin Institute of Technology(New Series)》2008年第6期757-762,共6页哈尔滨工业大学学报（英文版）

基　　金：Sponsored by the National High Technology Research and Development Program of China(Grant No.2006AA040203);The National Natural Science Foundation of China(Grant No.60475032 and 60775062);the Program for New Century Excellent Talents in University(Grant No.NCET-07-0538)

摘　　要：A new scan matching method for mobile robot localization is presented, which takes line segment as the feature and matches the real scans in the given reference map by relationships of the directional-defined line segments. The alignment was done by hierarchically identifying the multiple relationships and the result was recorded in a correspondence matrix, where the best match is defined and selected for localization. It is indicated that the searching algorithm of the best match can find the ambiguities and get rid of them. This method with less computational cost works well in occluded environment, and can correct the error in pose estimation without the need for the estimation itself. The efficiency, accuracy and robustness of this method were verified by experiments of localization in an occluded environment and a long-distance indoor navigation.A new scan matching method for mobile robot localization is presented, which takes line segment as the feature and matches the real scans in the given reference map by relationships of the directional-defined line segments. The alignment was done by hierarchically identifying the multiple relationships and the result was recorded in a correspondence matrix, where the best match is defined and selected for localization. It is indicated that the searching algorithm of the best match can find the ambiguities and get rid of them. This method with less computational cost works well in occluded environment, and can correct the error in pose estimation without the need for the estimation itself. The efficiency, accuracy and robustness of this method were verified by experiments of localization in an occluded environment and a long-distance indoor navigation.

关 键 词：scan matching laser range finder map-based localization mobile robot 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]