一种基于视觉补偿的移动机器人地图匹配室外定位算法（英文）  

An outdoor localization method based on visual compensation for mobile robot map-matching

作　　者：李修云[1,2] 梁山[2] 赵聃敏[2] 

机构地区：[1]重庆工程职业技术学院,重庆402260 [2]重庆大学自动化学院,重庆400044

出　　处：《机床与液压》2016年第12期122-128,135,共8页Machine Tool & Hydraulics

基　　金：supported by Chongqing Municipal Education Committee Science and Technology Project (KJ1403209) Support Program

摘　　要：在室外环境中,由于车轮打滑、里程计存在严重的累积误差,同时由于二维激光测距仪扫描到环境信息量少,在进行地图匹配时,机器人常常陷入相似或对称环境,导致定位失败。针对以上问题,在二维激光地图匹配的基础上,提出了利用单目视觉信息补偿激光信息与里程计信息的定位方法。该方法利用单目摄像机获取环境中的色块信息补偿二维激光信息;摄像机同时作为视觉里程计实时辨识车轮打滑,校正里程计的误差;最后利用粒子滤波算法融合多传感器信息。该实验利用Pioneer3-AT机器人平台在大学校园进行,实验证明该方法具有较高的定位精度。Due to the wheel slip phenomenon of the mobile robot in the outdoor environment,the odometer has terrible accumulation errors.The two-dimensional laser range finder only scans little environment information,of which the robot would be located in a similar or symmetry environment during the map-matching process.It leads to the localization failure.According to the above problems,this paper proposed a novel localization method based on the two-dimensional laser map-matching method by using the monocular visual to compensate laser data and odometer data.This method utilizes the monocular camera to collect the color lump data to compensate the two-dimensional laser data.The camera is also used as the visual odometer to observe the wheel slip phenomenon,by which the odometer error can be corrected.Finally,particle filter algorithm is applied to fuse the multi-sensor information.Based on the Pioneer 3-AT mobile robot,experiment is proposed in a campus of university.The results show that the proposed localization method has good localization accuracy.

关 键 词：二维激光地图匹配 视觉补偿 粒子滤波 定位 

分 类 号：TP242.62[自动化与计算机技术—检测技术与自动化装置]