基于激光雷达与视觉信息融合的SLAM方法  

SLAM method based on lidar and visual information fusion

作　　者：王光庭 曹凯[1] 刘豪 WANG Guangting;CAO Kai;LIU Hao(School of Transportation and Vehicle Engineering,Shandong University of Technology,Zibo 255049,China)

机构地区：[1]山东理工大学交通与车辆工程学院,山东淄博255049

出　　处：《山东理工大学学报（自然科学版）》2019年第1期9-13,19,共6页Journal of Shandong University of Technology:Natural Science Edition

基　　金：国家自然科学基金项目(61573009)

摘　　要：单线激光雷达传感器的扫描范围是二维平面,在室内环境中,服务型移动机器人在进行同时定位和地图构建过程中,不在激光雷达扫描平面范围内的障碍物无法被识别,构建的地图中会缺少相应的环境信息。为了获取缺失的环境信息,通过改进粒子滤波Rao-Blackwellized的方法得到移动机器人位姿的估计,将Kinect获取的环境三维信息数据和激光雷达数据融合,使用载有ROS(机器人操作系统)的移动平台Turtlebot2在实际场景中进行实验。实验结果证明,使用融合后的数据信息能得到更加准确的建图和导航效果。When the Service Robots locates themselves in the indoor environment, the scanning range of single-line lidar sensor is two-dimensional plane. The obstacles that are not in the plane range of the lidar scan can not be identified, and thus a lot environmental information is missing in the constructed map. In order to get the missed environmental information, we estimated the pose of a mobile robot by the Rao- Blaekwellized particle filter, and analyzed the shortcomings of SLAM in this ease. Then, three-dimensional environmental information data from Kineet was added to the lidar data to get fusion map. Finally, we used the mobile platform Turtlebot2 of ROS （Robot Operation System） to do experiments in real scene, which proves that using fusion information can get better mapping and navigation effect.

关 键 词：同时定位和地图构建 数据融合 机器人操作系统 导航 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]