面向移动机器人SLAM的双RGB-D传感器数据融合方法与应用
于宁波; 徐昌; 王石荣
2017
会议名称36th Chinese Control Conference
会议日期July 26-28, 2017
会议地点Dalian, China
关键词双RGB-D 移动机器人 SLAM 数据融合 点云地图构建 全景图
其他题名Fusion of Dual RGB-D Sensors for Mobile Robot SLAM
页码6863-6868
通讯作者王石荣
中文摘要移动机器人在执行给定的任务时,需要通过传感器获取环境信息,从而构建地图、定位并在环境中自主运动。RGB-D传感器能够以较低的成本提供环境的色彩与深度信息,得到了广泛的应用,但是其存在视野受限的问题。为此,本文提出了一种面向移动机器人SLAM的双RGB-D传感器数据融合方法,对双RGB-D传感器的数据流进行同步处理,实时构建环境点云地图,并融合多视角RGB信息在线构建全景图。然后,将这一方法应用于室内环境下移动机器人SLAM与运动控制任务。实验表明,本文的双RGB-D传感器数据融合方法能够较好地构建环境点云地图,移动机器人能够根据地图信息自主地从起始点运动至目标点并规避障碍物,构建的全景图能够提供具有视觉临场感的环境信息。由此,为RGB-D传感器和移动机器人的拓展应用提供了一种新方法。
英文摘要Mobile robots need sensors to perceive environment information for map construction, self-localization and motion control. RGB-D sensors provide an economic solution to acquire color and depth information of the environment, and have received wide interest in recent years. Nevertheless, the small angle of view restricts its application. In this paper, we propose a method to fuse the data from two RGB-D sensors, so that the angle of view can be extended. Data streams from the two RGB-D sensors are synchronized, the point cloud map is constructed in real-time, and the environment panorama is built up online with the fused multi-view RGB information. The proposed method has been validated through SLAM and motion control tasks of the mobile robot in indoor environment. Experimental results showed preferable capabilities of point cloud map construction and self-localization, good autonomous motion control and successful obstacle avoidance. Meanwhile, the built panorama provided detailed environment information with visual telepresence. Therefore, the proposed method promises effective potentials to promote further applications of RGB-D sensors and mobile robots.
产权排序3
会议主办者中国自动化学会控制理论专业委员会
会议录Proceedings of the 36th Chinese Control Conference
语种中文
内容类型会议论文
源URL[http://ir.sia.cn/handle/173321/21070]  
专题沈阳自动化研究所_机器人学研究室
作者单位1.南开大学机器人与信息自动化研究所
2.机器人学国家重点实验室(中国科学院沈阳自动化研究所)
3.南开大学天津市智能机器人技术重点实验室
推荐引用方式
GB/T 7714
于宁波,徐昌,王石荣. 面向移动机器人SLAM的双RGB-D传感器数据融合方法与应用[C]. 见:36th Chinese Control Conference. Dalian, China. July 26-28, 2017.
个性服务
查看访问统计
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。


©版权所有 ©2017 CSpace - Powered by CSpace