一种基于RGB-D SLAM的移动机器人定位和运动规划与控制方法

LOCALIZATION AND MOTION PLANNING AND CONTROL OF MOBILE ROBOTS WITH RGB-D SLAM

移动机器人通常需要搭载多种传感设备以获取自身和周边环境的信息,导致系统过于复杂,且成本高昂.文章建立了一种低成本移动机器人系统,使用基础运动底盘,搭载单一的通用RGB-D传感器,在ROS开源环境下,运行RGB-D SLAM算法、点云分割算法和反馈镇定控制器,实现了移动机器人的自身定位、障碍检测、运动规划与控制,并能同时建立三维空间地图.实验表明,文章系统具有出色的定位与运动控制效果,成功检测并规避障碍物,建图效果良好,并且精度较高,重复性好.此外,系统结构简便,通用性好,可在家庭服务等环境和任务方面取得广泛应用.

Mobile robots need various sensors to acquire necessary information of the environment, resulting in a complicated and expensive system. This paper pro- posed a low-cost system, consisting of one mobile platform, one RGB-D sensor, and one control computer. The RGB-D SLAM algorithm for localization and maDDing.point cloud processing for obstacle detection, together with the feedback stabilization controller for motion planning and control are developed and implemented in the ROS platform. Experimental results showed good motion control accuracy and repeatability, successful obstacle avoidance and 3D map construction capability. This promises feasibility for service applications in home and office conditions.