为解决移动机器人在环境未知条件下,利用单一传感器自主导航时不能及时定位、构建地图不精确的问题,提出采用一种改进RBPF算法,在计算提议分布时将移动机器人的观测数据(视觉信息与激光雷达信息)和里程计信息融合;针对一般视觉图像特征点提取算法较慢的问题,采用基于ORB算法对视觉图像进行处理以加快视觉图像处理速度的方法;最后通过在安装有开源机器人操作系统(ROS)的履带式移动机器人进行实验,验证了采用该方法可构建可靠性更高、更精确的2D栅格图,提高了移动机器人SLAM的鲁棒性.%Given an unknown environment,a mobile robot cannot achieve a timely autonomous navigation and an ac-curate map with a single sensor. To handle this issue,a modified RBPF algorithm is proposed,which calculates the proposed distribution based on the fusion of the observation data(visual Information or lidar Information)and odom-eter information. Since it is slow to extract information in general visual images,a visual image processing method is further proposed based on ORB algorithm to speed up the processing of visual images. Finally,experiments are con-ducted via crawler mobile robots embedded with open source robotic operating system(ROS)and the results show that more reliable and accurate 2D raster maps could be built by the proposed method,which improves the robust-ness of SLAM for mobile robots.
展开▼
