An improved scan matching algorithm in SLAM

Simultaneous localization and mapping (SLAM) technology has always been the research focus of robot navigation in unknown environment. Aiming at the problem of cumulative errors of robot pose in the localization process of SLAM algorithm based on particle filter, a loop detection algorithm based on graph-SLAM was proposed. The algorithm uses constraints to adjust the robot attitude at different moments. In this paper, the constraint refers to the scanning matching of lidar. In the process of drawing, when the robot returns to the known area, if the current laser scanning is successfully matched with the previous laser scanning, the robot's posture can be adjusted to eliminate the accumulated errors caused by the odometer. In the process of laser scanning matching, the method of grouping step threshold value judgment is proposed to match the laser point cloud, which can effectively reduce the computation. Experimental results show that the proposed algorithm can effectively eliminate the cumulative errors of positioning and achieve a better mapping effect.