Abstract:

In view of the problem of low positioning accuracy and poor robustness of the lidar SLAM(Simultaneous Localization and Mapping), this study proposes a SLAM method that combines IMU(Inertial Measurement Unit) data with the three-dimensional point cloud registration process. Based on the research of LeGO-LOAM(Lightweight and Ground-Optimized Lidar Odometry and Mapping on Variable Terrain), IMU data is introduced in the ground point extraction link, and the point cloud is mapped to the world coordinate system to reduce the influence of carrier jitter on the ground point extraction. The output information of IMU is used to eliminate the distortion of the point cloud due to the movement of the carrier and enhance the robustness of the algorithm. The three-point clustering method is used to perform cluster analysis on a frame of point cloud, which reduces the interference of noise, speeds up the point cloud registration process and improves the positioning accuracy of the algorithm. Meanwhile, closed-loop detection is introduced to reduce the cumulative error in the matching process and obtain the global optimal solution. The results show that in a large-scale outdoor interference environment, the improved SLAM algorithm reduces the trajectory fluctuations obtained by the solution, improves the point cloud registration accuracy, and enhances the robustness of the algorithm.

Key words: multiple information fusion, laser, IMU, three-point clustering, LeGO-LOAM, trajectory solution, laser SLAM