Research on Mapping Algorithms for Underground Coal Mine Mobile Robots Based on Lidar Inertial Fusion

The coal mine underground space is narrow and cramped, the lighting conditions are uneven, so it poses challenges to the mapping of mobile robots in coal mines. In response to the above problems, a modified algorithm SC-LIO-SAM (the fusion of Scan-Context and LIO-SAM) is proposed, which realizes the mapping of robots in underground coal mines. Firstly, the Lidar odometry is used to eliminate the accumulated error of the IMU (Inertial Measurement Unit) and the point cloud distortion is removed through IMU pre-integration, forming a tightly coupled mapping system between the Lidar and IMU. Secondly, by adding the Scan-Context loop closure detection factor, Lidar odometry factor and IMU pre-integration factor for back-end graph optimization, the problem of low Lidar scan matching can be solved, thereby enhancing the global consistency of mapping and reducing the pose estimation error. Finally, taking the developed underground coal mine mobile robot as a platform, the algorithm proposed in this paper was verified by using the KITTI data set and the self-collected data set of simulated coal mine roadways. The results of the comparative experiments show that, the RMSE (Root Mean Square Error) of SC-LIO-SAM is reduced by 15.5%, 47.0% and 29.0%, respectively, compared with ALOAM, LeGO-LOAM and LIO-SAM.