Research On Laser SLAM Algorithm For Complex Environment In Coal Mines

With the rapid advancement of smart mine construction,many new mine construction methods and positioning and navigation technologies have emerged.In the process of intelligent mining in coal mines,various autonomous mobile robots need to perform accurate positioning and navigation to ensure accurate control and management of mine less humanized mining by staff.The autonomous navigation of mobile robots in the mining environment relies on Simultaneous Localization and Mapping(SLAM).SLAM provides environmental structure information for the positioning and navigation of coal mining robots by scanning the 3D information of the mine and constructing point cloud maps.Point cloud map construction is one of the key technologies of SLAM.The coal industry is a high-risk industry with high intensity and danger factor of staff operation,and many positions are short of operators and need to adopt robot operation.In this paper,based on the independently designed multi-source information perception platform,we use the sensors such as LIDAR and IMU of the multi-source information perception platform to sense the environment around the mine,and realize the autonomous positioning of the multi-source information perception platform and the construction of the point cloud map;based on the created offline point cloud map,we use the normal distribution transformation algorithm to realize the repositioning of the multi-source information perception platform in the point cloud map of the mine.The research contents of this paper are as follows.(1)Design and implementation of a multi-source information sensing platformIn this paper,a platform is built for simultaneous data collection by sensors such as Li DAR and IMU.The communication between sensor nodes is implemented in the ROS operating system.To address the problem of data format error in the LIO-SAM framework for the mine data collected by this experimental equipment,the LIDAR software rslidar＿sdk is configured to add ring and timestamp attributes to the LIDAR point cloud format to transform the LIDAR data format to XYZIRT(x,y,z,intensity,ring,timestamp)format,add rs＿to＿velodyne transformation node in rslidar＿sdk,and then transform the data format to Velodyne radar format.(2)A point cloud map construction method in mining environment is designed and implemented based on the LIO-SAM frameworkIn this paper,we use LIO-SAM framework to build a point cloud map,and for the problem that LIO-SAM framework has poor map building in the unstructured environment of mine,we calibrate the LIDAR and IMU external parameters by collecting data sets with LIDAR and IMU,converting the LIDAR data to IMU coordinate system,and optimizing the calculation so that the data points of the two scans before and after can overlap as much as possible The calibration is completed.By calibrating the IMU internal reference,the quantization noise analysis and zero bias instability noise analysis of the collected IMU data are performed by Allan’s variance method,and the error parameters of the accelerometer and the error parameters of the gyroscope are obtained after calibration.The experiments show that the LIO-SAM framework calibrated by this experimental equipment has better map building effect in the mining environment.(3)Design and implementation of repositioning algorithm based on a priori map of mine environmentThe generated point cloud map has a large number of noise points and a large amount of data.The method to remove the noise points is based on statistical robust filtering,and the point cloud downsampling method is voxel filtering,which reduces the number of point clouds and improves the efficiency of alignment while keeping the point cloud form basically unchanged.Through the normal distribution transformation algorithm,the NDT model describing the spatial surface is established,and the objective function is constructed and solved to achieve the alignment and complete the relocation in the mining environment.The experiments show that the relocation algorithm adopted in this paper can achieve more accurate positioning in the complex environment of the mine.