Research On 3D Reconstruction Technology Of Mine Roadway Based On LiDAR And UWB Information Fusion

With the rapid advancement of the intelligent construction of coal mines,the research and application of autonomous driving technology for mining vehicles have become the key development target.The positioning and mapping technology in the mine roadway is the basis for realizing the automatic driving of mining vehicles.The UWB positioning module currently used can only be used for low-precision positioning of people and vehicles.Also,it cannot meet the high-precision requirements of automatic driving.In this paper,lidar and UWB technology are fused to complete the three-dimensional reconstruction of mine roadway.The main research contents are as follows.(1)Aiming at the pose estimation problem of mining vehicles,a pose estimation method fusing lidar and UWB information is proposed.In order to improve the extraction and matching efficiency of roadway feature points,the feature point cloud extracted by point-to-edge and point-to-point methods is firstly registered by the improved Iterative Closest Point(ICP)algorithm,and then used in the ICP algorithm.The filter filters some redundant points to improve the matching efficiency,and eliminates the abnormal points to ensure the matching accuracy.The distance and position-related covariances of UWB sensor data are obtained through the FIM optimal position calculation method.The method based on the Error State Kalman Filter(ESKF)is designed to fuse lidar and UWB data to obtain UWB pose estimation with directional information.(2)A loosely coupled LU-SLAM method is proposed for accurate 3D roadway maps construction with real geographic coordinate information.Firstly,the UWB global coordinate system is established,and the UWB anchor points are associated with the coordinates in the real environment,so that the global coordinate system established by the UWB positioning system contains the actual geographic coordinate information.Then,the UWB-related global position constraint factor,distance constraint factor and local factor composed of lidar constructed,and add them to the constructed global factor graph model.The global constraint factor and local constraint factor are coupled asynchronously.The back-end optimization is used to improve the accuracy and ensure the robustness and stability about the entire system.(3)The loosely coupled LU-SLAM algorithm proposed is experimentally verified by using the indoor corridor to simulate the single structure of the mine roadway.The results show that the loosely coupled LU-SLAM algorithm can still ensure good positioning accuracy and 3D mapping accuracy in the simulated roadway environment with a single structure and few feature points.