基于矿区巷道巡检机器人的LOAM-SLAM地图重建改进算法的研究

Research on Improved Algorithm of LOAM-SLAM Map Reconstruction Based on Mine Roadway Inspection Robot

由于井下路况复杂,矿区巷道巡检机器人在井下获取点云数据时,点云中的点会随着激光雷达运动产生运动畸变,即点云中的点会相对实际环境中的物品表面上的点存在位置上的误差,从而无法获得正确的里程计信息,导致既定路线精度较低,容易造成井下安全问题。通过研究SLAM建图算法-LOAM算法,分析该算法利用三维空间中运动的两轴单线激光雷达将定位和建图分别进行,从而提高精度。在该算法的基础上,同时结合井下成本,选择单轴单线激光雷达进行数据采集,进行点云匹配时,采用Harris 3D角点检测,选取角点作为点云帧数据中的关键点进行匹配,在保证计算量的同时,去除部分冗余数据。定位部分则利用EKF(扩展卡尔曼滤波),消除运动畸变,通过状态方程和预测方程,对机器人进行运动估计,提高定位精度,达到建图所需要求。

Due to the complex underground road conditions,when the mining roadway inspection robot obtains the point cloud data underground,the points in the point cloud will produce motion distortion with the movement of lidar,that is,the points in the point cloud will have position errors relative to the points on the object surface in the actual environment,so it is unable to obtain the correct odometer information,resulting in low accuracy of the established route,It is easy to cause downhole safety problems. By studying SLAM mapping algorithm LOAM algorithm,this paper analyzes that the algorithm uses the two axis single line lidar moving in three-dimensional space to locate and map respectively,so as to improve the accuracy. On the basis of this algorithm,combined with the underground cost,the single axis single line lidar is selected for data acquisition.When matching the point cloud,Harris 3D corner detection is used,and the corner is selected as the key point in the point cloud frame data for matching. While ensuring the amount of calculation,some redundant data are removed. In the positioning part,EKF(Extended Kalman Filter) is used to eliminate the motion distortion. Through the state equation and prediction equation,the motion of the robot is estimated to improve the positioning accuracy and meet the requirements of map construction.