Research On Technologies Of Mapping And Autonomous Navigation For Tunnel Drilling Robot

As an important energy source in China,the intelligent development of coal mining techniques is the only way for its high-quality development.Intelligent coal robot is a new engine to drive the high-quality development of coal industry,which has certain economic and social benefits.As an important member of the coal robot,the drilling robot has a wide range of application scenarios.The visualization of its working environment and its autonomous navigation are the key to intelligence.This thesis proposes a tunnel mapping based on Li DAR and inertial sensors and the autonomous navigation strategy of drilling robot.Through the sensor fusion method,the three-dimensional point cloud map of the underground tunnel is constructed,which effectively realizes the visualization of the unknown environment,and provides map guidance for the autonomous navigation of drilling robot.Based on the point cloud map,the navigation and obstacle avoidance is realized,the environmental obstacles are effectively distinguished,and the safe and accessible path is planned,This will provide a solution strategy for autonomous navigation of drilling robot.The main research contents of this thesis are as follows:According to the special environment of coal mine tunnels and the specific working conditions of drilling robots,this thesis analyzes the autonomous navigation requirements of rock drilling robots,completes the autonomous navigation function design of drilling robots,and proposes an autonomous navigation strategy based on Li DAR and inertial fusion.A mobile robot autonomous navigation system experimental platform was designed and built.Data processing and calibration of selected Li DAR and inertial sensors have been carried out,laying the foundation for subsequent mapping and autonomous navigation.In view of the excessive noise point cloud in the tunnel environment,effective geometric features are extracted to remove redundant noise and complete point cloud feature matching.To explore the degradation characteristics of tunnel point cloud,aiming at the problem of incomplete state estimation of degraded environment,the robust optimization of point cloud feature matching in degraded environment is carried out,and a feature matching method based on degradation detection and intensity assistance is proposed.The SLAM framework of Li DAR-inertial sensor fusion based on graph optimization is established,which is composed of two steps of optimization process of inertial odometry and Li DAR-inertial odometry,to achieve high-precision and high-robustness 3D mapping.In view of the serious accumulated error of localization based on track calculation,and the ambiguity of similar scenes in localization based on known map,the research on Li DAR-inertial sensor fusion navigation and localization with known map constraints is carried out to output the absolute coordinates under the global map coordinate system.For robot path planning under obstacle constraints,point cloud ground segmentation is used to ensure that the robot can distinguish between obstacles and accessible areas in space,and a curvature continuous obstacle avoidance path is output through a cubic spline curve.The mapping and autonomous navigation experiments were conducted on an autonomous navigation experimental platform.The experiment results showed that in an environment such as an underground garage,the error of the point cloud map is0.04 m,and the error of the odometry is 0.017 m;The navigation error is 0.09 m.In a simulated tunnel environment,the error of the point cloud map is 0.06 m,and the error of the odometry is 0.030 m;The navigation error is 0.12 m.The SLAM experiment on a drilling robot shows that the improved SLAM method still has high applicability even under severe vibration and harsh environmental conditions of the drilling robot,with an odometry error of 0.031 m.The experimental results indicate that the method proposed in the thesis can provide high-precision environmental maps and complete autonomous navigation,providing a theoretical reference for achieving autonomous navigation of tunnel excavation equipment.This thesis has 71 pictures,14 tables,108 references.