Path Planning for Autonomous Unmanned Ground Vehicles in Underground Mining

Autonomous robotic navigation in underground mining environments poses significant challenges due to confined spaces, poor lighting, and the absence of GPS signals. This study presents the design and implementation of an autonomous navigation system for the Husky unmanned ground vehicle (UGV), utilizing LIDAR-based Simultaneous Localization and Mapping (SLAM) within the Robot Operating System (ROS) framework. The system enables real-time mapping, obstacle avoidance, and both global and local path planning in GPS-denied environments. The performance of navigation system was validated through simulations in Gazebo and field tests in two physical environments: the Bunker Lab at New Mexico Tech and the Missouri S&T Experimental Mine. These tests confirmed the Husky’s ability to navigate complex terrain and generate accurate 2D occupancy maps without prior environmental knowledge.