LiDAR navigation in underground openings

In recent decades, the technological advances in the fields of artificial intelligence, data analytics, computer architecture, and wireless communication enable the ever-increasing integration of autonomous solutions into the mining industry infrastructure. Smart mines are anticipated to increase the mine workers safety and health, as well as increase the performance of the smart mine sites. One of the ways to fulfill that goal is to integrate autonomous vehicles into the mining cycle. Autonomous navigation in underground, hence GPS-denied, environments is a complex and challenging task. Nevertheless, structured underground environments, such as room-and-pillar mines or underground tunnel networks, allow for lightweight solutions. The current research demonstrates the framework for autonomous navigation inside a room-and-pillar panel. An efficient navigation system that incorporates data management, LiDAR mapping, path planning, and control in real-time has been developed for a lab-scale prototype equipped only with 2D LiDAR scanners. A multiple Random Sample Consensus (RANSAC) algorithm extracts salient features from the 2D LiDAR maps, which are input to a Stanley controller to define the vehicle's motion. Simulations in a mock mine section have shown the reliability of the prototype to navigate around underground pillars. The same concepts can be applied to any system of underground openings with linear segments