Dynamic Mapping in Confined Spaces: Robotic SLAM with Human Detection and Path Planning

The mining sector holds significant potential for the deployment of advanced autonomous robots, offering enhanced safety, reproducible motion, and reduced human presence in hazardous areas. This paper aims to reduce the response time of mine rescue teams by employing Mine Rescue Robots equipped with LD_LIDAR19 laser scanners and AMG8833 thermal imaging cameras to conduct environmental inspections and rescue operations. The SLAM method facilitates the generation of 2D point clouds, aiding in the creation of precise maps for locating individuals trapped in underground tunnels. The integration of robotics and autonomous systems has the potential to revolutionize exploration in both indoor and outdoor environments. Mobile ground robots with autonomous capabilities find applications in surveillance and reconnaissance. The objective is to create a prototype and show off autonomous capabilities in a mobile robot, including first-person non-line-of-sight exploration, automatic collision avoidance, real-time 2D generation using laser scanning and thermal cameras, waypoint tagging, and autonomous navigation to predefined waypoints. The robot operates on ROS on Raspberry Pi 3B, is connected to an Arduino controller for motor system control, and is equipped with a 2D LiDAR LD19, thermal camera, and IMU. The Hector SLAM algorithm is employed for mapping, while Visual Inertial Odometry (VIO) with cameras enhances mapping capabilities. We will consider incorporating extra functionalities like autonomous target recognition, waypoint designation, collision prevention, and iterative path refinement. The project will reach its conclusion with practical demonstrations highlighting our achievements in autonomous navigation and remote control capabilities, effective collision avoidance, precise waypoint following, and the creation of reliable maps even during high-speed maneuvers.