Abstract:

With the continued expansion of the mining industry, underground mining has become increasingly vital. However, underground environments are among the most hazardous and complex workplaces, characterized by confined spaces, dense obstacles, high humidity, and pervasive dust-factors that significantly hinder automation and pose substantial safety risks. In light of these challenges, robotic systems equipped with articulated arms have emerged as a promising solution to improve operational efficiency, enhance worker safety, and support intelligent mining operations. This study proposes a collision avoidance framework for robotic manipulators operating in underground settings. The kinematic model of the robotic arm is formulated using the Denavit–Hartenberg convention, incorporating both forward and inverse kinematics to enable precise motion planning and obstacle assessment. A visual sensing system based on RGB cameras is integrated to detect obstacles and estimate distances in real time. Building on this sensor data, a proactive control algorithm is developed to avoid potential collisions and ensure safe arm movement. The proposed approach is implemented and evaluated through simulations on the Niryo Ned 2 robotic arm within the Robot Operating System (ROS) environment. Simulation results demonstrate the effectiveness and feasibility of the method in constrained and dynamic underground scenarios, indicating strong potential for real-world deployment in modern mining automation systems.