Research on mine robot hazard avoidance method based on layer normalized long short-term memory

The safety and efficiency of mine robots are critical in ensuring autonomous operations in hazardous mining environments. As artificial intelligence and deep learning technologies evolve, there is a growing need for intelligent hazard avoidance syste ms capable of handling the complexity of underground conditions. This paper presents a novel hazard avoidance method for mine robots based on Layer Normalized Long Short-Term Memory (LSTM) networks. The proposed method leverages the ability of Layer Normalized LSTM to process sequential data with improved stability and faster convergence, making it particularly effective in dynamic and unpredictable underground environments. By integrating this LSTM-based approach with traditional hazard avoidance techniques, the system can more accurately predict potential risks and make real-time adjustments to the robot’s path. The model is trained on various environmental and operational datasets, allowing it to generalize across differ ent mine conditions. Experimental results demonstrate that the Layer Normalized LSTM-based method significantly enhances hazard avoidance accuracy and operational safety compared to conventional methods, offering a reliable and energy-efficient solution for autonomous mine robot navigation. This research contributes to the development of intelligent underground robotic systems, optimizing both hazard avoidance and operational efficiency.