Design and Demonstration of Multi-agent System for Underground Mine Rescue Missions

PROJECT ABSTRACT: When the lives of miners are in danger, or the workers self-escape from a hazardous environment is impeded, mine emergency response systems must respond and deploy rapidly. The post-accident mine environment may significantly increase the mission’s risk and the stress on the rescue team. Any delays in the rescue mission can cost the lives of the trapped miners. Self-escape and mine rescue heavily rely on each individual’s level of skills and preparedness to make the correct decision. Introducing an intelligent system to assist mine personnel and mine rescue team during an emergency will significantly improve the likelihood of success while reducing the risks to miners. This proposed research is an interdisciplinary collaboration to improve self-escape and mine rescue through innovative robotic and autonomy solutions. This project brings together a diverse and highly reputable group of faculty from three mining schools (New Mexico Tech, University of Arizona, University of Nevada, Reno), researchers from NIOSH-Mining, Sandia National Lab, NASA, stakeholders, mining companies, and technology developers/licensors. The project goals were developed to address key NIOSH-NORA mining intermediate goal “to improve mine disaster response (objective 2).” We propose to design and demonstrate intelligent mine evacuation and mine rescue systems for underground mining applications. First, we will design a multi-agent robotic system to assist mine rescuers during rescue missions in underground mines. The multi- agent system is a combination of ground and aerial vehicles with complementary capabilities to complete a rescue mission. Second, we will develop algorithms to assist miners to find the safest and fast paths to safety. The algorithm relies on the existing communication and tracking system incorporating human factors into the simulation models. Third, we will design a new communication system to optimize locating, tracking, and communicating with trapped miners and provide AI-assisted self-escape when possible. Finally, we will create and deliver technologically driven hybrid training products to improve self-escape and mine rescue that incorporate the tools and strategies developed in this project. The ‘hybrid’ modules that serve as a bridge between traditional tabletop activities, hands-on training activities and full-scale synthetic learning environments. This effort will also be focusing on developing expertise in the area of mine health and safety by supporting eight Ph.D. and five M.S. students. The research to practice (r2p) plan for this research include dissemination of the project findings through publications, training, and workshops, collaboration with industry partner to technology development and enhancement, and technology implementation with collaboration of NIOSH-Mining and stakeholders. The intermediate outcomes of the proposed work include adoption of the developed technologies and strategies through collaboration with partner mining operations, and technology developer/licensor, new competency-based training products that incorporate project results, as well as publications and patents. The end-goal of the project is to introduce new autonomous solutions to improve self-escape and mine rescue.