While a body of work exists on intelligent systems for guiding occupants out of buildings during an emergency, the problem of guiding miners out of a mine during an underground emergency has not been explored. Underground evacuation fundamentally differs from building evacuation because of differences in the regulatory regime and uncertainty regarding the number of occupants. The potential for implementing an intelligent miner guidance system is buttressed by the fact that miner tracking systems, along with communication and sensor networks, already exist in underground mines. This work presents a method based on a local positioning system, in the form of a network of "nodes" at major "intersections," which can then be used to guide miners to safety. We have developed a path-planning algorithm, based on Dijkstra's algorithm and the local positioning system that can provide miners with the optimal escape path in an emergency. The algorithm is validated with scenarios at the Missouri University of Science & Technology Experimental Mine demonstrating the potential to include smart technology in mine emergency management and response planning.