Vehicle Positioning in Underground Mines

This paper describes the problem of locating a vehicle relative to a digital map of an underground mine. The mines of interest produce potash, and they are at a depth of approximately 1000 meters, being relatively large, covering an area greater than 10 kilometers by 10 kilometers. A major application of vehicle positioning is in tracking ground penetrating radar. Ground penetrating radar is used to ascertain the current condition of the mine ceiling and evaluate its risk of delamination (i.e. collapse). A ground penetrating radar system is driven along a tunnel and measurements are logged. It is necessary to record position information along with the radar signal, and this can be an expensive task without a positioning system, since there are many kilometers of mine tunnels. Two possible problem solutions are proposed each using different absolute measurements. For the first case the distance to the mine walls is used. In the second case an existing wireless ethernet network furnishes the positioning system with distances from local access points. The system model is the same for both formulations and uses a state space representation. Both formulations also make use of inertial measurements such as acceleration. A particle filter, which is a recursive Bayesian technique using sequential Monte Carlo sampling, is used to estimate the state and, therefore, the position of the vehicle.