Vehicle localization in underground mines using a particle filter

In this paper a particle filter is applied to the problem of tracking the position of a moving vehicle in an underground mine using only inertial sensors and a digital map of the mine. This is useful for several applications, one of which is tracking the position of a ground penetrating radar system. The ground penetrating radar is used to ascertain the current condition of the mine ceiling and evaluate its risk of failure (i.e. collapse). A ground penetrating radar system is driven along a tunnel and measurements are logged. It would be useful to record position information along with the radar since there are many of kilometers of mine tunnels. Although particle filters have been applied to vehicle tracking before, they have not been applied to the problem of tracking a vehicle in an underground mine environment. A particle filter is a probabilistic approach to estimating the next state based on the current observation. There are three main components to a particle filter. They are particle generation (sampling), determining and normalizing the importance weights (importance), and resampling the particles (resampling). Each of these components are described for the given application. The vehicle tracking model is presented and its particle filter implementation is described