Using laser range data to model tunnel curvature for the automatic guidance of a mining vehicle

The paper presents an automatic guidance system for semi-structured environments such as mine tunnels. This system uses a laser range finder to obtain information about the environment. In order to guide the vehicle properly, a perception system estimates, from the range data, the tunnel curvature and the lateral and orientation offsets of the vehicle with respect to the center of the tunnel. The vehicle considered in the study represents a small scale articulated mining vehicle. Two different methods are proposed and compared for the perception system. The first one, based on the least-squares approach, requires much range data. The second method makes use of a Kalman filter and needs only two range data. Different tests in simulation illustrate the performance of the two approaches for automatic guidance