A mobile robot localization based on a multisensor cooperation approach

Abstract

The authors present an original method of localization based on using complementary measurement data. They use, for their application, a new omnidirectional device made of a panoramic video camera coupled with a rotative LASER range finder. The complementarity of information obtained from these two sensors is used to improve safety in the algorithm of position determination. This algorithm is based on a matching of the straight line projections and the natural vertical beacons taken in the global environment. In order to reduce the computing time for this matching phase, one can extract, from the rotative LASER range finder, a reduced set of possibilities for this matching. The method has been tested on the authors' "SARAH" mobile robot. Results obtained from this experiment are quite promising, especially for real-time applications.