Probabilistic Methods for Mobile Robot Mapping

Abstract

The problem of map building is the problem of determining the location of entities-of-interest in a global frame of reference. Over the last years, probabilistic methods have shown to be well suited for dealing with the uncertainties involved in mobile robot map building. In this paper we introduce a general probabilistic approach to concurrent mapping and localization. This method poses the mapping problem as a statistical maximum likelihood problem, and devises an efficient algorithm for search in likelihood space. We furthermore address the problem of using occupancy grid maps for path planning in highly dynamic environments. The approaches have been tested extensively and several experimental results are given in the paper.