T-SLAM: Registering topological and geometric maps for robot localization in large environments

This article reports on a map building method that integrates topological and geometric maps created independently using multiple sensors. The procedure is termed T-SLAM to emphasize the integration of Topological and local Geometric maps that are created using a SLAM algorithm. The topological and metric representations are created independently, being local metric maps associated with topological places and registered at the topological level. The T-SLAM approach is mathematically formulated and applied to the localization problem within the Intelligent Robotic Porter System (IRPS) project, which is aimed at deploying mobile robots in large environments (e.g. airports). Some preliminary experimental results demonstrate the validity of the proposed method.