A cognitive modeling of space using fingerprints of places for mobile robot navigation

In this work we address the problem of perception, spatial cognition and topological navigation for a mobile robot. The objective of this work is to enable the navigation of an autonomous mobile robot (or vehicle) in an indoor (or outdoor) structured environment without relying on maps a priori learned and without using artificial landmarks. A new method for incremental and automatic topological mapping and global localization using fingerprints of places is presented. The fingerprint-based representation permits a reliable, compact and distinctive environment-modeling. Experimental results for mapping indoor and outdoor environments with a mobile robot and a "SMART" vehicle, both equipped with a multi-sensor system composed of two 180deg laser range finders and an omnidirectional camera are also reported