Uncertainty Treatment in a Surface Filling Mobile Robot

We have elaborated an autonomous mobile robot capable of exploring and navigating the entire accessible area in a closed a-priori unknown environment. In order to carry out this mission several modules execute in parallel as the robot evolves. The main modules developed are: the navigation strategy, the localisation system and the map building procedure. This paper shows the models and mechanisms used to overcome uncertainty of the sensors and of the motion system. Robustness of the system is augmented both by considering uncertainty inside each module and by an intrinsic co-operation between modules. A high level of performance has been reached at low hardware requirements.