|
From the understanding of visual navigation mechanisms in mammals to autonomous mobile robot applications| old_uid | 5926 |
|---|
| title | From the understanding of visual navigation mechanisms in mammals to autonomous mobile robot applications |
|---|
| start_date | 2009/01/08 |
|---|
| schedule | 14h |
|---|
| online | no |
|---|
| summary | This seminary aims at highlighting by two concrete examples, how the interplay between robotics and the sciences of the living can promote both the understanding the cerebral circuitry and the achievement of very robust algorithms for robotics.
The problem of the development of a neural architecture, inspired by neurobiological models of the cerebral circuitry, for autonomous navigation of mobile robots in unknown indoor and outdoor environments, will be first addressed . A localization system based on a model of purely visual place-cells is presented. The place-cells enable to learn reactive sensory-motor behaviors and planning behaviors. The system is able to learn behavioral tasks autonomously or by interacting with a human teacher. An orientation system, merging a visual compass and path integration information, will also be proposed.
The second example will illustrates how the great performances of attentional tracking systems and bio-inspired depth estimation algorithms allow to consider purely visual-based pursuit-evasion games in presence of obstacle in unknown environment. A short survey on the problem of pursuit-evasion games in cluttered unknown environment is first proposed. Some conditions for the pursuer/evader to switch its policy from minimizing/maximizing a heuristic to a direct-chase/run-away strategy are considered. New heuristics called the temporal and the spatial hidability dealing both with capture and visibility maintenance are proposed and compared to other heuristics relying on the escape risk and the vantage time against a smart evader, in order to highlight the advantages and the drawbacks of these methods.
Ch. Giovannangeli is currently involved in the European project Visiontrain as a postdoctorant researcher for the Technion, Haifa, Israel, where he works on pursuit-evasion games and computer vision for mobile robots. He obtained his PhD diploma, 2007 December the 13th, in the Neurocybernetic team of the Laboratory ETIS (Equipe de Traitement d’Image et de Signal). The subject, co-financed by the DGA (Délégation Générale pour l’Armement) and the CNRS (Centre National pour la Recherche Scientfique), raises the problem of neurobiologically inspired algorithms for both indoor and outdoor navigation, the learning of sensory-motor associations, the planning, and the interactive learning. In 2003, he received a master degree of engineering in Information Processing Sciences from the EISTI (Ecole Internationale des Sciences du Traitement de l’Information), Human Machine Interfaces as specialization. The same year, he has been graduated with a master of Sciences in Signals and Images Processing from the ENSEA (Ecole Nationale Supérieure de l’Electonique et de ses Applications) the University of Cergy-Pontoise, (specialization in "natural and artificial vision" and "learning and classification architecture"). His researches now focuses on insect and mammalian navigation strategy for the autonomous navigation of UGVs and UAVs : pursuit-evasion games in presence of obstacles, online environment learning, landmarks selection and recognition, sensors fusion, planning, action selection and developmental approaches ; or more generally on all the subjects that could be linked with the design of biologically inspired robots able to assist Humans in numerous domain. |
|---|
| responsibles | Fenouil |
|---|
| |
|