Thesis of Gerardo Ayala Jaimes

Supervision of Over-actuated Steering System of Mobile Omnidrive HeavyVehicle

This work focuses on the modeling and analysis of an over-actuated steering system of a mobile omnidrive heavy vehicle, that is part of a wheeled robot called RobuTAINeR. This robot is used for transporting 12-meters containers inside confined spaces of maritime terminals. In this class of robots, the steering and wheel play a crucial role for the autonomous driving; hence, there is a necessity to understand and model the omnidirectional motion capabilities, switching dynamic behaviors, and its multi-domain areas: electrical, mechanical, hydrodynamic, etc. The Bond Graph technique is a unified methodology to perform and analyze physical systems, where there is power exchange, not only suitable for modeling but also for Fault Detection and Isolation. The validation of this model has been done with real-time simulation, based on real data integrated into the professional simulator SCANeR Studio and compared with the off-line simulation using 20Sim. This latter is used to model complex systems such as the hydraulic steering; the traction system, the longitudinal, the lateral, the yaw and the tire-ground dynamics of the omnidrive robot robot. The fault diagnosis of the steering system is applied based on its hybrid Bond Graph in presence of parameter uncertainties.

Jury

Directeur de thèse : Rochdi MERZOUKI Rapporteurs : Kamal MEDJAHER, Hichem ARIOUI Examinateurs : Jean-Charles MARE, Anne-Lise GEHIN, Kumar Arun SAMANTARAY, Pushparaj Mani PATHAK

Thesis of the team defended on 15/12/2016