Doctorant sous la direction de C. STOICA

Titre de la thèse: Fault-tolerant predictive control techniques for multi-vehicle formations
Résumé de la thèse: One of the priorities of this PhD thesis will be the robust predictive control for multiple agents’ formations in the presence of faults. In the context of rescue missions (large-scale fire extinguishing missions, object research problems etc.), the coordination and control of a fleet of heterogeneous vehicles become an important issue. These multi-agent missions will imply task allocation of the agents, trajectory/path planning, real-time control issues under constraints and uncertainties, state estimation etc. Centralized/decentralized/distributed receding horizon control techniques based on set-theoretic tools will be used to deal with possible faults, constraints, uncertainties, delays, formation reconfiguration strategies. The safety of the missions is a priority and thus the developed techniques will be essentially built on sensor redundancy necessary for diagnosis and fault-tolerant control for multi-agent systems. The developed control laws will be tested and validated on a multi-vehicle formation, for which simplified models are already available.