I am interested in the design and understanding of intelligent systems that have to operate in an environment they can perceive through sensors and act on it. These kind of systems are called situated agents (such as robots or decision-support systems).
My research focuses on fondation for building or understanding such systems. The core of my research addresses techniques for decision-making (planning, learning, spatial computing) that enable single situated agent, a teams of agents or a huge number of agents (swarm intelligence, distributed computing, web, …) to act intelligently for solving problems.
Theory of Markov Decision Processes (MDPs) is particularly powerful in that context. In particular, I’m highly interested in extending the Markov decision process framework to problems of decentralized control. One studied framework is a generalization of a partially-observable MDP (POMDP), which we call decentralized partially-observable Markov decision process (DEC-POMDP).
An other interesting approach I’m interested in, is to design collective systems whose behavior imitate natural complex systems such as ant colonies, flocking of birds, …
Applications of my research includes planatary exploration, decision support for medecine (especially for telemonitoring elderly people or people suffering from chronical pathologies), signal understanding, designing and understanding large-scale, distributed, robotic, computer and natural systems (new issues I want to address more deeply in the futur) with potential applications in robotics, urban security or transportation, military reconnaissance, and environmental monitoring.
This work is a pluridisciplinary research on areas such as Artificial Intelligence, Decision theory, Stochastic control, Biology, Medicine, complex systems.