Star Field

research interests

My research is mainly centered on autonomous decision making, with a focus on uncertainty (using Markov Decision Processes). Here are some past and present topics of my work:

In Nancy,

during my PhD, I have worked on coordination in a Multi-Agent System through shaping, and on selecting and learning basic behaviors for an agent with multiple simultaneous objectives.

In Canberra,

my work was oriented toward a large planning project (DPOLP), in collaboration with Doug Aberdeen and Sylvie Thiébaux.
We have developped a brand new --and quite successful-- probabilistic planner (FPG, see available software below). Plus, I have studied robust planning (how to find the best policy for the worst model when the model is uncertain).

An other australian project I was involved in is the SuperCom project, which focusses on model-based supervision (monitoring and control) of composite systems (networks of computing components: web services, telecommunication/distribution networks...). I have worked on a planning algorithm adapted to such systems (in collaboration with Elena Kelareva, Sylvie Thiébaux and Jinbo Huang).

In Toulouse,

my main focus is on making space machines (satellites, rovers) autonomous. This involves designing an appropriate models and algorithms for state estimation and decision-making.

Related links


Last modified: Wed Aug 22 22:15:19 CEST 2010