Who am I ?
At Aldebaran I am a researcher-engineer working in the field of Artificial Intelligence at large. My goal is to look at what is done in the labs and transfer the interesting bits to our robots. A second, but equally important goal is to contribute to research based on our experience on real robots. All this work is conducted thanks to the numerous collaborations between and Aldeberan and multiple universities.
I worked previously in multiple laboratories on the use of learning algorithms for multi-agent systems. I targeted the challenges arising when agents have to adapt to the environment, and produce a satisfying behaviour at the level of the population.
I worked on swarms of robots learning behaviours in unknown environments. This was first done during my Ph.D. in the context of the European project Symbrion (IP,FP7), and then during a Post-Doc at the NTNU in the CRAB team under the supervision of Pauline Haddow.
I worked also on the modelisation of the Roman trade network through multi-agent methods in the EPNet project. The project is investigating the co-evolution of culture and trade. I worked on the development of original modeling methods relying on multi-agent models and behaviour learning.
I am participating to the development of two simulation platforms:
- Roborobo : A fast 2D robotic simulator. https://github.com/nekonaute/roborobo3
- Pandora : A parallelized flexible multi-agent simulator. http://xrubio.github.io/pandora/
I took an active part in the development of the ChIRP robotic platform. More details can be found here ChIRP
You can find more about my technical developments on my github page:
montanier [dot] jeanmarc [at] gmail [dot] com