Description
Date depot: 30 octobre 2019
Titre: Evolutionary robotics for morpho-functional swarm robotics
Directeur de thèse:
Nicolas BREDECHE (ISIR (EDITE))
Directeur de thèse:
Olivier DAUCHOT (Gulliver lab)
Domaine scientifique: Sciences et technologies de l'information et de la communication
Thématique CNRS : Non defini
Resumé:
We are looking for a Ph.D student to work on an ANR-funded multi-institution project “Morpho-functional Swarm Robotics' under the supervision of Prof. Nicolas Bredeche (http://pages.isir.upmc.fr/~bredeche/).
You will participate to a swarm robotics project at the crossroad of evolutionary robotics and active matter. This exciting project explores both the physical and logical capabilities of small-scale robots (~2 cm in diameter) in order to design self-adaptive selg-organizing robot swarms. The goal is to exploit physical interactions between robots to perform collective behaviours using minimal logical computation (e.g. changing global alignement by modulating only the shape or the speed of the robots, collective transport through synchronised collision events, etc.).
You will design learning and exploration algorithms for large swarm robotics experiments, using both simulation and real robots. Your task will be to develop tools to explore the space of collective behaviours using exploration algorithms (QD, Map-Elites, ...) and to design new lifelong learning algorithms for swarm robotics (embodied ER, social learning, multi-agent on-line learning).
Skills: previous experience in Evolutionary Robotics or Artificial Life. Excellent skills in experimental design, experimental practice and rigorous data analysis. Excellent programming skills, preferably in C++ and Python. Previous experience with real robots (e.g. Kilobots) is welcome, but not mandatory.
Doctorant.e: Mirhosseini Yoones