Postdoc in swarm robotics and learning

Nov 05, 2019
Dec 05, 2019
Organization Type
University and College
Full Time
Open position for a postdoctoral researcher in swarm robotics and learning (2 years)

What? To explore and design on-line learning algorithms for a swarm of real robots

Where? Sorbonne Université, Paris, France Institut des Systèmes Intelligents et de Robotique (ISIR/CNRS)

We are seeking a postdoctoral researcher with skills in evolutionary robotics and/or swarm robotics, starting as soon as possible, to work in our ANR-funded "morpho-functional swarm robotics" project (joint project between Sorbonne Université and ESPCI).

The successful candidate will participate to a swarm robotics project at the crossroad of robotics, computer science and active matter. She/he will design learning and exploration algorithms for large swarm robotics experiments, taking into account both the physical and logical capabilities of small-scale robots (~2 cm in diameter). In particular, the candidate will explore how to exploit physical interactions between robots to obtain collective behaviours with minimal logical computation (e.g. changing global alignement by modulating only the shape or the speed of the robots).

The main task of the candidate will both develop tools to explore the space of collective behaviours using exploration algorithms (QD, Map-Elites, and alike) and design new on-line distributed learning algorithms for swarm robotics (embodied ER, multi-agent machine on-line learning).

Most (preferably all) of the work is expected to be conducted with real robots.The robots considered are capable of simple communication and self-propulsion, and function in swarms of 1000+ units (similar to Kilobots, with extensions). We are currently using a swarm of 300 Kilobot robots, and developing our own new robotic platform with extended capabilities at a similar size. Participation to the on-going development is possible, but not mandatory.

Required qualifications: evolutionary robotics, swarm robotics, quality-diversity algorithms

Optional: evolutionary game theory, background in mechatronics and/or electronic engineering, statistics physics of active matter.

Salary depends on the candidate's number of years of experience and includes the national health insurance. The location is at Sorbonne Université, campus Pierre et Marie Curie in central Paris, France (Metro: Jussieu, RER C: gare d'Austerlitz).

The candidate will integrate the project team at Sorbonne Université, that includes two Ph.D, one engineer and several interns at master level all working on this project. He/she will also benefit from the larger ecosystem of the AMAC team (e.g. working coffee machine, and social environment with 25+ people working on topics from computational neuroscience to learning in robotics), the ISIR laboratory (e.g. support team of engineers, 3D printers, etc.) and of the University (e.g. sport facilities).

The project is lead jointly with a statistical physics team at Gulliver lab, ESPCI (Ecole Supérieur de Physique et Chimie Industrielle), with weekly meetings. Both teams can access each other resources. The project is funded by the Agence Nationale de la Recherche until 2022. All production done in the project will be open-source (software and hardware).

The job can start as early as October 2019, and the position will remain open until filled.

Review of applicants will begin immediately, and will continue until the position is filled.

Similar jobs

Similar jobs