Post-Doctoral Research Visit F/M Swarm behaviour: from mathematical models to experimentation with CrazyFlies drones

Download job offer in PDF format

Contract type : Fixed-term contract

Level of qualifications required : PhD or equivalent

Fonction : Post-Doctoral Research Visit

Level of experience : Up to 3 years

About the research centre or Inria department

The Inria research centre in Lyon is the 9th Inria research centre, formally created in January 2022.  It brings together approximately 320 people in 19 research teams and research support services.

Its staff are distributed in Villeurbanne, Lyon Gerland, and Saint-Etienne.

The Lyon centre is active in the fields of software, distributed and high-performance computing, embedded systems, quantum computing and privacy in the digital world, but also in digital health and computational biology.

Context

This post-doc will take palce in the  Chroma team in Lyon. The team studies algorithms and models to enable autonomous agents/robots to perceive, decide and adapt to their environment, by combining probabilistic methods, machine learning, planning techniques, multi-agent decision-making and constrained optimisation tools. The team develops and maintains several experimental platforms, and is involved in several academic and industrial projects in the field of mobile and aerial robotics.

This two-year post-doc is part of the INRIA/AID  BioSwarm project: Bio-inspired algorithms for collective research and decision-making in drone swarms. The post-doc will be carried out in collaboration with Olivier Simonin (CITI/Chroma) and Emanuele Natale (Inria Sophia Antipolis), co-PIs of the BioSwarm project.

Assignment

In this post-doc, the main objective is to study the transition from simplified discrete models of behaviours observed in nature, particularly in insects, to behaviours that can be exploited by a swarm of UAVs. Some properties or tasks can be performed with a high degree of robustness without requiring heavy or centralised communications. Mathematical models of some of these behaviours make it possible to study and demonstrate these properties, but in return it remains difficult and rarely studied how they can be applied to real autonomous systems, such as a fleet of mobile robots. This transition from the mathematical model to the robotic model, supported by experimentation, is the object of study of this post-doc (as for example in [3]). The experimental part will be carried out using the CrazyFlies mini-UAV platform available in the Chroma team.

The work will focus on two recent models from the literature:

  • The evaluation by agents of their density in a limited space according to a model inspired by the behaviour of ants [1]. This article proposes an algorithm based on random walks. The challenge will be to study how this model, proposed in a discrete 2D framework, can be extended to the continuous 3D context of a fleet of CrazyFlies mini-drones.
  • In [2], the mathematical formalisation of a neurological mechanism in the olfactory circuit of Drosophila enabled a new implementation of a classifier for nearest neighbour clustering (NNC) in the difficult configuration of federated learning (FL). We will examine how this model can be adapted to a learning task in a fleet of autonomous drones with limited communications.

Interactions with the PhD student from the BioSwarm project, working on mathematical modelling, is desirable. In addition, the experimental work could be based on prior simulation work, in particular with the [Gym PyBullet Drones software] tool (https://utiasdsl.github.io/gym-pybullet-drones/), used in the thesis.

The post-doc recruited will be part of the Chroma team working on several robotics projects, and will work in collaboration with the research engineer in charge of the UAV platform (Nicolas Valle). He/she will aim to publish work in international conferences and journals in AI and robotics.

[1] C. Musco, H-H. Su, and N. A. Lynch , "Ant-inspired density estimation via random walks" Proc. Natl. Acad. Sci. (PNAS), 114 (40), pp 10534-10541, 2017

[2] P. Ram and K. Sinha "Federated Nearest Neighbor Classification with a Colony of Fruit-Flies", Proceedings of the AAAI Conference on Artificial Intelligence, 36(7), pp 8036-8044. 2022.

[3] A. Glad, O. Simonin, O. Buffet, F. Charpillet "Influence of Different Execution Models on Patrolling Ants Behavior : from Agents to Robots", 9th International Conference on Autonomous Agents and Multiagent Systems (AAMAS'10), 2010.

Main activities

The activities of the post-doc recruited will consist of :

  • Studying and adapting mathematical models and bio-inspired algorithms to the control of a fleet of aerial robots.
  • Implementing and carrying out experiments with a fleet of CrazyFlies drones. A preliminary simulation study should be necessary.
  • Evaluate the performance of the tasks carried out with drones, and analyse the properties initially defined in the mathematical model.
  • Disseminate the results via publications in AI and robotics conferences and journals.

Skills

PhD in Robotics or AI/Computer sciences or related fields

  • Good theoretical and experimental background in robotics and/or AI

The following qualifications would be an advantage :

  • Experience in deploying mobile or aerial robots
  • Experience in using robotic simulators and/or modelling multi-agent or living systems
  • Ability to work in a team and with other researchers
  • Fluency in French and English (written and spoken)

Benefits package

  • Subsidized meals
  • Partial reimbursement of public transport costs
  • Leave: 7 weeks of annual leave + 10 extra days off due to RTT (statutory reduction in working hours) + possibility of exceptional leave (sick children, moving home, etc.)
  • Possibility of teleworking (90 days / year) and flexible organization of working hours Social, cultural and sports events and activities
  • Access to vocational training
  • Social security coverage under conditions

Remuneration

2788 € gross salary / month