Internship: humanoid robotics / parametric-task solver to connect reinforcement learning and model-based control

Contract type : Internship agreement

Level of qualifications required : Master's or equivalent

Fonction : Internship Research

Context

 

The HUCEBOT team is a new team of the Center Inria at the University of Lorraine. 

The team is dedicated to advancing algorithms for human-centered robots: robots that are not working autonomously in isolation, but that instead react, interact, collaborate, and assist humans. To do so, these robots need to intertwine a multi-contact whole-body controller, a digital simulation of the interacting humans, and machine learning models to predict and respond to human movements and intentions. In a crescendo of complexity, the team will tackle scenarios that involve collaboration with cobots,  assistance with exoskeletons, and whole-body teleoperation of humanoid robots. The application domains span from industrial robotics to space teleoperation.

The main robots of the team are the Tiago++ bimanual mobile manipulator, the Unitree G1 humanoid, and the Talos humanoid robot. The team also works with Franka cobots and exoskeletons.

The intern will work with a team of about 25 members, including permanent researchers, PhD students and post-doctoral researchers.

Assignment

The topic of this internship will be to work on bridging the gap between reinforcement learning and model-based approaches for humanoid control. It will be focused on developing a new kind of solver that can solve many similar problems in parallel more efficiently that the current quadratic programming or non-linear programming solver.

The starting point will be the recently introduced Parametric-task MAP-Elites [1], that will need to be connected with modern solvers like ProxQP that can take advantage of warm-starting [2].

[1] Anne T, Mouret JB. Parametric-Task MAP-Elites. In Proceedings of the Genetic and Evolutionary Computation Conference 2024 Jul 14 (pp. 68-77).

[2] Bambade A, El-Kazdadi S, Taylor A, Carpentier J. Prox-qp: Yet another quadratic programming solver for robotics and beyond. InRSS 2022-Robotics: Science and Systems 2022 Jun 27.

Main activities

Main activities (5 maximum):

  • programming experiments (most probably in C++, possibly in Python)
  • analyze and report results

 

Skills

Technical skills and level required:

  • Python programming
  • C++ would be appreciated

Languages: English (French is not required -- the official language of the team is English)

 

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 (after 6 months of employment) and flexible organization of working hours
  • Professional equipment available (videoconferencing, loan of computer equipment, etc.)
  • Social, cultural and sports events and activities
  • Access to vocational training
  • Social security coverage

Remuneration

€4.35/hour