Les commentaires sont clos.

HUMAR - PhD on Kinodynamic Human Skeleton tracking for activity recognition at LIRMM Montpellier, France

24 Mars 2022

Catégorie : Doctorant

HUMAR - PhD on Kinodynamic Human Skeleton tracking for activity recognition at LIRMM Montpellier, France


In the context of our EU project H2020 SOPHIA, we are looking for a highly motivated PhD candidate to collaborate with us on the development of a real-time vision-based estimator of human activities for Human-Robot Interaction.

The selected candidate will work at LIRMM, a cutting edge laboratory of Robotics located in vibrant and sunny Montpellier.

The PhD will start in the fall of 2022.



Design a vision-based (RGB-D) estimator (software library) of the human skeleton, to be used for human activity recognition and contact point/force estimation between human and environment.

The framework must operate at a frame rate that is sufficiently high for robot control.

It will rely on the output of openpose, a state-of-art skeleton detector which operates on single RGB-D frames.

To make the skeleton consistent over time and space, the PhD student should exploit inverse kinematics (expressed as constrained optimization), computer vision, and signal filtering.

S/he will use state-of-art adaptive filtering and the hierarchical optimisation solver developed at LIRMM and LAAS-CNRS, based on Pinocchio (

S/he will enrich this library by participating in the development of a general open-source toolbox to perform human kinematics and kinetics state estimation.

Depending on his/her profile and research interests, s/he may focus on robot control, and exploit human activity recognition in the context of human-robot-interaction.



- Master in a Robotics, Control or Signal/Image Processing (but with strong programming skills)

- expertise in Computer Vision, Optimization, synthesis of observers (Kalman, etc.) and noise filtering

- excellent programming skills in Python/C++/Linux (for real-time application)

- excellent analytical skills & critical thinking

- above average team & communication skills

- very good English skills, written and spoken

Not required but beneficial: hands-on experience with robotic platforms, robotic software frameworks (ROS), and sensors




Interested candidates should submit the following by email before 31st March 2022 in a single PDF file to:

1. Curriculum vitae with 2 recommendation letters

2. One-page summary of research background and interests

3. List of units studied over the past 2 years

4. Previous publications or student projects demonstrating expertise in one or more of the areas mentioned above (optional)




The PhD student will be supervised by Andrea CHERUBINI (Prof, LIRMM), Robin Passama (Res. Eng CNRS) and Vincent Bonnet (Prof, LAAS)





Andrea Cherubini

Full Professor - PU

Head of the IDH Group

Responsable Master Robotique

LIRMM / Université de Montpellier