Postdoctoral position in robotics (M/F): Modeling the Adaptation of Human Movement to an Assistive Device

Job Description

Missions

Wearable robotic devices designed to assist with physical tasks are a potential solution for physically supporting humans at work, thereby reducing exposure to risk factors for musculoskeletal disorders. Numerical simulation of the human-robot system is a promising approach for guiding the design of these assistive devices. However, current simulation approaches synthesize human movements in a similar manner regardless of whether assistance is present, which is unrealistic. While the synthesis of human movements has been the subject of extensive research in several fields, no model currently exists that can predict motor adaptation with robotic assistance.

The objective of this project is to simulate the changes in human movement induced by the use of a wearable assistive device. This will enable the development of predictive models of human movement with assistance, which are necessary for realistic motion synthesis. The project will focus on upper limb assistive devices (e.g., shoulder assistance), which are the most common for professional applications, as musculoskeletal disorders primarily affect the upper limbs.

The synthesis of human movements generally relies on optimal control methods or data-driven approaches. Inverse optimal control (IOC) or inverse reinforcement learning (IRL) make it possible to identify criteria for movement optimality based on a limited amount of experimental data. These criteria can then be used to synthesize movements in situations not observed experimentally (provided the movement remains similar in nature to those studied). However, when a wearable assistive device is added, two elements may change, both of which affect the solution to the optimization problem: the system's internal model and/or the cost function.

In this project, we will use techniques based on optimality principles (IOC/IRL) to determine whether and how changes in the cost function can predict changes in human movement when using an assistive device, depending on the level of modeling adopted for the human-robot system. The first phase of the project will consist of acquiring experimental motion data (kinematic and dynamic) recorded with and without an assistive device, in order to quantify and characterize the postural changes induced by the assistance. The targeted activities are overhead work and load handling, which are the primary tasks where professional assistive devices can be used. Particular attention will be paid to motor adaptation based on the load carried with and without assistance.
An IOC or IRL analysis will then be performed on the collected data to identify and compare the relevant biomechanical optimality criteria in cases with and without assistance. The candidate cost functions, as well as the representation of the system dynamics required for IOC/IRL, depend on the level of modeling chosen for the human-robot system. We will therefore compare different levels of modeling, in particular purely skeletal models (without muscles) and musculoskeletal models, to determine the most appropriate model for balancing computation time and the generalizability of the results. We will explore the inclusion of a comfort cost related to forces at the human-robot interface in the cost function, which will raise questions regarding the modeling of force transmission between the human and the assistive device.
Given the variability of human movement, it is possible that different strategies will be observed among the subjects. The IOC/IRL analysis can then be supplemented by a clustering analysis to identify and characterize different motor strategies. Finally, the results will be validated by comparing the results of the predictive simulation—using the developed motor adaptation models—with experimental data different from that of the initial data collection. To do this, we will synthesize human-robot system movements, using optimal control or reinforcement learning, based on the previously identified cost functions, and compare the kinematics and dynamics of the resulting movement to the experimental data.

Activity

- State of the art regarding the models and cost functions used for IOC and IRL applied to upper limb motion, as well as the available IOC and IRL tools.
· Selection and implementation of an IOC or IRL solution based on existing libraries.
· Definition of the experimental protocol (including drafting the application for ethics committee approval) and collection of data on human movements assisted and unassisted by assistive devices.
· Application of IOC/IRL to the experimental data and analysis of the results.
· Drafting of a scientific article presenting the approach and the results obtained.
· Presentation of the work at at least one national or international conference.

Your Profil

Skills

- Education: Ph.D. in Robotics (with an interest in biomechanics) or in Biomechanics (with a computational component and an interest in simulation);
- Skills: Modeling of multi-joint systems (kinematics and dynamics), robotic control, C++/Python programming, motion capture, experience with human participants; knowledge of optimal control and reinforcement learning is appreciated; knowledge of biomechanics and human motion analysis is a plus;
- Languages: English and French.

Your Work Environment

- Team Hucebot of LORIA in Nancy : https://team.inria.fr/hucebot/team/
- The postdoctoral researcher will have access to the lab cafeteria. The employer will cover part of the cost of public transportation.

Compensation and benefits

Compensation

About 3000€ gross salary

Annual leave and RTT

44 jours

Remote Working practice and compensation

Pratique et indemnisation du TT

Transport

Prise en charge à 75% du coût et forfait mobilité durable jusqu’à 300€

About the offer

About the CNRS

The CNRS is a major player in fundamental research on a global scale. The CNRS is the only French organization active in all scientific fields. Its unique position as a multi-specialist allows it to bring together different disciplines to address the most important challenges of the contemporary world, in connection with the actors of change.

CNRS

The research professions