Ready-to-go Biomechanical Models

Summary

The current maturity of musculoskeletal modelling and personalizing tools, and the democratisation of medical imaging data, mean that in the near future it will be possible to create detailed biomechanical models of humans that can be personalised with just a few measurements and a few clicks. This objective aim at representing specific subjects (high-level athletes, specific disabilities), or particular populations (workers in a given sector, age cohorts, etc.). To achieve this objective, it is necessary to develop both statistical learning methods to reduce the specific data required to establish a model representative of a given population, and model-informed methods based on anatomical and physiological knowledge to obtain unique characteristics representative of the subject’s physical capabilities.

Long term scientific objective

In this research axis, the aim is to propose musculoskeletal models (models of rigid polyarticulated solids, actuated by muscles and ligaments) validated for the analysis of physical activity (sport, people at work, clinical exercise…), which can be individualised within a few clicks and are statistically representative of specific populations. The challenge of individualising and generalising musculoskeletal models is a relatively old problem, and one that is still poorly solved. The emergence of high-precision motion capture tools, efficient statistical learning methods and the availability of medical imaging databases with large cohorts mean that new methods of individualisation and statistical representation of populations are now possible. In addition, the increase in computing capacity means that models can be made more complex and more physiologically realistic. Such models will be intensively used in all other axes of the team, being used for biofeedback in XR experimentations, as a digital twin in simulations, and as an evaluator for biomechanical quantities for field analyses.

Figure 1: Illustrations of previous work around designing and personalizing of musculoskeletal models. From left to right, top to bottom: a) A lower limb musculoskeletal model (A.Muller), b) An imagery-based model used to validate a geometrical calibration method (P.Pucchaud, c) A muscle path modeling method (C. Livet), d) Kinematical data superposed on video images and subsequent configuration of a musculoskeletal model of the hand and forearm e) Experimental setup aiming at personalizing muscular models (D. Haering) f) Musculoskeletal simulation of muscle moment arm during maximal voluntary flexion-extension aiming at personalizing muscular models.

Short term goals and actions

We are currently working on the following goals:

Range of models for backpack carriage analysis: In this specific PhD thesis, funded by the chaire Safran-Saint Cyr “The augmented soldier in the numerical battlefield”, we will develop a specific approach able to provide a range of models statistically representing the French soldier on the basis of a large database. Such range of models will be useful to design the optimal backpack carriage rules able to relieve the biomechanical constraints endured by the soldiers during their mission.
Personalized upper limb disabled models: Within the frame of the Exoptim post doctoral project, we aim at developing personalized musculoskeletal models of people presenting muscle weaknesses. In particular, we will combine medical imaging, isokinetic data and machine learning methods to adjust the strength capabilities of the model to the subject, in order to adjust the control of an assistive device to the subject.
Integrated motion analysis tools: Our motion analysis library CusToM (Mullet et al. 2019), currently developed in matlab, and currently being ported in python, is already guaranteeing compatibility between model parts at any descriptive levels. This objective will be to go further in guaranteeing a larger application spectrum of the models by being inter-operable among the main musculoskeletal platforms available. We are also seeking at developing partnerships with several Inria teams interested in such features to develop a common library/platform able to provide motion analysis/synthesis and human-system interaction assets.
Detailed upper limb musculoskeletal model: Within the frame of the CAPACITIES project, we are still finalizing a detailed musculoskeletal model of the upper limb and the shoulder to provide meaningful biomechanical quantities related to wheelchair locomotion. This model uses the most advanced works we developed on musculoskeletal modeling to achieve a high level of personnalization and fidelity.