The accurate quantification of internal forces in the human body in motion is still a challenge and a necessity to better understand motion-linked pathologies. In this context, the objective of this research is to develop a method to quantify the joint efforts and the corresponding muscle forces in the human body in motion. The main originality of the proposed method for the muscle force quantification is the cautious use of electromyographic (EMG) data information, known to be noise-corrupted. The experimental protocol and calculation are divided into two steps:​

1. Muscle force static calibration :​ this process provides a relation between the EMG measurements and the corresponding muscle forces, on the basis of the Hill model, enabling to produce a good initial condition of each muscle force time history for the subsequent optimization process.
2. Muscle force dynamical quantification :​ this optimization process improves the scale and offset of the above force time history during motion in order to compute the muscle forces and particularly their distribution between the flexor and extensor sets. The objective of this optimization process is to make correspond the resulting joint net torque, Q_emg, to the corresponding joint net torque, Q_inv, given by inverse dynamics.

This method is presently applied to a benchmark-case :​ the quantification of the flexor and extensor muscle sets of subjects engaged in weightlifting and performing forearm flexion/extension. The results clearly and fortunately show a gradation of the forces according to the weights carried. More fundamentally, an experimental analysis of this muscle effort quantification method was performed with six male and six female subjects carrying five different weights (from 0 to 4 kg) with several flexion/extension frequencies (½, ⅓ and ¼ Hz). This validation shows a good inter-test reproducibility (also showing a fatigue effect) and a very good correlation (correlation factor r = 0.99) between Q_inv and Q_emg at the end of the identification process. Finally, let us point out that our method is based on muscle activity measurements, which is a promising alternative to the methods based on strategies usually used in the literature, e.g. which maximize the endurance.

Finally, we have initiated an extension of this method to the whole human body, enabling to compare the intervertebral torques between a healthy and a scoliotic subject during gait on a treadmill, and also to analyze the forces of muscles that overactuate the knee joint during pedaling on an exercise bike. As a perspective, we suggest a model refinement and an extension of this method, in order to improve the quality of the model and to increase the comprehension of the mechanisms of muscular activation. In parallel, the present applications and database will be extended to complete the present rehabilitation evaluations with internal effort assessments during daily motions, e.g. the assessment of joint efforts of scoliotic patients during gait or the assessment of muscle force of spastic patients during forearm flexion/extension.