A reliable evaluation of muscle forces in the human body is highly desirable for several applications in both clinical and research contexts, especially for the follow-up of musculoskeletal pathologies in rehabilitation. Several models of muscle force distribution based on non-invasive measurements have been proposed since 1836, amongst which
Crowninshield et Brand [1]’s (1981), which maximizes a cost-function representing the muscle fiber endurance, is the most popular. It is worth noting that this model is the most widely adopted notwithstanding its major limitations of physiological coherence.
Forster et al. [2] (2004) pointed out that these conventional optimization criteria are inadequate in predicting muscle co-contraction, and proposed an improved model to deal with this problem. Moreover, electromyographic (EMG)-driven models have been proposed to assess individual muscle forces, which are based directly on the measured EMG patterns. However, this approach has not been broadly adopted, because of its complexity and the necessity of calibration before each test. Nevertheless, the EMG-driven approach could lead to the identification of more advanced cost-functions, which would be more in line with the muscle physiological activations compared to the EMG-free cost-functions, and easier to use than the evaluation using directly the EMG signals. The objective of this paper is to propose the first cost-function combining kinematic and EMG data for the quantification of muscle forces during movement. The muscle force prediction of our method performs 18.8% higher coherence with the EMG solution than the prediction of Crowninshield’s method when tested on a database of 17 subjects.