Multi-muscle force estimation using data fusion and HD-sEMG: An experimental study

Abstract

This paper presents a new multi-muscle force estimation approach based on a HD-sEMG/force model in experimental context. Accordingly, we will try to estimate the muscle force of the three elbow flexors: Biceps Brachii (BB), Brachialis (BR) and Brachioradialis (BRD) in isometric isotonic conditions. This will be done by using a recently proposed sEMG/force equation model validated by simulation. For representing muscle activation, we will the sEMG Root Mean Squared value acquired after data fusion via Watershed image segmentation algorithm. This data fusion method allows us to obtain one RMS value per force level from the sEMG signals recorded from the HD-sEMG grids putted on each one of the three considered muscles. Five subjects participated in the experimental protocol, where we recorded the force simultaneously with the HD-sEMG signals for 9 contraction levels. After solving a linear equation system, the force of each muscle is estimated. Obtained results shown different muscle activation synergies with the dominance of the BB muscle. Finally, the feasibility of this approach is demonstrated in solving the load sharing problem in isometric isotonic context.