PHYSIOLOGICAL AND BIOMECHANICAL APPROACH FOR HUMAN FINGER MOVEMENT: MODELING, SIMULATION AND EXPERIMENTAL VALIDATION

The work presented in this paper deals with the description of an analytic modeling of the neuromusculoskeletal system responsible for the finger movement. This simulation task is complex due to the interacting processes (physiological and biomechanical) represented by muscles, joints and bones. In this study, we focused on the presentation of a complete model for the finger motion decomposed in quasi-static positions. In fact, this model can be considered as a preliminary step before dynamic modeling. The proposed model is composed of several compartments: biomechanical finger model, mechanical muscle model and muscle/neural activation model. The main objective of this study is to estimate, by inverse procedure, the muscle forces, muscle activations and neural activations that are responsible for generating a given finger movement decomposed in successive quasi-static positions. The anatomical model contains six muscles which control the decomposed movement of the three joints of the system. To estimate the model unknowns, an optimization technique is proposed for improving robustness to initial conditions and physiological reliability. After, an experimental protocol for recording surface electromyogram (sEMG) data, from three extrinsic muscles, according to specific finger positions is applied on five subjects to evaluate the model reliability. From analysis of the obtained results, both in simulation and experiment, the presented model seems to be able to mimic, in a realistic way, the finger movement decomposed in quasistatic positions. Finally, results, model limitations and further developments are discussed.