N-step predictive algorithm based on fractional multimodel for rat muscle

Peroneus digiti quarti rat muscle responses are studied for different pattern stimulations. The multimodel structure presented in previous studies (L. Sommacal et al., 2006) is performed to take into account dynamic and amplitude variations, which depend not only on fatigue but also on amplitude. The multimodel includes fractional sub-models. Fractional orders used in the sub-models lead to transfer functions with few parameters. This multimodel structure allows distinguishing contraction and relaxation phases. This study presents the identification results obtained with the final multimodel structure, for IIA, IIB and I fibres at 34degC. The validity and reliability are tested for different motor units (MU) stimulation types (low/high excitation frequency, long experimental durations)