A Method for Identification of Electrically Stimulated Muscle

We present a model structure and a method for identifying the dynamics of electrically stimulated muscle. The model structure is sufficiently rich to describe a wide set of muscle behavior. It consists of (i) an input static nonlinearity representing the muscle's recruitment properties, (ii) a linear dynamical system representing the contraction dynamics, (iii) an output static nonlinearity representing generalized force-length and force-velocity relationships, and (iv) prefilters for the mechanical input that capture impedance and history dependence properties of the muscle. It is assumed that each of the subsystems is linearly parameterized. We present parameter estimation methods, and verify via simulation successful convergence of the estimates to their true values with small variances