Recursive Identification of Hammerstein Systems with application to Electrically Stimulated Muscle

Download

PDF Download (769Kb)

Description/Abstract

Two methods for recursive identification of Hammerstein systems are considered. In the first method, the recursive least squares algorithm is applied to an overparameterized representation of the Hammerstein model and a rank-1 approximation is used to recover the linear and nonlinear parameters from the estimated overparameterized form. In the second method, the linear and nonlinear parameters are recursively estimated in an alternate manner. The superiority of the second method is confirmed using a numerical simulation example, together with experimentally measured data from electrically stimulated muscles.