Filtering based multi-innovation stochastic gradient identification algorithm for multivariable nonlinear equation-error autoregressive systems

This paper considers the parameter estimation problem of the multivariable Hammerstein systems with colored noise. By means of the adaptive filtering technique, a new adaptive filtering configuration consisting of a noise whitening filter and a recursive algorithm is developed. Then an adaptive filtering based multi-innovation stochastic gradient (F-MISG) identification algorithm is presented. The simulation results show that the proposed F-MISG algorithm has a higher parameter estimation accuracy and a faster convergence rate than the multi-innovation stochastic gradient algorithm for the same innovation length.