An ICA approach to semi‐blind identification of strictly proper systems based on interactor polynomial matrix
Jani Even, Kenji Sugimoto · International Journal of Robust and Nonlinear Control · 2006
Abstract This paper presents an identification method for strictly proper systems with input signals partially unknown but mutually independent. Due to recent progress in independent component analysis (ICA), it has been feasible to identifybipropersystems with unknown inputs by resorting to a finite impulse response (FIR) approximation. By means of an interactor polynomial matrix, the paper generalizes the existing method to strictly proper systems. The method is further generalized to the case where some input components are known, which is often the case in control engineering. Copyright © 2006 John Wiley & Sons, Ltd.