Identification of scalar errors-in-variables models with dynamics

Abstract

The paper considers the task of identifying a causal linear dynamic system excited by stationary Gaussian zero mean noise of unknown spectrum, and given measurements of the system input and output contaminated by independent additive stationary noise signals of unknown spectra. While the solution is normally not unique, finite-dimensional parameterizations of the solution set are given (in some cases involving just one parameter), even though the various spectra may not be rational. Show more