Set of q-Markov covariance equivalent models of discrete systems

Abstract
One criterion that has been developed for obtaining reduced-order models is the matching of a finite number of terms in the impulse and autocorrelation sequences. In this paper a model-reduction technique which preserves the first q-Markov parameters and covariances is developed for discrete systems. This method obtains a set of reduced-order models by projection and contains earlier models as a special case.

This publication has 6 references indexed in Scilit: