On the simultaneous realization problem - Markov-parameter and covariance interpolation

• An efficient algorithm for determining the unique minimal and stable realization of a window of Markov parameters and covariances is derived. The main difference compared to the Q-Markov COVER theory is that here we let the variance of the input noise be a variable, thus avoiding a certain data consistency criterion. First, it is shown that maximizing the input variance of the realization over all interpolants yields a minimal degree solution-a result closely related to maximum entropy. Secondly, the state space approach of the Q-Markov COVER theory is used for analyzing the stability and structure of the realization by straightforward application of familiar realization theory concepts, in particular the occurrence of singular spectral measures is characterized.

Ämnesord

TEKNIK OCH TEKNOLOGIER  -- Elektroteknik och elektronik -- Signalbehandling (hsv//swe)

ENGINEERING AND TECHNOLOGY  -- Electrical Engineering, Electronic Engineering, Information Engineering -- Signal Processing (hsv//eng)

Nyckelord

realization theory

moment interpolation problems

covariances

Markov

parameters

Q-Markov COVERs

entropy

duality theory

auto-correlation sequences

equivalent realizations

linear-systems

identification

approximation

Publikations- och innehållstyp

ref (ämneskategori)

art (ämneskategori)