17th International Symposium on
Mathematical Theory of Networks and Systems
Kyoto International Conference Hall, Kyoto, Japan, July 24-28, 2006

MTNS 2006 Paper Abstract

Close

Paper TuP09.2

Enqvist, Per (Royal Inst. of Tech.)

On Minimal MIMO Realization of Covariance and Markov Parameters.

Scheduled for presentation during the Regular Session "Linear System Identification II" (TuP09), Tuesday, July 25, 2006, 15:45−16:10, Room J

17th International Symposium on Mathematical Theory of Networks and Systems, July 24-28, 2006, Kyoto, Japan

This information is tentative and subject to change. Compiled on June 23, 2024

Keywords Linear system identification, Convex optimization, Filtering and estimation

Abstract

We study design of filters from covariances and Markov parameters in the multivariable setting where the input noise variance is considered as a design variable. This makes this approach different from the classical Markov COVER approach and allows realizations of lower degree. In the scalar case a unique minimal solution always exists, but in the MIMO case this is not true and the problem is much more complex. A difficult rank minimization problem is encountered and a heuristic method for solving it and obtaining a minimal realization is proposed.