Brief paper: Spectral estimation by least-squares optimization based on rational covariance extension

Authors:
Giovanna Fanizza;Ryozo Nagamune
Affiliations:
Department of Mathematics, Division of Optimization and System Theory, Royal Institute of Technology, SE 100 44 Stockholm, Sweden;Department of Mechanical Engineering, The University of British Columbia, Vancouver, BC, Canada V6T 1Z4
Venue:
Automatica (Journal of IFAC)
Year:
2007

Citing 9
Cited 1

System identification: theory for the user

System identification: theory for the user
Canonical correlation analysis, approximate covariance extension, and identification of stationary time series

Automatica (Journal of IFAC)
Model Reduction for Control System Design

Model Reduction for Control System Design
On the Duality between Filtering and Nevanlinna--Pick Interpolation

SIAM Journal on Control and Optimization
From Finite Covariance Windows to Modeling Filters: A Convex Optimization Approach

SIAM Review
A new approach to spectral estimation: a tunable high-resolutionspectral estimator

IEEE Transactions on Signal Processing
Cepstral coefficients, covariance lags, and pole-zero models forfinite data strings

IEEE Transactions on Signal Processing
A covariance extension approach to identification of time series

Automatica (Journal of IFAC)
Kullback-Leibler approximation of spectral density functions

IEEE Transactions on Information Theory

A genetic algorithm approach to the spectral estimation of time series with noise and missed observations

Information Sciences: an International Journal

Quantified Score

Hi-index	22.16

Visualization

Abstract

This paper proposes a new spectral estimation technique based on rational covariance extension with degree constraint. The technique finds a rational spectral density function that approximates given spectral density data under constraint on a covariance sequence. Spectral density approximation problems are formulated as nonconvex optimization problems with respect to a Schur polynomial. To formulate the approximation problems, the least-squares sum is considered as a distance. Properties of optimization problems and numerical algorithms to solve them are explained. Numerical examples illustrate how the methods discussed in this paper are useful in stochastic model reduction and stochastic process modeling.