Spectral estimation via adaptive filterbank methods: a unified analysis and a new algorithm

Authors:
Erik G. Larsson;Petre Stoica;Jian Li
Affiliations:
Department of Electrical and Computer Engineering, University of Florida, P.O. Box 116130 458 ENG Bldg. 33, Gainesville, FL;Department of Systems and Control, Uppsala University, Uppsala, Sweden;Department of Electrical and Computer Engineering, University of Florida, P.O. Box 116130 458 ENG Bldg. 33, Gainesville, FL
Venue:
Signal Processing - Signal processing with heavy-tailed models
Year:
2002

Citing 8
Cited 0

Matrix analysis

Matrix analysis
Digital spectral analysis: with applications

Digital spectral analysis: with applications
Matched-filter bank interpretation of some spectral estimators

Signal Processing
Forward-only and forward-backward sample convariances — a comparative study

Signal Processing
Fast Implementation of Two-Dimensional APES and CAPON Spectral Estimators

Multidimensional Systems and Signal Processing
An adaptive filtering approach to spectral estimation and SARimaging

IEEE Transactions on Signal Processing
Generalized quadratic minimization and blind multichanneldeconvolution

IEEE Transactions on Signal Processing
MIMO system identification: state-space and subspace approximationsversus transfer function and instrumental variables

IEEE Transactions on Signal Processing

Quantified Score

Hi-index	0.00

Visualization

Abstract

The problem of estimating the amplitude spectrum of a signal is of interest in a number of applications ranging from radar imaging to time-series analysis. The so-called adaptive filterbank-based nonparametric spectral estimators have recently received renewed interest as potential solutions to this problem. In essence, the adaptive filterbank methods determine an estimate of the spectrum for a frequency of interest by computing a finite impulse response filter according to a certain criterion and fitting a sinusoid to the filtered data sequence. In this paper, we first analyze the asymptotic estimation accuracy of the amplitude spectrum for various filterbank estimators. Next, we propose a new adaptive filterbank estimator based on a minimum mean square error criterion. Numerical examples indicate that the new estimator can have a better resolution capability than previously known filterbank estimators.