Weighted least-squares blind deconvolution

Author

Fiori, Simone ; Piazza, Francesco

Author_Institution

Dipt. di Elettronica e Autom., Ancona Univ., Italy

Volume

5

fYear

1999

fDate

1999

Firstpage

2507

Abstract

This paper presents a new cost function for blind deconvolution of non-minimum phase systems. The proposed criterion arises as a natural consequence of a fundamental theorem proved by Benveniste, Goursat and Ruget (1980), and appears to be the weighted square of the difference among two spectra, thus its minimization leads to a weighted least-squares blind deconvolution technique. In order to assess the new theory some simulation results both on ideal (noiseless) and noisy channels are presented

Keywords

blind equalisers; deconvolution; filtering theory; least squares approximations; spectral analysis; telecommunication channels; blind equalization; characteristic functions; cost function; equalizing filter output; ideal noiseless channels; noisy channels; nonminimum phase systems; simulation results; source signal; spectra; weighted least-squares blind deconvolution; Additive noise; Cost function; Crosstalk; Deconvolution; Delay; Digital systems; Finite impulse response filter; Nonlinear filters; Sampling methods; Transversal filters;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1999. Proceedings., 1999 IEEE International Conference on

Conference_Location

Phoenix, AZ

ISSN

1520-6149

Print_ISBN

0-7803-5041-3

Type

conf

DOI

10.1109/ICASSP.1999.760640

Filename

760640