Kalman filtering algorithm for blind source separation

Author

Lv, Qi ; Zhang, Xian-Da ; Jia, Ying

Author_Institution

Dept. of Autom., Tsinghua Univ., Beijing, China

Volume

5

fYear

2005

fDate

18-23 March 2005

Abstract

The paper presents a Kalman filtering algorithm based on nonlinear principal component analysis (PCA) with prewhitening for blind source separation (BSS), and compares the new algorithm with other algorithms. Simulations show that, for BSS, the Kalman filtering algorithm has a faster convergence rate and a much better tracking capability, compared with the existing natural gradient algorithm for independent component analysis (ICA), the RLS algorithm and the natural gradient based RLS-type algorithm for nonlinear PCA.

Keywords

Kalman filters; blind source separation; gradient methods; independent component analysis; least squares approximations; principal component analysis; Kalman filtering algorithm; RLS algorithm; blind source separation; convergence rate; natural gradient algorithm; nonlinear PCA; nonlinear principal component analysis; prewhitening; tracking capability; Blind source separation; Convergence; Covariance matrix; Filtering algorithms; Independent component analysis; Kalman filters; Principal component analysis; Resonance light scattering; Signal processing algorithms; Source separation;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 2005. Proceedings. (ICASSP '05). IEEE International Conference on

ISSN

1520-6149

Print_ISBN

0-7803-8874-7

Type

conf

DOI

10.1109/ICASSP.2005.1416289

Filename

1416289