DocumentCode :
2164743
Title :
Neural network schemes for blind separation of sources from nonlinear mixtures
Author :
Woo, W.L. ; Sali, S.
Author_Institution :
Newcastle upon Tyne Univ., UK
Volume :
2
fYear :
2002
fDate :
2002
Firstpage :
1227
Abstract :
Most existing BSS algorithms are based on the ideal situation where the mixture is merely a linear transformation of the source signals and the demixer is simply a linear network. Nonlinear techniques are presented for instantaneous blind signal separation using an information theoretic approach combined with (nonlinear) neural networks. Firstly, we address the issue of modelling the mixture for both linear and nonlinear transformation of the source signals. Secondly, we derived the required algorithm to train the variable gradient multilayer perceptron (MLP) based on a Lie group. In the past, most demixers employed a fixed gradient. Finally, computer simulations are carried out to compare the performance of the linear and nonlinear demixer when the underlying mixture of the source signals is either linear or nonlinear.
Keywords :
Lie groups; blind source separation; gradient methods; information theory; multilayer perceptrons; nonlinear systems; Lie group; blind signal separation; blind source separation; demixer; information theoretic approach; nonlinear mixtures; nonlinear neural networks; variable gradient multilayer perceptron; Cost function; Entropy; Independent component analysis; Mathematical model; Neural networks; Satellites; Signal processing; Speech; Transponders; Wrapping;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Digital Signal Processing, 2002. DSP 2002. 2002 14th International Conference on
Print_ISBN :
0-7803-7503-3
Type :
conf
DOI :
10.1109/ICDSP.2002.1028315
Filename :
1028315
Link To Document :
بازگشت