Modified Herault–Jutten Algorithms for Blind Separation of Sources

Andrzej Cichocki , Robert E. Bogner , Leszek Moszczyński , Kenneth Pope

Abstract

We present several modifications of blind separation adaptive algorithms which have significant advantages over the well-known Herault–Jutten learning algorithm in handling ill-conditioned signals. In particular, the proposed algorithms are more stable and converge to the correct solutions in cases where previous algorithms did not. The problem is the classical one in which several independent source signalssj(t) (j = 1, 2,…, n) are linearly combined via unknown mixing coefficients (parameters)aijto form observationsxi(t) = ∑j=1naijsj(t),i= 1, 2, …,n. The synaptic weightswijof a linear system (often referred to as a single-layer feedforward neural network) must be adapted to combine the observationsxi(t) to form optimal estimates of the source signalsŝp(t) =yp(t) = ∑i=1nwpixi(t). The optimal weights correspond to the statistical independence of the output signalsyp(t) and they simultaneously ensure self-normalization of these signals. Starting from the modified Herault–Jutten recursive neural network model, we have derived a family of on-line adaptive learning algorithms for feedback (fully recurrent) and feedforward architectures. The validity and high performance of the proposed neural network are illustrated by simulation.
Author Andrzej Cichocki (FoEE / ITEEMIS)
Andrzej Cichocki,,
- The Institute of the Theory of Electrical Engineering, Measurement and Information Systems
, Robert E. Bogner
Robert E. Bogner,,
-
, Leszek Moszczyński (FPE / IoMP)
Leszek Moszczyński,,
- The Institute of Manufacturing Processes
, Kenneth Pope
Kenneth Pope,,
-
Journal seriesDigital Signal Processing, ISSN 1051-2004
Issue year1997
Vol7
No2
Pages80-93
DOIDOI:10.1006/dspr.1997.0281
URL http://www.sciencedirect.com/science/article/pii/S1051200497902815
Score (nominal)30
Publication indicators WoS Impact Factor: 2006 = 0.889 (2) - 2007=0.986 (5)
Citation count*49 (2015-06-01)
Cite
Share Share

Get link to the record


* presented citation count is obtained through Internet information analysis and it is close to the number calculated by the Publish or Perish system.
Back
Confirmation
Are you sure?