In this work, we deal with the elimination of artifacts (electrodes, muscle
, respiration, etc.) from the electrocardiographic (ECG) signal. We use a n
ew tool called independent component analysis (ICA) that blindly separates
mixed statistically independent signals. ICA can separate the signal from t
he interference, even if both overlap in frequency. In order to estimate th
e mixing parameters in real time, we propose a self-adaptive step-size, der
ived from the study of the averaged behavior of those parameters, and a! tw
o-layers neural network. Simulations were carried out to show the performan
ce of the algorithm using a standard ECG database. (C) 1998 Elsevier Scienc
e B.V. All rights reserved.