A fixed-point algorithm for independent component analysis which uses a priori information

Independent component analysis is a powerful tool for separating signals from their mixtures. In this field, many algorithms have been proposed, but they poorly use a priori information in order to find the desired signal. Besides, they provide many outputs, from which we have to choose the one of interest. Here we propose a fixed point algorithm which uses a reference input to find the signal of interest. We applied the algorithm to electrocardiographic (ECG) interference cancellation. In simulations, the algorithm successfully found the desired component even if it was spectrally overlapped by the interference signals. Moreover, the algorithm was applied to an actual situation consisting of an eight channel ECG obtained from a pregnant woman. As a result, the algorithm could either obtain the ECG signal from the baby or from the mother by just changing one parameter: the fundamental frequency of the desired ECG.