Extraction of Fetal Electrocardiogram from Maternal Skin Electrodes using Affine Projection Algorithm (APA), Recursive Least Square (RLS), and QR-RLS Algorithms

The abdominal ECG is a composite signal because it contains both the mother's and the fetus' ECG signals, but the chest ECG is supposed to have only maternal (MECG) component. In this paper, we use three different adaptive algorithms to remove FECG signal from the abdominal ECG. All procedures are applied to real multichannel ECG recordings obtained from a pregnant woman. We explain and compare the use of multichannel APA, RLS, and QR-RLS in time-domain for fetal electrocardiogram (FECG) extraction from four ECG signals recorded at the abdominal and chest areas of mother's skin. The results show that the multichannel QR-RLS is more effective than multichannel APA and RLS in this important biomedical application.