Non-Invasive Extraction of Fetal and Maternal ECG Signals using Adaptive Filter for Pregnancy Monitoring System

This paper proposes an innovative way for the successful extraction of fetal ECG from abdominal mixed ECG of the mother. This investigation was carried out using the adaptive filter algorithm. For this algorithm, data was taken from an online database called DAISY Database. A total of eight ECG signals have been taken out of which three ECG signals were maternal Thoracic ECG signals and five were abdominal ECG signals. These ECG signals were fed into an adaptive filter algorithm that has successfully suppressed the overlapping maternal ECGs from the abdominal mixed ECGs and has successfully been able to obtain the fetal ECG. Three kinds of adaptive filter algorithms have been used in two different systems namely SISO (Single input single output) and MISO (Multiple input single output) systems. Fetal heart rate has then been acquired from the extracted fetal ECG. The heart rate of the mother has been obtained from the maternal thoracic ECG. Performance analysis has been done by comparing the performance parameters like cross correlation and magnitude squared coherence of the three algorithms of both SISO and MISO systems.