Multiple Sub-Filter Adaptive Noise Canceller for Fetal ECG Extraction

Non-invasive monitoring of fetal cardiac signs through the signal processing of maternal abdominal signals has been emerging as a promising technique in the field of obstetrics and gynaecology. The underlying challenges in obtaining the Fetal Electrocardiogram (FECG) from maternal abdominal signals include elimination of dominant maternal ECG and other noises. The simple technique based on adaptive noise cancellers uses Single Long Filter (SLF) as adaptive filter for extracting the FECG. In this paper, we propose an algorithm for implementation of adaptive filter in adaptive noise canceller (ANC) as Multiple Sub-Filters (MSF). Each of the adaptive sub-filters are updated using the Least Mean Square (LMS) algorithm. The proposed MSF architecture is compared with traditional SLF ANC by testing using a real data from Daisy database. From the obtained results, it is observed that MSF-ANC performs better than SLF-ANC with 91.66% positive predictive value and 84.61% accuracy. (C) 2019 The Authors. Published by Elsevier B.V.