ANALYSIS OF FETAL ECG SIGNAL EXTRACTION USING THE LEAST MEAN SQUARES ALGORITHM FOR FETAL HEARTBEAT AND HYPOXIA INDEX DETECTION

• Main Author: Fachrezi, Dzaky
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/80683
• Institution: Institut Teknologi Bandung

Fetal hypoxia is a condition in which the fetus does not receive sufficient oxygen supply in the womb, and it potentially increases the risk of fetal death. Monitoring fetal conditions is one strategy to prevent the adverse effects of fetal hypoxia. One evolving and promising method for monitoring fetal conditions is NI-FECG, employing the least mean squares (LMS) adaptive filtering extraction technique. The LMS algorithm utilizes the mother's chest ECG channel to extract fetal ECG from the mother's abdominal ECG. In this project, the LMS algorithm successfully obtained fetal ECG, resulting in an average increase in fetal ECG SNR of 3.43 dB. Subsequently, we performed fetal heartbeat detection, fetal heartbeat variation calculation, and fetal hypoxia index computation as parameters related to the risk of fetal hypoxia. Testing on 10 data from the NIFEADB dataset showed that fetal heartbeat detection yielded satisfactory results with a relative error of 5.1%. However, fetal heartbeat variation calculation and fetal hypoxia index computation did not yield favorable results, with relative errors of 50.3% and 50.1%, respectively. The detection and calculation of these three parameters in the NR class were better than in the ARR class data. Further research is needed to develop algorithms for fetal ECG extraction, fetal peak R detection, and fetal R-R interval validation to achieve more precise and accurate detection of parameters related to the risk of fetal hypoxia in the future.