The Design Of Fetal ECG Extraction System Based On Blind Source Separation

With the development of society, people pay more and more attention on fetal monitoring.As a kind of physiological signal which reflects the fetal cardiac activity, Fetal Electrocardio (FECG) has become a hot topic in the field of fetal monitoring.There are two difficulties to obtain FECG. The first one is the acquisition of FECG. Since FECG is a very weak signal it is affected by a lot of interference when collecting from maternal abdomen, such as the mother ECG, frequency interference, EMG interference, baseline drift. The second difficulty is the extraction of FECG. As FECG has small amplitude it is always lost in the maternal ECG. In current research, it is a major problem to extract the FECG from the maternal ECG. To solve these two difficulties, this paper designs a system to collect FECG and extract the FECG. In this paper, we are designed an embedded hardware circuit to collect the FECG and BSS algorithms which is very popular in recent years is applied to extract the fetal ECG.The main works of this paper are as follows:(1).Design an embedded system for acquiring FECG signals from the maternal body. The system uses embedded STM32F103 series ARM chip. This chip has 6410 ports, and has fast processing speed while has low power consumption. The A/D converter of this chip uses professional bioelectrical processing chip ADS 1294 which has high conversion efficiency. When designing the circuit, filtering technology is used to filter the FECG noise signal, such as frequency interference and baseline drift. The processor UART port and SPI port are fully utilized to transmit the data. We use a Bluetooth module to communicate between the embedded system and the PC. The short-distance wireless transmission is very convenient in data acquisition.(2).Implement a blind source separation algorithm which based on oblique projection,for extraction FECG in the MFC framework. We can extract clean FECG from mixed maternal abdominal signal by this algorithm. Since the algorithm requires the location of the maternal and FECG QRS,so in this paper we implement two algorithms for locating mother ECG QRS and FECG QRS respectively.(3).Design an interactive interface which can control the operation of the embedded system as well as can increase the visibility. People can see the real-time ECG data which is collecting and the FECG data has been separated.