Fetal Electrocardiogram (FECG) Acquisition And Extractionalgorithm Based On Blind Source Separation Theory

The field of electrocardiography (ECG) has been in existence for over a century, yet despitesignificant advances in adult clinical ECG, the analysis of fetal Electrocardiogram (fECG) isstill in its infancy. fECG signal contains potentially precise information on the health of fetusand could assist clinicians making more appropriate and timely decisions during gestation andlabor. The object of the subject is to get the pure fECG signal for clinical analysis whichmainly involves two aspects: first, analog front-end circuit designs which mainly refer to therelatively mature design of adult ECG, aim to amplify the abdominal surface signal linearlywith no distortion as much as possible. Second, the robust and precise fECG extractionalgorithm, which can not only effectively cancel the mECG and other clutters, but also notdamage every fECG waveforms. The extraction of fECG can be solved by the booming BlindSource Separation (BSS) theory, combined by statistical signal analysis and adaptive signalprocessing techniques, we give some effective Blind Source extraction(BSE) algorithms.Based on the contents of the two aspects, the innovations we have made are listed as follow1. Introduce the structure of fetal heart, and the generation, propagation and mark system offECG, reveal the inner relationship between fECG waveform and heart illness.2. Raise an effect electrode lead position scheme, model the contact between electrode andskin for simulation verification of every circuit module, especially the offset voltage,noise and feedback stability. Raise a low gain high resolution ADC scheme which greatlysimplify the design and be verified valid by the real clinical ECG collection.3. Based on the specificality of ECG: Quasi-Periodic, linear temporal mixed model andstatistically unrelated, we exploit the cyclicity of ECG signal autocorrelation and present aBSE way, two real-time and on-line processing objective functions are constructed.Meanwhile, we observe the change of the fECG period, and present a way to find varyingperiods in frequency field, combined by a new objective function based on oblique project,we obtain a more excellent fECG BSE algorithm.