Theory And Application Of Bayesian-based Blind Signal Separation

Blind signal sepearation (BSS) or blind source separation is a technology which can separate souce signals from unkown mixed systems and unknown circumstances. BSS is a research hotspot in signal processing. It has been widely used in many fields. One of its applications on fetal ECG extraction is described in this paper..Firstly, this paper introduces several basic methods about BSS. Secondly, it introduces the Bayesian theory. Finally, it puts these knowledgements together for extracting fetal ECG signal, which is a typical application on BSS. The main coutributions of this paper are:1)Wavelet denoising combined with neural network is used for the extraction of fetal ECG signal. Firstly, it extracts denoised ECG signals from a pregnant by using wavelet denoising method. Then it uses neural network to approximate the nonlinear fuction. Lastly, it extracts fetal ECG signal successfully. Experiments show that theory of BSS combining with wavelet denoising and neural networks is an effective method when extracting fetal ECG signal from a pregnant.2)Bayesian theory is used for the extraction of fetal ECG signal. Firstly, it extracts ECG sinals synchronization from the pregnant woman's chest and abdomen. Then it uses Bayesian inference to estimate the fetal ECG signal.Experiments show that by using Bayesian theory for extracting fetal ECG has broad application prospects.