Extracting Fetal ECG Based On Adaptive Filtering

Fetal electrocardiogram (FECG) is the common one of physiological signal. It has much clinical significance in diagnosis of fetal oxygen deficit in womb, fetal cardiopathy and so on. But the FECG acquired from the surface of mother body often contains many noises such as mother electrocardiogram (MECG), power line interference (50Hz) and baseline wander. This article takes fetal electrocardiogram signal as research object and uses the adaptive filtering algorithm to extract FECG from the complex noise.The article first introduces research background and its sense, as well as international research condition of extracting FECE based adaptive filtering. The paper also researches the characteristic of FECG signal and the noises.After analyzing the characteristic of FECG signal and the noises, the noises is filtered in view of the characteristic. The article first discusses the widely used Least Mean Square (LMS) adaptive filtering algorithm, and discusses the Recursive Least Square (RLS) adaptive filtering algorithm because of slow convergence rate of LMS algorithm. Although the RLS algorithm has good qualities of convergence rate and filter effect, the RLS algorithm is very complexity (RLS:M~2, LMS:2M, M is the step of filter). So this thesis uses the Fast Transversal Filter (FTF) adaptive algorithm. It has good qualities of convergence rate and small computation quantity (FTF: 8M, RLS:M~2, M is the step of filter). Further more the filter effect of FTF adaptive algorithm is better than RLS algorithm.Having done the simulation research on filtering and extracting FECG signal using the computer and Matlab software, and discussed the simulation result. Finally carries on the FTF algorithm to the measured data. It can be seen that the MECG is well restrained from the result. The satisfying result is obtained.At last, research results are summarized and expectation is provided.