Research Of Noninvasive Fecg Extraction Methodand The Application On DSP System

With the development of science and technology, in pregnant time the fetal electrocardiogram real-time wardship has been paid to more and more great attention. Analyzing the FECG signals and clinical diagnosis are helpful for the doctor knowing the growth situation of the fetal. Therefore it is very important of capturing the clear FECG. . Due to the weakness of FECG and the effect of various noises, catching the FECG has become a big problem for signal process the academic circle. Recently more and more international experts start using the ICA technology to solve that problem.My study comprises two parts. One is a new method of noninvasive FECG extraction; another is DSP system that is designed to extract the fetal electrocardiogram.My paper introduces the basic framework, the basic principle and correlative knowledge of ICA. It also presents the fast fixed-point algorithm using negentropy in detail. Comparing the tradition ICA arithmetic that is lack of extract FECG, my paper discusses a new method of noninvasive fetal electrocardiogram (FECG) extraction. The new method uses the FIR high filter to deal with the signals on the abdomens of pregnant women; it is debase the non-linear relativity of the signals. Then it uses a traditional FastICA algorithm based to maximum negentropy to extract the FECG. Simulation experiments show that our method of extracting the FECG is improved, compared with the direct traditional ICA arithmetic. The advantage of my method is less noise and has more effect on extracting FECG.The DSP system, using floating-point processor TMS320VC33 of TI as the kernel, is designed to sample and extract the fetal electrocardiogram from abdomen of pregnant woman. Sampling, extract and transportation of FECG signals are accomplished by embedded system programming. The extract result of the DSP real-time embedded system is the same performance of simulation.