微型机与应用Issue(9)：92-94,97,4.
一种基于广义特征值最大化的胎儿心电盲提取算法
A fetal electrocardiogram blind extraction algorithm based on generalized eigenvalue maximization
摘要
Abstract
Extracted fetal electrocardiogram (FECG)from the maternal abdominal signal is a complex signal processing problems, obtained the clear FECG is a great significance for perinatal fetal monitoring. FECG has periodic time structure , according to the autocorrelations of FECG, and statistical independence between FECG and other sources, this paper minimize the ratio between the varianceof the difference periodic and expected to extract the signal , and established theobjective function of blind extraction under the condition of noises.Then derived the FECG extraction method of maximizing generalized eigenvalue. The algorithm calculates the autocorrelation matrix with a dynamic delay, it does not need to be iterative calculation but calculate generalized eigenvalue and eigenvector, algorithm is simple. Simulation experiments show that the algorithm can extract the clear FECG , and the performance is stable.关键词
盲提取/胎儿心电/广义特征值/自相关矩阵/动态延迟Key words
blind source extraction/fetal electrocardiogram/generalized eigenvalue/autocorrelation matrix/dynamic delay分类
信息技术与安全科学引用本文复制引用
苏博妮,谢宏,化希耀..一种基于广义特征值最大化的胎儿心电盲提取算法[J].微型机与应用,2014,(9):92-94,97,4.基金项目
国家自然科学基金（61162018）；上海科学技术委员会项目支持（12441903500）；中国交通部应用基础研究项目（2012319810190）；上海海事大学研究生创新基金 （）
