一种用于远程监护中心电信号检测的新方法  

A new method of electrocardiograph signals detection in remote monitoring

作　　者：潘静[1] 郭兴明[1] 陈旻[1] 谭新[1] 

机构地区：[1]重庆大学生物工程学院,重庆市400030

出　　处：《中国组织工程研究与临床康复》2008年第9期1713-1715,共3页Journal of Clinical Rehabilitative Tissue Engineering Research

基　　金：国家教育部春晖计划项目资助(Z2004-1-55006)~~

摘　　要：目的:远程心电监护过程中如何降低各种噪声干扰,将心电信号快速准确表达是实验拟解决的问题。方法:将人工神经网络和自适应噪声抵消原理相结合,用一个三层BP网络来代替自适应抵消中常用的线性滤波器,并根据实际情况改进其获得参考输入的方法。通过Matlab/Simulink进行建模仿真,运用MIT/BIH数据库中的数据验证该方法的有效性。结果:①对用余弦波模拟的50Hz工频干扰的滤除有显著作用,并且可以很好地保留原始信号的波形特征。②对MIT/BIH数据库中含噪声较严重的108号数据进行滤波,能够有效的消除基线漂移和其他因素引起的干扰,提高信号的信噪比。结论:由于神经网络具有自学习和非线性映射能力,该方法能够比一般的自适应滤波更好的适应噪声的非线性特性。在远程心电监护中,可以有效滤除运动和环境因素引起的各种干扰,效果满意。AIM： To investigate efficient method of decreasing noise disturbance in remote electrocardiograph （ECG） monitoring and detect the ECG signals exactly. METHODS： The method was based on both elements of Artificial Neural Network （ANN） and self-adaptive noise cancellation. A three-layer BP network was built to replace the linear filter of a self-adaptive noise cancellation and the structure of the referential input was also changed. Emulated by Matlab/Simulink, the validity for ECG was detected by MIT/BIH data. RESULTS： （1）For the signal which had 50 Hz disturbance simulated by cosine waves, the method was very efficient for decreasing the noise and reserving the characteristics of the original signals.（2）For the No. 108 signal form MIT/BIH, which had heavily noise, it was good for removing the base-line excursion and other interference, so as to increase signal-to-noise ratio. CONCLUSION： Since the ANN has good ability of self-study and nonlinear mapping, the self-adaptive noise cancellation based on ANN can be very available for adapting the non-linear characteristic of the noise and satisfactory for reducing the influence induced by exercise and environment in remote ECG monitoring.

关 键 词：远程监护 心电信号检测 神经网络 自适应噪声抵消 非线性 医学工程 

分 类 号：R318[医药卫生—生物医学工程]