基于EMD自相关的表面肌电信号消噪方法  

De-noising method of the s EMG based on EMD autocorrelation

作　　者：席旭刚[1] 武昊[1] 罗志增[1] 

机构地区：[1]杭州电子科技大学智能控制与机器人研究所,杭州3100118

出　　处：《仪器仪表学报》2014年第11期2494-2500,共7页Chinese Journal of Scientific Instrument

基　　金：国家自然科学基金(60903084;61172134);浙江省自然科学基金(LY13F030017);浙江省科技计划(2014C33105)项目资助

摘　　要：为了更好地消除混杂在表面肌电信号(s EMG)中的噪声,提出一种基于噪声统计特性的EMD自相关消噪方法。首先,对含噪s EMG信号进行EMD分解,并根据噪声统计特性降低低信噪比的高频IMF分量的能量后重组信号。其次,对重组后的信号进行自相关函数特性的EMD分解,并对自相关函数方差低于阈值的高频IMF分量进行小波去噪。最后,把处理后的高频IMF分量和低频IMF分量重构,得到的信号即为消噪信号。实验结果表明,该方法不仅能更好的消除噪声,而且在低信噪比情况下有良好表现。A de-nosing method based on the noise statistical EMD autocorrelation is proposed to eliminate the noise mixed with s EMG. Firstly, the s EMG signals are recombined through EMD based on energy reduction of low SNR high-frequency IMF according to the noise statistics. Then, the recombined signals are decomposed by EMD, and high-frequency IMFs whose autocorrelation function variance is below the threshold are denoised by the wavelet. Finally, the processed high-frequency IMF and low-frequency IMF are reconstructed to get the signal denoised. The experimental results show that the method not only can well eliminate noise, but also a good performance is achieved under the condition of low signal-to-noise ratio.

关 键 词：表面肌电信号 经验模态分解 小波变换 自相关函数 噪声统计特性 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置] TH7[自动化与计算机技术—控制科学与工程]