基于EEMD与SVD相结合的弱信号提取方法  

Method for Extraction of Weak Signal Based on EEMD and SVD

作　　者：殷德龙[1] 段志善[1] 郭宝良[1] 王苗[1] 

机构地区：[1]西安建筑科技大学机电工程学院,西安710055

出　　处：《煤矿机械》2016年第2期154-156,共3页Coal Mine Machinery

基　　金：国家自然科学基金青年科学基金项目(51105292);教育部博士点基金项目(20126120110009);陕西省科技攻关项目(2013K07-09);陕西省教育厅专项基金项目(2013JK1032)

摘　　要：提出了基于集合经验模态分解(EEMD)和奇异值(SVD)相结合的微弱信号提取方法和高低频部分的判别准则:采用EEMD把信号分解成几部分,将IMF分为高频段与低频段2部分,对2部分分别进行奇异值处理,叠加得到降噪信号,做出其频谱图,得到所需要的有用信号。此方法可以在未知原信号的情况下提取,并且可以提取信噪比为-15 dB的信号。仿真结果和对比分析表明,此方法能更好地提取微弱特征信号。A method for the extraction of weak signal based on EEMD and 5VD are proposed, lhe signals are decomposed into the IMFs by EEMD ,Secondly, IMF component are divided in two partshigh frequency and low frequency. Thirdly two part of signal are de-noised by SVD. Adding two part of signals to get denoised signal and its frequency spectrum. The useful signals can be got. This method can extract the signal that SNR is -15 dB under the unknown primary signal. Simulation results and comparative analysis show the weak feature signal can be extracted by this method.

关 键 词：强噪声 微弱信号 集合经验模态分解 奇异值 

分 类 号：O235[理学—运筹学与控制论]