基于ICA-CEEMD小波阈值的传感器信号去噪

发布时间：2018-03-01 07:40

本文关键词： ICA-CEEMD 小波阈值传感器去噪 σ原则　出处：《振动与冲击》2017年04期 　论文类型：期刊论文

【摘要】：针对传感器在采集信号时混入不同的噪声,提出一种基于ICA-CEEMD小波阈值的组合去噪算法。该方法是对一维含噪信号进行剪切分段、平移和拼接,得到几个不同的含噪信号作为独立分量分析(ICA)的输入通道信号。通过ICA的盲源分离技术使得信号和噪声进行初步分离。再利用互补集合经验模态分解(CEEMD)对分离信号进行分解去噪,由于不同的高频和低频噪声,需要对分解的高阶和低阶固有模态函数(IMF)进行处理。对第一层和最后一层IMF利用3σ原则提取细节信息,进一步抑制模态混叠影响,重构去噪信号。最后,利用小波阈值对重构信号做去噪处理,提升去噪效果和性能指标。为验证该方法的有效性,进行了仿真和中北大学汾机实测实验,结果表明,该方法在去噪效果和性能指标上都优于小波软阈值去噪和基于CEEMD的小波阈值去噪方法,是一种有效的信号去噪新方法。
[Abstract]:This paper presents a combined denoising algorithm based on ICA-CEEMD wavelet threshold, which is used to cut, translate and splice one-dimensional noisy signals. Several different noise-containing signals are obtained as input channel signals for independent component analysis (ICA). The signal and noise are preliminarily separated by blind source separation technique of ICA. The separation signals are separated by complementary set empirical mode decomposition (CEEMD). The signal is decomposed and de-noised, Due to different high and low frequency noises, the decomposition of high and low order intrinsic mode functions (IMFs) should be processed. For the first and last layers of IMF, the detailed information is extracted by using the 3 蟽 principle to further suppress the influence of mode aliasing. Finally, the wavelet threshold is used to Denoise the reconstructed signal to improve the denoising effect and performance index. In order to verify the effectiveness of the method, the simulation and the experimental results of Fen machine of Zhongbei University are carried out. The results show that, This method is better than wavelet soft threshold denoising method and wavelet threshold denoising method based on CEEMD in denoising effect and performance index. It is a new effective signal denoising method.
【作者单位】：中北大学理学院;
【基金】：国家自然科学基金(61275120) 山西省研究生教育创新项目(2016SY050)
【分类号】：TP212;TN911.7

【相似文献】