摘要
研究了一种基于集成经验模态分解的滤波技术,可更准确地从惯性导航系统中对输出数据进行信号去噪处理,以便提取出更有效的有用数据.EEMD去噪是根据信号自身尺度来分解信号,再基于分解产生本征模态函数IMF的Fourier变换频谱,最后构造EEMD时空滤波器.文中对非平稳仿真信号用小波去噪与EEMD去噪两种方法相比较,EEMD滤波效果显然优于小波去噪,并且避免了小波去噪时参数的选择,具有更好的自适应性.通过对高塔测试数据的消噪处理,证明了EEMD滤波技术能够有效地消除高塔测试数据中的高频随机噪声,并根据各分解层的频谱分布提取有用信号,为下一步研究提供了更加有效的数据.
This article presents a kind of filtering technology based on integration of empirical mode decom- position. This technology is more effective on denoising the signal of output data through inertial navigation system, in order to extract more reliable and realistic data. Firstly EEMD denosing method decomposes the signal based on the different scale of signals, then generating the Fourier transform frequency spectrum based on the intrinsic mode function（IMF）, finally constructing the EEMD spatiotemporal filter. Comparing with the wavelet denoising, EEMD method is more effective on denoising the simulated non-stationary signals on tower test and also avoids the parameter selection of wavelet denoising, indicating that EEMD has better self- adaptability on signal denoising. The tower test data denoising process testifies that the EEMD filtering techn- ology can effectively remove the high frequency random noise, extract useful signal through studying the spec- trum distribution of each decomposition layer, and provide more reliable and realistic data for further research.
出处
《测试技术学报》
2013年第5期419-424,共6页
Journal of Test and Measurement Technology
