EEMD滤波技术在GPS动态监测数据去噪中的应用

为了削弱结构动态监测中GPS随机噪声的影响,本文研究了一种基于集成经验模态分解(EEMD)技术的滤波方法。根据信号自身尺度分解信号,基于分解产生的本征模态函数(IMF)的Fourier变换频谱特征,构造了EEMD时空滤波器。对不同信噪比的仿真非平稳数据进行去噪处理并与小波去噪法相比较,各项指标表明基于EEMD滤波器的去噪方法与小波去噪方法效果相当,但避免了小波基的选择,具有更大的自适应性。应用于GPS动态监测数据的去噪结果表明该方法能有效分解信号消除GPS高频噪声及低频噪声的影响,提取有用振动信号,为进一步结构分析提供有效数据。

Application of EEMD filter in GPS dynamic monitoring data de-nosing

To reduce the effects of random noise to GPS signal in structure monitoring,a technique based on EEMD(Ensemble Empirical Mode Decomposition) was studied in the paper.With the EEMD the extremums of the signal were used as scales to decompose data in the time domain and get the IMFs(Intrinsic Mode Function) with their frequencies from high to low.Then based on the frequency spectrum from FT(Fourier Transform) of each IMF a new filter could be establish to de-nose signals.Compared with the wavelet de-nosing method for non-stationary signal with different SNR(Signal to Noise Ratio),it was shown that the result with these two methods were on the same level,but the one with EEMD was more adaptive and could avoid selecting the wavelet.With EEMD filter,the high and low noise in GPS dynamic monitoring signal would be reduced to extract useful vibration response for structure analysis.