经验模态分解中的频域分辨率及其改进方法

经验模态分解(EMD)的主要目的是提供满足Hilbert变换要求的单组分或窄带信号.针对EMD中由于模式混淆以及信号间相互作用带来的单个本征模态函数带宽过大的不足,对单次EMD分解结果本征模态函数的带宽进行了研究,计算了其瞬时频率分辨率,以此为依据提出了经验模态分解中限制当前信号带宽的改进屏蔽信号方法.此方法完全解决了模式混淆的问题,尽可能地减少了经验模态分解中信号相互作用的不利影响,有效地提高了本征模态函数经H ilbert变换后其瞬时频率表达的频域分辨率.

Frequency-Domain Resolution and Its Improvement During Empirical Mode Decomposition

The main task of Empirical Mode Decomposition(EMD) is to offer the monocomponents or narrow band signals prior to Hilbert transform.During the EMD,the frequency band of intrinsic mode function(IMF) is rather wide due to the mode mixing and the signal interaction.In order to solve this problem,the band width of IMF resulting from a single EMD is investigated,and the corresponding instantaneous frequency resolution is calculated,on the basis of which an improved signal masking method is proposed to narrow the frequency band of signal before the standard EMD.Experimental results show that the proposed method can not only overcome the mode mixing in EMD completely but also avoid the disadvantages caused by the signal interaction,thus effectively improving the frequency-domain resolution expressed by IF after the Hilbert transform of IMF.