A novel denoising method and its application to MEMS gyro signal

A novel denoising method based on empirical mode decomposition (EMD) is put forward in this paper. With the new method, the nonlinear and non-stationary signal is decomposed into intrinsic mode functions (IMF) via EMD, then, high frequency IMFs will be processed by threshold denoising approach. The difference between the proposed method and the wavelet transform (WT) is that the EMD method can be applied to nonlinear processes and non-stationary processes adaptively. With the new method, there is no need selecting the base wavelet and it can be applied easily to analyze signals. In order to solve the problem of end effect in EMD, an approach of data extending based on minimum mean-square deviation approach is adopted in the present study. As the result of using the real MEMS signal for experiment the standard deviation was reduced to 0.027 times the original one in new ways. By analyzing the Allan variance, types of noises existing in the signal of MEMS gyro are restrained.