Compensation for Random Errors in MEMS Based on AOA-VMD

In order to reduce the random error of the gyroscope, a new denoising algorithm is proposed. This algorithm first preprocesses the original signal using SG (Savitzky-Golay) smoothing filter. Then, an Archimedean Optimization Algorithm (AOA) is employed to optimize the two key parameters of Variational Mode Decomposition (VMD), decomposition level, and quadratic penalty factor. Subsequently, the VMD method is used to decompose the signal processed by SG filtering into multiple intrinsic mode functions (IMFs). Based on the correlation coefficients of each mode, the IMFs are classified into information IMFs, mixed information and noise IMFs, and high-frequency noise IMFs. The high-frequency noise IMFs are discarded, and an improved wavelet threshold denoising method is applied to the mixed information and noise IMFs. Finally, the denoised signal is reconstructed by merging the denoised IMFs obtained using the improved wavelet thresholding technique and the information IMFs. Through simulation experiments, the denoised signal is improved from 12 dB to 24 dB, and the root-mean-square error is reduced by 75.5 %. Additionally, denoising comparative experiments are conducted with standalone wavelet soft and hard thresholding methods, as well as SG smoothing filtering. The experimental results demonstrate that the proposed method exhibits superior denoising performance.