Adaptive MEMS gyroscope denoising method based on the à trous wavelet transform

Based on the a trous wavelet transform, an adaptive MEMS gyroscope angular velocity denoising method is proposed. In order to overcome the phase distortion resulted by decimation used in the traditional wavelet transform and to cater for the requirement that future data values cannot be used in the calculation of the wavelet transform, the Haar a trous wavelet transform is selected to filter the signal. And for further retaining the singularities of signal other than noise, a time varying signal denoising is performed on the segmented detail coefficients at different scales. The experimental results show that this method not only reduces the variance of signal efficiently when the platform is stationary but also keeps track the real signal timely when the platform is dynamic.