MEMS gyro de-noising based on coefficients dependency in wavelet transform domain

Wavelet de-noising is widely recognized as an effective tool for gyro signal processing. In order to reduce the MEMS gyro random drift error, a wavelet transform de-noising technique based on inter and intra scale coefficients dependency in the wavelet transform domain is introduced in this paper. Firstly, based on the correlation not only between neighboring coefficients but also across the scales wavelet coefficients dependency, a wavelet coefficients dependency proportional shrinkage de-noising method with best approximation is proposed. Then this wavelet de-noising method is applied to practical gyro output signal and compared with some threshold de-noising methods. The results of the noise error analysis after application show that the proposed method is very effective.