Research of gyro signal de-noising with stationary wavelets transform

In the strapdown inertial navigation system, the accurate and quick determination of the carrier initial azimuth angle and attitude angle is an important problem. The main factor that affects its accuracy is the gyroscope drift. Therefore, how to eliminate the gyroscope drift effectively becomes very important to guarantee the north-finding accuracy. As wavelets have preferable time-frequency resolving characteristics, they are applied to the de-noising of gyro signals and better de-noising results are obtained. However, because classic wavelets do not have the translation invariant property, the de-noising results have some limitations. In this paper, the stationary wavelet is, for the first time, applied in the gyro signal de-noising field. The de-noising result is compared with that of classic wavelets. It shows that the north finding accuracy by using stationary wavelet is obviously better than that by using classic wavelet. It also proves the better de-noising performance of stationary wavelet in such applications.