The de-noising of gyro signals by bi-orthogonal wavelet transform

It is important for the strapdown inertial navigation system to determine the carrier initial azimuth angle and attitude angle accurately and quickly. The main factor that affects its accuracy is gyroscope drift. Therefore, how to eliminate the gyroscope drift effectively is the key to guarantee the north-searching accuracy. In general, the random drift is weak nonlinear and slow time invariant. And at the same time it is often affected by some indeterminate factors such as exterior environment noise. So the random drift ratio becomes the most important factor that measures the accuracy of gyroscope. Wavelets analysis is especially suitable to non-stationary signal processing and with relatively good filtering results. So it is appropriate for gyro signal filtering. In this paper, bi-orthogonal spline wavelet with symmetry is utilized to gyro signals to perform multi-resolution analysis. According to the analysis results, we know that wavelet coefficients are generated by noise. Therefore a reasonable gyro signal de-noising method is proposed in this paper. Compared with other traditional signal de-noising methods, the filtering results by this method is relatively good enough to satisfy the requirements of practical use.