Gyro Signal De-noising Based on a New Wavelet Threshold Function

Abstract

Gyro random drift is a remarkable factor which influences the precision of north seeker, commonly used wavelet threshold method for gyro signal de-noising, but wavelet threshold function has some flaws, which affects the de-noising performance. Through analysing the advantages and shortcomings of soft and hard threshold functions and several existing improved threshold functions, puts forward a new double and adjustable threshold function, which has simple expression, without partition value, convenient adjustment, and excellent mathematical properties. The more important is that wavelet coefficients were processed with rapid contraction when absolute value is below the threshold in the threshold function, which retains true signal to a certain extent. The seminal work on "Doppler" signal and gyro to measurement signals have shown that the new threshold function overcome effectively the shortcomings of soft and hard threshold function, and restrain significantly the noise. Thus, it has certain application value.