基于LMS与二代小波变换的光纤陀螺去噪算法

Denoising algorithm for FOG based on LMS and second-generation wavelet transform

为了降低光纤陀螺输出中的噪声分量,提出一种基于最小均方法与二代小波变换相结合的去噪方法。首先利用LMS算法进行前端预处理,提高信号的信噪比;然后使用SGWT去噪算法降噪,考虑到SGWT去噪算法易受阈值函数的影响,将模糊与平滑因子引入到传统软阈值法,以缩小估计小波系数和原小波系数两者之间的常值偏差;最后,将本文提出的算法应用于某型光纤陀螺的去噪研究中。实验结果表明,相对于SGWT去噪算法,采用LMS-SGWT算法处理后,光纤陀螺的信噪比从0.1698d B提高到2.0521 d B,方位对准误差从0.33°降低到0.13°。

To denoise the FOG output signal, a denoising algorithm based on LMS (Least Mean Square) and the Second Generation Wavelet Transform (SGWT) is proposed. Firstly, the LMS algorithm is used to pretreat and improve the signal's signal-noise ratio (SNR). Then the SGWT algorithm is used to further denoise the signal. Considering that the SGWT algorithm is easily influenced by threshold function, a fuzzy factor and a smooth factor are introduced to reduce the constant deviation between estimated wavelet coefficient and original wavelet coefficient. A certain kind of FOG is used to verify the LMS-SGWT algorithm. Experiment results show that the SNR of FOG is increased from 0.1698 dB to 2.0521dB, and the alignment error of heading is decreased from 0.33° to 0.13°. ©, 2014, Editorial Department of Journal of Chinese Inertial Technology. All right reserved.