基于经验模态分解的光纤陀螺1/fγ类型分形噪声滤除方法

提出了一种可有效滤除输出信号中干扰信号的新的光纤陀螺信号降噪方法.中高精度光纤陀螺的随机噪声中包含了白噪声和具有长程相关性、自相似性及1/fγ型谱密度特点的一种非平稳随机噪声--1/fγ型分形噪声.利用传统的滤波方法无法有效去除该类分形噪声,而新方法以经验模态分解(EMD)和阈值降噪方法为核心,结合提升小波理论对光纤陀螺的输出信号进行软阈值滤波,进而提高光纤陀螺的精度.利用这种新方法对多组实测数据进行滤波仿真实验,结果中仅偏置稳定性一项性能指标就已表明,使用新滤波方法滤波后数据精度可提高一个数量级,而相较传统的小波滤波方法可提高2.5倍,故此验证了新方法的有效性.

Elimination of Noises From Fiber Optic Gyro Based on Empirical Mode Decomposition (EMD)

A new noise reduction method which can effectively eliminate fractal noises in the output signals from a fiber optic gyro is proposed. The stochastic noises in a high precision fiber optic gyro not only contain the white noise but also contain the l/fì fractal noise which is an unstable stochastic noise with the features of long-term correlation, self-similarity and l/fì spectral density. The l/fì fractal noise can not be eliminated with traditional filtering methods. The new method proposed is based on empirical mode decomposition (EMD) and threshold demoisi吨and is combined with a lifting wavelet theory. By using this method, the soft threshold filtering can be done for the output signals from a fiber optic gyro and hence the precision of the fiber optic gyro can be improved. A filtering simulation experiment is made for several sets of measured data. The experimental result shows that the precision of the filtered data is improved by an order of magnitude. Compared with the traditional wavelet filtering method, the precision of the data filtered by the new method is 2.5 times higher. Thus, the effectiveness of the new method is verified.