基于提升小波的光纤陀螺1/f~γ类型分形噪声滤除方法

Elimination of 1/f~γ Fractal Noise from Fiber Optic Gyro Based on Lifting Wavelet

光纤陀螺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类型谱密度特点的一种非平稳随机噪声——1/fγ类分形噪声。传统的滤波方法无法有效去除该类噪声,而由于分形信号在小波域所具有的特性,使小波变换成为研究分形噪声的有力工具。提升小波相较传统小波变换结构更简单,速度更快,计算更简便,结果更准确,对存储空间要求较低,故采用一种新的基于提升小波的滤波方法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,对多组实测数据进行仿真,结果验证了这种方法的有效性。

Random noise for a fiber optic gyro mainly consists of white noise and 1/f^γ fraetal noise, which characterized by long-term correlation, self-similarity and 1/f^γ spectral density with 1/f^γ power law. It is difficult to eliminate fractal noise with traditional method, but with fractal noise＇s special property in wavelet domain, wavelet analysis has become a powerful tool for studying fractal noise. Compared with the traditional wavelet transform, with the advantages of faster speed and higher precision, wavelet transform based on lifting scheme has a simple structure, needs no much storage space and is very easy to implement. In this paper, a new simple and effective method for eliminating fractal noise from a fiber optic gyro was proposed, and it can be further proven by simulation results.