基于广义交叉验证的光纤陀螺动态信号小波阈值去噪  

Wavelet Threshold De-Noising of FOG Dynamic Signal Based on Generalized Cross-Validation

作　　者：董明[1] 贾学东[1] 陈坡[1] 吕静 

机构地区：[1]信息工程大学测绘学院 [2]61416部队

出　　处：《信息工程大学学报》2011年第4期482-487,共6页Journal of Information Engineering University

摘　　要：动态环境下,光纤陀螺的噪声更加难于处理并且是制约光纤捷联惯导精度的重要因素。针对光纤陀螺动态信号的复杂特性难以准确估计,提出一种利用广义交叉验证对信号的小波分解层分层确定阈值的方法。在信号先验统计特性未知的情况下,利用广义交叉验证原理,该方法能够识别信号和噪声所在的小波分解层,从而分层自适应给出合理阈值。对模拟数据和动态信号仿真算例的分析表明,该方法可有效解决在动态条件下抑制光纤陀螺噪声与保留高频信号的矛盾,提高光纤陀螺动态信号处理的精度。In dynamic environment, it is hard to process fiber optic gyro （FOG） noise, so it becomes an important constraint factor of inertial navigation accuracy. Considering the uncertain and complicated character of FOG dynamic signal, this paper develops a new method based on the technique of generalized cross-validation to adaptively determine the threshold of every wavelet layer. This method can identify wavelet signal and noise layer automatically and determine ideal threshold of wavelet decomposition level without statistic character of signal. Experiments with simulated data series and dynamic data show that the method can solve the contradiction between suppressing noise and reserving signal in dynamic environment and improve the accuracy of FOG dynamic signal processing.

关 键 词：光纤陀螺 动态信号 广义交叉验证 小波阈值去噪 小波变换 

分 类 号：U666.1[交通运输工程—船舶及航道工程]