基于经验模态概率分布的光纤陀螺信号处理  

Fiber optic gyro signal processing based on empirical mode probability distribution

作　　者：崔冰波[1] 陈熙源[1] 龚政仰 

机构地区：[1]东南大学仪器科学与工程学院微惯性仪表与先进导航技术教育部重点实验室,南京210096 [2]重庆延锋彼欧富维汽车外饰有限公司

出　　处：《中国惯性技术学报》2015年第5期690-695,共6页Journal of Chinese Inertial Technology

基　　金：国家自然科学基金(51375087;50975049);江苏省普通高校研究生科研创新计划资助项目(KYLX_0106)

摘　　要：为了抑制光纤陀螺随机漂移，基于改进的经验模态分解(EMD)和新型模态筛选标准提出了一种自适应的区间阈值滤波方法。首先分析加入高斯噪声对 EMD 分解结果的影响，提出有界噪声辅助以改善 EMD 分解质量，然后针对本征模态函数的概率分布特征提出了基于样本熵的模态筛选标准，最后采用数据驱动的阈值选择方法实现自适应的区间阈值滤波。为了验证算法的有效性，采集一款干涉型光纤陀螺静态漂移信号进行实验分析，结果表明本文方法较基于平稳小波变换和 EMD 的阈值滤波有更好的去噪效果。仿真分析表明该去噪算法减小了捷联惯性导航系统的航向角误差，均方根误差较平稳小波变换去噪算法改善了约 78.6%。Based on an improved empirical mode decomposition（EMD） and a newly proposed mode selection criterion,an adaptive interval threshold filtering method is developed to mitigate the random drift of fiber optic gyroscope. First,the effect of adding Gaussian noise to assist EMD decomposition is analyzed,and a bounded assist is developed to improve decomposition quality. Then,aiming at the probability distribution feature of intrinsic mode function,a novel mode selection criterion is proposed. Finally,an adaptive interval threshold filter is developed based on data-driven threshold selection. The experiment analysis,which employs static data detected from an interferometric FOG,is performed to verify the proposed algorithm,and the results show that,compared with threshold filtering methods based on stationary wavelet transform（SWT） and EMD,the proposed method significantly improves the denoising result. In addition,the simulation results show that,compared with SWT-based method,the proposed method improves the heading accuracy of strapdown inertial navigation system by 78.6% in term of mean square error.

关 键 词：光纤陀螺 经验模态分解 概率密度函数 样本熵 

分 类 号：V241.5[航空宇航科学与技术—飞行器设计]