光纤陀螺随机漂移的建模与实时滤波方法  

A Modeling and real-time filtering method for random drift of fiber optic gyroscope

作　　者：刘鑫[1,2] 乔彦峰[1] 

机构地区：[1]中国科学院长春光学精密机械与物理研究所,长春130033 [2]中国科学院大学,北京100049

出　　处：《黑龙江大学自然科学学报》2013年第1期129-134,共6页Journal of Natural Science of Heilongjiang University

基　　金：国家863高技术研究发展计划资助项目(2010AA8080202)

摘　　要：光纤陀螺的随机漂移具有非平稳和非零均值的特点,传统时间序列分析法必须对原始数据进行预处理后再进行建模,无法满足在线建模的要求。针对这一问题,采用差分自回归移动平均模型和递推最小二乘法相结合的方法对光纤陀螺的随机漂移构建动态模型,再通过基于该模型的无迹卡尔曼滤波算法对随机漂移实时滤波,并与自适应卡尔曼滤波进行实验对比。实验结果表明,该方法具有较高的滤波精度,能更有效地抑制光纤陀螺的随机漂移。Random drift of fiber optic gyroscope has characteristics of nonstationarity and nonzero mean, the traditional time series analysis must pretreat the data before modeling, it cannot meet the real-time requirement. To solve this problem, the method is adopted, that combines autoregressive integrated moving average model and recursive least square algorithm to construct dynamic model, then use unscented kalman filter based on this model to filter for the random drift, and contrast with the adaptive Kalman filter algorithm based on ARIMA model. The resuits show that the proposed method possesses higher filter precision and inhibits the random drift of fiber optic gyroscope more effectively.

关 键 词：光纤陀螺 差分自回归移动平均模型 递推最小二乘法 无迹卡尔曼滤波 

分 类 号：V214.5[航空宇航科学与技术—航空宇航推进理论与工程]