基于新型经验模分解的激光陀螺漂移趋势项提取

Trend Extraction from Laser Gyro Drift Data Based on Modified Empirical Mode Decomposition

为了提高激光陀螺建模精度进而提高惯性导航系统的精度,需要对激光陀螺漂移时间序列的趋势项进行精确提取。针对陀螺漂移时间序列非线性、非平稳的特点,提出一种基于自适应正弦延拓的新型经验模分解方法对其趋势项进行提取。该方法根据信号端点附近数据变化趋势,在其两端加上适当相位、幅值和频率的正弦延拓函数,从而抑制端点效应,提高分解精度。数字仿真和实例应用结果表明该方法具有筛选次数少,自适应能力强,分解精度高的优点,是提取趋势项的一种有效方法。

To improve the modeling accuracy of laser gyro and hence the precision of inertial navigation system, trend item extraction from laser gyro＇s drifting data was needed. Considering the nonlinearity and non-stationarity of laser gyro＇s drifting data, a modified EMD method based on adaptive sine extending to extract the trend item of Gyro＇s drifts was proposed. The proposed method added sine functions with proper phase, amplitude and frequency to the ends of the signal according to its tendency, so as to restrain end effects and improve the decomposition accuracy of normal EMD. Numerical simulation and practical application results indicated that the proposed method is efficacious with the merits of strong adaptability and high decomposition accuracy.