惯性导航EEMD区间阈值降噪方法

陀螺随机误差是影响惯性导航系统精度的主要因素。在经验模态分解（EMD）和阈值降噪的基础上，提出一种基于集合经验模态分解（EEMD）的区间阈值的陀螺信号降噪方法。该方法利用EEMD方法将陀螺信号分解多个本征模态函数（IMF）分量和1个残余分量，基于信号和IMF分量的概率密度函数的2范数距离方法剔除纯噪声IMF分量，利用改进的区间阈值降噪方法实现信号的降噪。仿真和实测试验表明，该方法不仅能有效抑制EMD中的模态混叠问题，而且能有效削弱陀螺的随机误差，从而提高惯性导航系统的精度和可靠性。

EEMD Interval Threshold De-noising Method for Inertial Navigation

The stochastic error of gyro is the significant factor that affect the precision of inertial navigation system (INS).On the basis of empirical mode decomposition (EMD) method and threshold de-noising method, proposed an ensemble empirical mode decomposition(EEMD) interval threshold de-noising method for gyro signal.The gyro signal is decomposed into several intrinsic mode function(IMF) components and one residual component, then rejecting the IMFs that contain pure noise by the 2-norm measures between the probability density function of IMFs and signal, using an improved interval threshold de-noising method to complete the signal de-noising process.Simulation and real test experiments show that the proposed method can not only effectively restrain the mode mixing phenomenon in EMD, but also effectively reducing the stochastic error of gyro, thus improving the precision and reliability of INS.