经验模态分解阈值消噪方法及其在惯性导航系统数据处理中的应用

摘要

：Since colored noise is predominant in sensor errors, fractional Gaussian noise model is established and the model parameter estimation method by power spectral density is given. Noise variance in Jntrinsic mode functions （IMFs） from empirical mode decomposition（EMD） is derived. Noise thresholds of IMFs are estimated through variance and EMD threshold de-noising method is established. The method is applied in INS and compared with wavelet de-noising method. It is shown that wavelet threshold de-noising is poor at suppressing colored noise while EMD threshold de-noising is effective on reducing sensor errors for its connection with proper noise model. INS accuracy is improved through EMD threshoJd de-noising.%针对惯性元件误差中有色噪声影响远大于白噪声的情况，建立元件误差的分形高斯噪声模型，利用功率谱密度方法估计模型参数。基于噪声模型推导经验模分解（EMD）的各固有模态函数（1MF）分量中噪声的方差，以此估计各分量相应的阈值，建立EMD阈值消噪方法。将该方法应用于INS中，并与小波阈值法进行比较。结果表明，小波阈值法难以控制元件中有色噪声的影响，EMD阈值法与噪声模型紧密结合，能够更有效地削弱元件中的随机误差，提高INS精度。