基于小波变换剔除数据野点的方法  

A Method of Removing the Outliers in Data Processing of Miss-distance Based on Wavelet Transform

作　　者：葛尧[1] 孟庆慈[1] 安玉华[1] 

机构地区：[1]92941部队,辽宁葫芦岛125000

出　　处：《飞行器测控学报》2006年第3期64-67,共4页Journal of Spacecraft TT&C Technology

摘　　要：在脱靶量数据处理中,野点的存在对采用非线性最小二乘法估计多普勒频率的精度会产生恶化影响。本文研究利用野值点和多普勒频率在小波域上系数的不同,对不同数据段的系数设定不同阈值进行处理,从而达到去除野值点和恢复多普勒频率的目的.试验结果表明,采用小波变换去除了随机测量误差与野值点对多普勒频率的影响,对于后续的脱靶量参数估计没有任何不利影响.During the data processing of miss-distance, because the nonlinear least square method （LSM） is adopted in estimating Doppler frequency, the outliers would worsen its precision. We found the different characteristics of wavelet coefficients between the outliers and ideal Doppler frequency. By applying the threshold method to the wavelet coefficients piecewise, the outliers are removed successfully. The experimental result shows that the influence of stochastic errors and outliers can be removed by adopting the wavelet transform, which is helpful to the successive estimation and optimization of miss-distance parameters.

关 键 词：小波变换 多普勒频率 野值点 脱靶量 

分 类 号：V557[航空宇航科学与技术—人机与环境工程]