基于拟合和小波分解的引信多普勒频率滤波  

Fuze's Doppler Frequency Filtering Based on Fitting and Wavelet Decomposition

作　　者：裴喆[1] 

机构地区：[1]92941部队92分队

出　　处：《探测与控制学报》2011年第1期65-69,共5页Journal of Detection & Control

摘　　要：防空导弹无线电引信多普勒频率经常包含野值,给试验结果分析带来一定困难,小波变换和时频峰值滤波去除方法只研究了野值所占比例小且成片野值少时的情形。针对该问题研究了一种基于非线性最小二乘拟合和小波分解的滤波方法,该方法首先对低频小波系数进行拟合滤除随机误差,然后将拟合和小波分解相结合判别和替换野值。经仿真验证该方法可以滤除随机误差和野值,多普勒频率序列仅含随机误差和孤立野值时滤波精度很高;若包含成片野值且野值总量小于序列总数的三分之一,滤波精度仍然满足要求。The Doppler frequency of anti-air missile radio fuze often contains outliers which causes difficulties of testing result analysis.The methods of wavelet and time-frequency peak filtering only concerned the situations of small proportion of outliers and fewer continuous outliers.A filtering method based on the nonlinear least square fitting and wavelet decomposition was put forward to solve this problem,which could filter stochastic errors by fitting of low frequency wavelet coefficient at first,then distinguish and replace outliers by combining fitting and wavelet decomposition.The simulation indicated that the method could filter stochastic errors and outliers,and with a high accuracy when the order of Doppler frequency data contained stochastic errors and isolated outliers only.The filter accuracy could fulfill the practical demand even if the data contained continuous outliers and the number of outliers was smaller than one-third of the total number.

关 键 词：引信 多普勒频率 滤波 拟合 小波 野值 

分 类 号：TJ430[兵器科学与技术—火炮、自动武器与弹药工程]