弹道目标动态全极化RCS数据的时间序列分析

摘要：基于微波暗室静态测量数据，利用目标的宏观运动学和微运动特性，反演了某弹头模型在全极化条件下的RCS回波数据。分析表明该时间序列具有非平稳性、相关性和拟周期性的特点，总体上可以表征为目标平动引起的慢变化和目标进动引起的拟周期性快变换。为了刻画RCS序列的时间演化行为，利用B样条函数基、广义自回归条件异方差（GARCH）模型和ARMA模型构造了RCS序列的迭合滤波模型。其中B样条函数用来表征趋势项，可以去除均值非平稳性；GARCH模型用来表征RCS的波动聚集性，可以去除方差非平稳性；ARMA模型用来表征受噪声污染的进动项，可以有效提取RCS序列的平稳分量。实验表明，所提方法能够有效分离数据的非平稳和平稳分量，其研究成果可辅助用于弹道目标跟踪、微动特征提取及识别评估等。

关键词：弹道目标；动态RCS；时间序列分析；非平稳性；微运动

中图分类号： TN95             文献标志码：  

Dynamic Full-polarization RCS Time Series Analysis of Ballistic Targets

Abstract: Based on microwave darkroom static measurement data and using features of macro/micro motion, this paper constructs the dynamic full-polarization RCS data for a specific ballistic target in full-polarization channels. The simulation indicates that the RCS time series has the characteristics of nonstationarity, correlation, and quasi-periodicity. The series can be divided into two parts, slowly varying terms induced by translation, and fast varying terms induced by quasi-periodical precession. In order to depict the RCS time evolutionary behaviors, we use B-spline function, generalized autoregressive conditional heteroskedasticity (GARCH) model, and autoregressive-moving average (ARMA) model to constitute a integrated filtering model, where the B-spline function represents the tendency term aimed for reduction of mean nonstationarity, GARCH model represents the volatility clustering term aimed for reduction of variance nonstationarity, and ARMA model represents the noise corrupted sinusoidal term. The experiment indicates that this decomposition method can distill the stationary and nonstationary components effectively. The results can be used for ballistic target tracking, precession feature extraction, and recognition evaluation etc.

Key words: ballistic missile; dynamic RCS; time series analysis; non-stationarity; micro-motion