雷达辐射源信号模糊能量分布特征提取及其分选机制

For a long time, the methods of pulse trains deinterleaving are generally based on five conventional parameters, i.e., time of arrival (TOA), radio frequency (RF), pulse width (PW), pulse amplitude (PA) and direction of arrival (DOA). However, as the signal environment is becoming more concentrated, more complex and more changeful, many conventional deinterleaving methods which only rely on the above five parameters are losing effectiveness.. In current study, the features extraction of ambiguity energy distribution and theirs deinterleaving mechanism will be profoundly investigated so as to find the effective characteristics set which meet the actual demand of signals deinterleaving. More specifically, we intend to build the quick searching method of the slice of ambiguity function main ridge (AFMR), which is based on intelligence optimization and the priori information of signals, so that the characteristics of ambiguity energy distribution supply the demand of real-time deinterleaving. And for the purpose of obtaining the essential characteristic parameters which represent the individual specialties of radar emitter signals, we will adequately extract the features of the slice of AFMR from three perspectives of globality, locality and variability. Finally, we will investigate the separation ability of these extracted characteristics and the five conventional parameters via the comprehensive analysis and evaluation of the capability of anti-noise, the accuracy of signals sorting, the sensitivity of deinterleaving and the timeliness of feature extraction. These researches can help us to make clear the deinterleaving mechanism of signal characteristics so as to build a new and effective characteristics set of deinterleaving.. In general, the research results of this project can provide direct theoretical basis to the technological advance of radar warfare equipment.

长期以来，雷达信号分选主要依靠到达时间、载波频率、脉冲宽度、脉冲幅度和到达方向五个常规参数，这已难于满足当前密集、复杂、多变信号环境的需求。本项目旨在模糊能量分布特征提取及分选机制研究的基础上，探索能满足信号分选实际需求的有效特征参数集。项目拟通过研究基于智能优化和信号先验信息的模糊函数主脊切面快速搜索方法，使模糊能量分布特征尽量满足信号实时分选的要求；通过从全局、局域和变化特性多个角度对信号模糊能量分布特性进行充分特征提取，以获得表征信号个体特性的本质特征参数；通过特征的噪声稳健性、分选准确性、灵敏度特性以及特征提取时效性对包括传统五参数在内的特征的可分选性进行综合分析与评价，以发掘特征的内在分选规律，探索新的有效分选特征参数集。项目的研究成果将为我国雷达对抗装备的技术进步提供直接的理论依据。

长期以来，雷达信号分选主要依靠到达时间、载波频率、脉冲宽度、脉冲幅度和到达方向五个常规参数，这已难于满足当前密集、复杂、多变信号环境的需求。本项目旨在模糊能量分布特征提取及其分选性能研究的基础上，探索能满足信号分选实际需求的有效特征参数集。项目从模糊函数主脊切面快速搜索方法、模糊函数主脊切面多角度特征提取方法、特征性能分析与评价方法三个方面展开研究，充分挖掘能表征信号个体特性的本质特征参数，提出基于智能优化的AFMR切面快速搜索、基于图形轮廓的AFMR切面特征提取、基于极坐标转换的AFMR切面形态特征提取、AFMR切面局域差分特征提取、模糊函数三维特征提取、模范函数等高线特征提取等多种有效方法，显著提升了搜索模范函数主脊切面的时效性，提取了AFMR图形轮廓特征、AFMR局域差分特征、模糊函数三维特征以及模范函数等高线特征等近十种有效分选特征，丰富了雷达辐射源信号分选特征参数集，并构建了基于先验信息库的多源混合信号快速识别模型，为我国雷达对抗装备的技术进步提供了直接的理论依据。