针对现代电子战环境下雷达辐射源信号的高度密集、复杂调制、交叠概率大等特点,基于时域、频域、时频域以及其他数学变换域的信号分析方法仍然不能满足雷达辐射源信号分选识别的要求,结合雷达辐射源信号具有可分辨性的确定数据结构表示,受时间序列的复杂网络分析启发,提出一种基于复杂网络的雷达辐射源信号脉内特征提取算法。该特征提取算法首先采用相空间重构方法对信号频谱序列数据进行预处理,将重构后的信号序列转换至复杂网络域进行表征;其次,通过复杂网络建模及其统计特征分析,研究雷达脉内信号时间序列动力学微观特性,从而挖掘信号的有效特征参数;最后,在不同噪声环境下,实现雷达辐射源信号序列的特征参数的分类性能及其准确性分析。仿真结果表明,所提取的特征参数在低信噪比环境下具有良好的抗噪能力和不错的聚类质量,验证了基于复杂网络时间序列分析的信号特征提取方法的可行性,为进一步丰富刻画雷达辐射源信号提供了有效的技术支持和手段。

Given the highly dense, complex modulation, and overlapping probability characteristics of radar emitter signals in the modern electronic warfare environment, signal analysis methods based on time domain, frequency domain, time-frequency domain, and other mathematical transformation domains still cannot meet the requirements of radar emission signal deinterleaving and identification. Therefore, combining the identifiable data structure representation of radar emitter signals with insights from complex network analysis of time series, a complex network-based algorithm for extracting intra-pulse features is proposed from radar emitter signals. The feature extraction algorithm first preprocesses the signal spectrum sequence data using phase space reconstruction to transform the reconstructed signal sequence into the complex network domain for characterization. Then, through complex network modeling and statistical feature analysis, the micro-dynamic characteristics of radar intra-pulse signal time series are studied to explore effective feature parameters of the signal. Finally, under different noise environments, the classification performance and accuracy analysis of the feature parameters of radar emission signal sequences are implemented. Simulation results show that the extracted feature parameters have good noise resistance and satisfactory clustering quality in low signal-to-noise ratio environments, verifying the feasibility of the signal feature extraction method based on complex network time series analysis and providing effective technical support and means for further characterizing radar emitter signals.

TN957

国家自然科学基金资助项目(61961042,61461051)