| In the information battlefield of modern radar,facing the electronic countermeasures environment with highly dense signals and increasingly complex new system radars,it has been difficult to achieve satisfactory results in radar emitter signal sorting and identification based on conventional parameters;although the radar intra-pulse feature extraction method based on mathematical transformation domain has high research enthusiasm,it is still a difficult task to establish an effective radar signal sequence characterization analysis model.This not only directly affects the performance of radiation source signal sorting and identification,but also brings new challenges to the modern radar intelligence reconnaissance.Therefore,in the complex electromagnetic environment,how to explore and study the description method of radar emitter signal and its feature extraction and analysis,to achieve rapid and accurate sorting and identification of complex signals,in order to meet the modernization of China’s radar countermeasures technology and electronic reconnaissance equipment,automation and intelligence needs,is the key problem of radar intelligence reconnaissance urgently need to solve at this stage.This paper takes the complex network theory as the entry point to study the complex network construction method and feature extraction method of radar emitter signal,to solve the network domain transformation and characterization problem of radar signal sequences,so as to gain an in-depth understanding of the dynamics of radar emitter inter and intra-pulse signal sequences,and to investigate the statistical properties of radar emitter signal network transformation domain and extract the corresponding network topology features.Finally,the feature extraction scheme is verified by simulation experiments.1.The main technical means for feature extraction from radar emitter signals are summarized in the literature.The development of signal sorting based on inter and intra-pulse modulated features is classified,the origin and characteristics of radiation source signal features are discussed,and the feasibility of applying complex networks to radar signal feature extraction is discussed by using complex network analysis time series as an example.2.With inter-pulse feature extraction as the goal,the complex network model is established by limited penetrable visibility graph(LPVG),and the best separability metric is selected to describe inter-pulse features to achieve the purpose of quantitative description of inter-pulse features from a complex network perspective.3.For intra-pulse feature extraction,a dynamic segmentation scheme is proposed for signal sequences.The segmented signal data is constructed into a complex network model,based on which the intra-pulse features of the signal are represented using network statistics to achieve the goal of obtaining the intra-pulse features of the signal from the network domain.4.the feature network of the radar signal is constructed from the extracted features,the community structure analysis is performed,and the sorting of the radar signal is achieved using the community structure of the network.The validity of the extracted features is verified. |
