Airborne Radar Emitter And Working Mode Identification

With the rapid development of science and technology,electronic warfare has gradually become the core of modern military war.As the precursor and foundation of electronic warfare,electronic reconnaissance is mainly to detect and intercept enemy radar radiation sources,in order to obtain enemy radar radiation source information,and provide technical information support for subsequent battlefield situation assessment and enemy attack.The main work includes:The radar pulse parameters and the modulation mode of main parameters are introduced,on this basis,the extraction of radar related parameter features is studied,and the common working modes of airborne radar are analyzed,finally,the basic concepts of transfer learning and domain adaptation are introduced..The identification method of unknown radar emitter is studied.Firstly,the problems existing in the current unknown radar emitter recognition algorithm are analyzed,and the solutions in this paper are proposed.The outlier detection method based on local projection fraction is used to distinguish the known radar emitter data from the unknown radar emitter data in the test samples.The modulation mode is predicted by using the decision tree algorithm,and then the domain adaptation transformation is carried out between the known class data and the unknown class data by using the migration component analysis method,and the working mode of the sample is predicted by the transformed data.Finally,the two-stage classification method based on neural network is used to determine the radiation source category of the sample.The division method of radar signal is studied.In the actual battlefield environment,the intercepted radar pulse is a continuous pulse sequence with unknown length.The working modes of different types of radiation sources are expressed in different ways,and the signal length contained in different working modes of the same radiation source is also different.Therefore,it is necessary to divide the radar signal before working mode recognition.Firstly,the working principle,advantages and disadvantages of common clustering algorithms are introduced.According to the characteristics of clustering algorithms and radar signals,select DBS CAN algorithm to divide the radar signal,and use KANN-DBSCAN algorithm to solve the problem that DBSCAN algorithm is sensitive to parameters.A binary search KANN-DBSCAN algorithm is proposed to reduce the time complexity.The recognition method of radar working mode is studied.Firstly,the hierarchical structure of "radar word radar phrase radar sentence" of radar signal is analyzed,and the radar word is extracted according to the characteristics of radar signal.Based on the defect that CNN and LSTM cannot process non European spatial data,the extracted radar word samples are represented by graph structure,and the GCN network model is constructed for working mode recognition.Finally,the GCN is extended from the two aspects of sampling neighbors and aggregation function,and the GraphSAGE network model is constructed for working mode recognition.