Researches On Identification Technology For Radar Emitter

Radar emitter individual identification technology is a hot issue in the field of electronic reconnaissance.Through the application of this technology,the radiation source and its platform can be uniquely determined,and its activity information and regulation can be accurately grasped.It is of great value and significance to carry out the research of individual recognition technology in the field of military and civil applications.In this thesis,the generation mechanism of individual characteristics,signal preprocessing and analysis of individual characteristics are systematically summarized.The main research contents are divided into the following parts:Firstly,the generation mechanism of radar individual features in three aspects of pulse envelope,spurious output and phase noise is analyzed,which provides the basis for individual feature extraction and recognition in the workflow;the performance requirements of the receiver in the realization of individual recognition technology are studied.Then,the work of radiation source signal in preprocessing stage is introduced.Through analyzing the difference between radar signal in actual detection and reception and theoretical situation,the necessity of noise reduction before processing is analyzed,and the process and application effect of noise reduction and normalization are explained;in the parameter estimation and suppression stage of multi-path signal,genetic algorithm and LWD based are studied method.And then,the thesis introduces the function and common methods of modulation recognition in individual recognition technology;studies the modulation recognition method in multi-path channel based on fuzzy function.It tries to compare the method with wavelet packet energy and high-order statistics.Through experiments on various modulation signals generated by simulation,the recognition rate comparison of three methods under different SNR is obtained.Finally,the time-frequency domain,wavelet packet transform,fuzzy function slicing,cyclic spectrum,cyclic bispectrum,time-frequency reconstruction and other methods of feature extraction in time-frequency domain are studied.Through the simulation,the performance of the method has a preliminary intuitive understanding,and the method selected in the follow-up experiment is determined.Through the experiment on the measured data,several of the most effective feature extraction methods are obtained and analyzed Some suggestions are given for further research.