Research Of Low-frequency Radiation Recognition

Identification of low-frequency radiation sources is one of the important research topics in the field of electronic reconnaissance.The classification of the radiation sources is determined by characteristics of the received radiation signals and the existing information.The differences of electronic components are common,so even the same type of radiation sources are a little bit different.And these differences are superimposed on the transmitted signal.The key of individual identification is to extract all kinds of subtle features which can distinguish different radiation sources and select the appropriate classifier to give the recognition result.The research object of this thesis is the same type of radiation source under the same working mode.The main contents of the study are as follows:First of all,the class-D power amplifier circuit and power circuit are designed.The experiment is carried out in the muffler pool to collect several kinds of signals.Secondly,the method of feature extraction is based on time-frequency analysis.Renyi entropy of time-frequency distribution is obtained from Ensemble Empirical Mode Decomposition(EEMD).EEMD and FFT is combined to extract spectral features.EEMD+FFT feature is steadier and more effective than Renyi entropy under the condition of white noise and simulated ocean ambient noise.Thirdly,the other aspect of feature extraction is based on higher-order statistics.The singular value features of cyclic bispectrum slice matrix and one-dimensional diagonal slices of fourth-order cumulants are studied here.The analyses of simulation signals and the real experimental signals demonstrate that these two features have good separability.Finally,the local linear embedding algorithm is used to reduce the dimensions of high-dimensional features and simplify the calculation.In the classification part,the support vector machine(SVM)and DS theory are studied.The DS-SVM combining classifier is used and basic probability assignment function is constructed by confusion matrix.Through simulation and real experimental analysis,this combining classifier obtains a good recognition result.The feature extraction and classification of low-frequency radiation sources are studied in this thesis which has a certain value in solving the problem of individual identification.