Analysis Of Subtle Characteristics Of Low-frequency Radiation

The individual identification of the low-frequency radiation sources is a very important researchtopic in the field of communications and electronic warfare. The hardware differences of differentradiation sources exhibit a variety of characteristics in the transmitted signal and certain features aredifferent in different radiation sources. The purpose of individual identification is to extract finefeatures which can distinguish different radiation sources and determine which radiation source thesignal comes from.In this paper, the work mainly including the following aspects:(1) The asymmetry of the power supply induced intermodulation distortion in the power amplifiersof sonar transmitters is analyzed. The behavioral modeling of the power amplifiers is studied and theElman wavelet neural network model is proposed by introducing the wavelet to the Elman neuralnetwork. The comparision of modeling performance through three kind of behavioral models, theVolterra-Laguerre model, the Elman neural network model and the Elman wavelet neural networkmodel, is done.(2) The spur fine feature extraction method of radiation sources under steady-state operation isstudied. Features such as the square integrated bispectra, the hexagon integrated bispectra and theirimprovement form are obtained from the received signal. Principal component analysis and kernelprincipal component analysis methods are used to reduce the dimension of the integrated bispectrafeatures. Features such as the energy entropy, the singular value etropy, the center frequency and theslope can be extracted by doing Hilbert-Huang Transform(HHT) to the received signals of transmitter.And analyszes the feasibility of these features for classification and recognition.(3) In the classification part, studies the support vector machine classifier and the radial basisfunction neural network classifier. By using the extracted steady-state fine features, the combinationclassifier is proved to get good recognition results.This article preliminarily explores the individual identification of different low-frequency radiationsources, which has the same type, the same production batch and the same operating mode. The workhas some theoretical sifnificance and practical applications.