Research On Communication Emitter Signal Subtle Feature Analysis And Individual Recognition Technologies

With the development of communication technology,people not only focus on the basic parameters of signal features,such as frequency,modulation mode and so on,but also attach more and more importance to the subtle characteristics produced by the differences of individual parameters and production debugging of sources.These characteristics are detectable,repeated,and will not affect the information transmission.Through advanced extraction and classification methods of signal subtle features,the emitter which the signal comes from,and the identify information can be identified in combination with actual requirement in specific application background.Special emitter identification(SEI)technology can also be applied in the fields of wireless network security,electromagnetic spectrum management,communication equipment fault detection.The initiative would be seized in the complicated electronic warfare environment,if an important communication emitter is identified with its subtle features.Therefore,it has important theoretical significance and application value to research on signal subtle feature analysis and individual identification technologies.In this paper,on the basis of analysis of the existing methods of SEI,how to accurately extract subtle characteristics of communication signal and how to improve the identification rate are deeply studied.The main works and academic achievements of the thesis include:1.The research background application requirements of signal subtle feature analysis,and individual identification technologies are briefly introduced.The research status of the SEI are summarized both from theory research and practical application.The existing technologies are classified,and the advantages and disadvantages of each method are summed up.2.The Hilbert-Huang Transform(HHT)is a widely used method of time-frequency analysis,which is a adaptive signal process method without any prior information on the signal.However,in practical application,HHT exists the weakness of endpoint effect and fails to explain the negative frequency and so on.Therefore,an improved HHT(IHHT)method is proposed to extract the signal features.It is proved that the IHHT method can extract accurate transient signal features by simulated experiments,and the endpoint effect is effectively suppressed.3.The conventional Square Integral Bispectra(SIB)features have several disadvantages,such as there are some negative-effect integral paths which reduce the identification rate.This paper proposes a new algorithm based on improved bispectra and time-domain analysis.The low-contribution and negative-effect bispectrum values are removed by redefining the bhattacharyya distance,obtaining the largest proportion of bispectra.Support Vector Machine(SVM)is used to realize the individual identification.Experiment results show that the method is able to identify the same model emitter under the environment of low signal-to-noise(SNR).4.Along with the increasing complex modulation modes of signals,the traditional features such as carrier frequency or symbol rate of signal already can not satisfy the actual needs.An unconventional feature of higher-order spectral skeleton is proposed.This paper models a higher-order spectrum skeleton model,which is based on the K-segments algorithm for principal curves.The information dimensions and box dismensions of the skeleton are extracted to be the transient features,and the time-frequency domain analysis is combined.Finally,the derived feature vectors are trained by SVM classifier in order to identify the signals emitted from different emitters.Experiment results show that the method is able to identify the same mode of emitters.The recognition rate can still reach 85% or higher under the condition of low SNR.5.The production of subtle features is a complex,nonstationary,and nonlinear process,which exists some chaos mechanisms.Therefore,a deterministic system theory technology is difficult to solve this problem.This paper proposes a new feature extraction algorithm of communication signals based on the fractal theory.After preprocessing the received signal,the correlation dimensions based on empirical mode decomposition(EMD)are researched to extract features,which is proved to be effective;As applying the multi-fractal to communication signals analysis,Several related parameters which represent emitters are extracted to be feature vectors from the multi-fractal spectrum,the combined features from correlation dimensions and several related parameters of multi-fractal spectrum are proved to be effective by identification experiments.The identification rate can achieve 99% under the SNR of 10 db.