Research On Blind Estimation Technology Of Wireless Communication Signal In Complex Electromagnetic Environment

With the rapid evolution of various radio communication systems,radio signals are becoming more and more dense,radio frequency utilization is becoming increasingly saturated,communication systems and communication signal categories show a trend of diversification,advanced modulation technology makes the form of communication signals more and more flexible and versatile,the development and application of various radio equipment makes a large number of dynamic overlapping radio signals throughout the electromagnetic space.The electromagnetic environment is becoming more and more complex with frequent occurrences.For the blind estimation problem and communication emitters identification of single-channel highly overlapped signals in complex electromagnetic environment,three solutions are proposed in this thesis.(1)For the blind estimation of the number of sources in the case of singlechannel highly overlapped signals,a blind estimation method based on spectral correlation features and local maximum gradient is proposed,and the simulation is verified,which can achieve the blind estimation of the number of sources in 0,1,2 and 3 kinds of overlapped signals.The simulation results show that when the number of sources is 1,the recognition accuracy is more than 97%at-lOdB SNR(signal-to-noise ratio),when the number of sources is 2,the recognition accuracy is more than 97%at-4dB SNR,and when the number of sources is 3,the recognition accuracy is more than 90%at 2dB SNR.(2)A maximum likelihood classification algorithm based on the asymptotic distribution of joint spectral features is proposed for the identification of the modulation mode of completely overlapped signals,and the simulation results show that the identification accuracy is about 90%at-2dB SNR when the number of sources is 1,and about 90%at 5dB SNR when the number of sources is 2.When the number of sources is 3,the recognition accuracy is about 90%under the condition of 5dB signal-to-noise ratio.(3)To address the underlying security authentication problem of wireless communication networks,this paper proposes a fractal theory-based spectrum slicing communication emitters identification algorithm,and conducts simulation verification.The shape of the spectrum slice of the radiation source signal contains the intrinsic differences of the roll-off coefficients of different communication emitter forming filters.In this paper,the fractal theory is used to give a quantitative formula for calculating the fractal features of the circular spectrum cross-section slice of the radiation source signal,and the class of individual communication emitters is classified by using a classifier based on the machine learning theory on the basis of the extracted fractal features of the spectrum slice,so as to realize the objective of the physical layer-based communication emitters identification based on nonlinear steady-state features.The simulation results show that the recognition accuracy can reach more than 95%for several different kinds of classifiers under the 4dB SNR condition.