Recognition Of Digital Modulation Signals Based On Higher Order Cumulants And Cyclic Spectrum Aanlysis

The automatic identification of communication signals' modulation methods has been widely used in signal detection,threat analysis,frequency spectrum monitoring and management in military and civil fields,which is a key technology for non-cooperative communication.Its main task is to process the intercepted signal,such as frequency conversion,carrier removal,parameter estimation,etc.,extract characteristic parameters from it,and then use appropriate algorithms to make decisions on the modulation method of the modulation signal,providing basis for subsequent demodulation and other processing.This dissertation mainly studies the modulation identification of digital communication signals in the traditional Gaussian and complex non-Gaussian environments.The main work is as follows:(1)Research on modulation identification of digital communication signals under traditional Gaussian noise environment.For the problem of limited identification signal and low recognition rate using a single cumulant quantity,by using the vector sets constructed by the features of the second,fourth,and sixth order cumulants of the signals to realize the inter-class identification of MASK,MPSK,MFSK and MQAM signals and the intra-class identification of 2ASK,4ASK,8ASK,2PSK,4PSK,8PSK,2FSK,4FSK,8FSK,4QAM,16 QAM,and 64 QAM.On the basis of this,for the problems of a poor recognition rate caused by using the cumulant method on the MFSK and MQAM signal,making improvement by using instantaneous characteristics and spectral correlation theory to construct and extract feature parameters,so as to achieve a better in-class identification of MFSK and MQAM.Finally,the simulation experiment was carried out in MATLAB environment.The simulation results show that when the SNR is 5d B,the recognition rate is above 90%.(2)Research on modulation identification of digital communication signals under non-Gaussian noise environment.Using Alpha stable distribution to simulate non-Gaussian noise,the mixed modulation signals' recognition of multi-carrier OFDM and single-carrier under Alpha stable distribution was studied.Firstly,generalized fourth-order cumulants,generalized instantaneous features and generalized cyclic spectrum features are defined;Then,based on the above concepts,by analyzing five characteristic parameters' difference of the generalized normalized fourth-order cumulants,the maximum value of the instantaneous amplitude power spectral density,the absolute standard deviation of the instantaneous phase on the section with weak signals,the position and numbers of the generalized cyclic spectrum's peak,giving a hybrid modulation method for identifying multi-carrier OFDM signals and single-carrier 2ASK,4ASK,BPSK,4FSK and 16 QAM signals with Alpha stable noise.Finally,the simulation experiment was carried out in MATLAB environment.The simulation results show that the recognition rate of the multi-carrier OFDM can reach 100% when the mixed signal to noise ratio(MSNR)is greater than-5 d B and the recognition rate can reach 90% for the single-carrier when the MSNR is greater than 2 d B.