Research On Identification Technology And Impact Factor Of Modulation Mode Of Digital Communication

With the continuous iterative upgrading of radio technology,digital communication modulation technology is becoming more and more complex and changeable.How to identify the modulation technology of digital communication in complex environment,and then demodulate,monitor and interfere has become an important premise for radio management institutions to effectively maintain radio wave order.To solve this problem,this paper combs the methods of modulation recognition studied by many scholars at home and abroad,and puts forward the method of this paper based on the experience and practice of previous studies.The main research objects of this paper are MASK ? MFSK and MPSK.Therefore,firstly,the main principles of the three modulation modes are introduced,and based on the principle,the differences between the three in instantaneous amplitude,instantaneous phase and instantaneous frequency are analyzed,which lays a foundation for the subsequent identification using instantaneous features.Secondly,from the perspective of inter class identification,this paper obtains the maximum spectral density of the normalized instantaneous amplitude of the zero center,which can distinguish the MASK modulation mode 100% from the three,and can distinguish the MFSK and MPSK by using the fourth-order moment energy of the normalized instantaneous frequency of the center.Then,from the perspective of intra class distinction,it introduces the characteristics of intra class distinction among MASK?MFSK and MPSK.Among them,the concepts of frequency domain spectral line cluster and instantaneous linear phase fitting curve slope are proposed for the first time.Based on the conclusion of feature analysis,this paper further analyzes the influence of predicted carrier frequency and symbol length on the final recognition result,and determines the maximum spectral density of Zero Center normalized instantaneous amplitude,the fourth moment of normalized central instantaneous frequency,the standard deviation of Zero Center normalized instantaneous amplitude absolute value,frequency domain spectral line cluster The five eigenvalues of the slope of the instantaneous linear phase fitting curve are used as the decision node to realize the 2ASK?4ASK?2FSK?4FSK?8FSK?2PSK?4PSK?8PSK.Based on the modulation recognition of decision tree,this paper further uses libsvm to realize the intra class recognition of MASK and MSPK,and studies the influence of using one eigenvalue under signal-to-noise ratio as the training set and multiple eigenvalues under signal-to-noise ratio as the training set on the prediction results.Then,GA algorithm is used to optimize the parameters of libsvm to improve the final recognition result.