Research And Application Of Automatic Modulation Recognition Of Digital Communication Signals

The research for communication signal modulation mode of automatic identification is one of the important subjects in the field of communication.With the development of electronic technology,especially with the modern software to manipulate and control of the traditional "pure hardware circuit" after the introduction of wireless communication technology,communication signl modulation mode of automatic identification is paid more and more attention.This paper proposes a transient characteristics and the power spectrum characteristics of the combination of digital communication signal modulation mode recognition method,through the neural network classifier to classify recognition,can effectively distinguish between 2ASK,4ASK,2FSK,4FSK,BPSK and QPSK.When the signal noise ratio(SNR)is low,the modulation recognition method based on instantaneous characteristic will be affected easily,this is mainly because that the influence of noise is affect by instantaneous parameters.Signal power spectrum can reflect many characteristics of modulation signal,when the signal carrier frequency,symbol rate and signal-to-noise ratio changes,power spectrum will be affected,but the same basic modulation signal power spectrum shape characteristic,by introducing based on power spectral characteristic parameters,can effectively improve the classification effect of low SNR situation.Using this classification recognition method based on combination of time-frequency characteristics can further improve the robustness of the algorithm.In this paper,the simulation process is including signal modeling,noise modeling,simulation of sampling,processing dealing,parameter extraction,classifier modeling,classifier training,classifier identifing and statistical results.The method use 1 to 2 times the Fourier transform and a small amount of multiply and then completed the correct identification.When the signal-to-noise ratio is more than 15 db,by using the neural network classifier and the text of the four parameters,can make the identification of modulation signal accuracy greatly increased.