Research On Modulation Recognition Technology Of Communication Signals

The information transmission in modern warfare depends mainly on radio communication.The main task of communication reconnaissance is to intercept,measure,analyse,identify and demodulate the signal of communication emitter,and modulation recognition was one of the key steps of communication reconnaissance.Only correct identification of the modulation type can provide a favorable condition for the subsequent demodulation and countermeasures.In this thesis,recognition of modulation mode of digital communication signal was researched,and modulation recognition algorithm was researched on feature extraction and classifier design respectively.The main summary of the thesis is as follows:(1)Researching on modulation recognition algorithm based on instantaneous signal characteristics.In this algorithm,seven kinds of instantaneous characteristics of communication signals,such as the peak amplitude of the amplitude spectrum and the standard deviation of the absolute amplitude,were selected.According to the classification performance of each characteristic parameter,judgment criterions of decision tree classifier were set up,and the decision threshold was selected via simulation results,then the decision tree classifier was built.The performance of the algorithm was analyzed by simulation experiments.(2)A modulation recognition algorithm based on the combination of high order cumulant and entropy was proposed.In this algorithm,in order to improve the anti-noise ability and stability of the algorithm,the extracted characteristics of high order cumulant was used to identifing the digital communication signal,meanwhile,the extracted entropy features were used to identifiing the signals which could not be classified by high order cumulant features.According to the classification performance of each feature,the appropriate decision criteria and threshold values were selected to build a decision tree classifier which was used to realizing the modulation recognition of the signal.The performance of the algorithm was analyzed by simulation experiment,and compared with the modulation recognition algorithm based on the instantaneous characteristics of the signal.(3)Researched on the application of Support Vector Machine(SVM)in modulation recognition.Particle swarm optimization(PSO)algorithm was used in optimizing the parameters of SVM,then the PSO-SVM classifier was constructed.The simulation results showed that the modulation recognition algorithm based on PSO-SVM could efficiently identify all kinds of digital communication signals under the condition of SNR of-3d B.And the classification performance of PSO-SVM was improved compared with SVM.(4)In order to improve the PSO was prone to local convergence,an improved particle swarm algorithm was proposed.Though introducing the concept of neighborhood search in the PSO,this method achieved the improvement of PSO algorithm.The improved PSO was used in optimizing the parameters of SVM,then an improved PSO-SVM classifier was built.The simulation results showed that the modulation recognition algorithm based on improved PSO-SVM could identify all kinds of digital communication signals under the condition of SNR of-4d B,and the identification rate of improved PSO-SVM was improved compared with PSO-SVM.