Research On Identification Technology Of Communication Modulation Signal Mode

Modulation signal mode recognition technology in communication system is to ensure that the signal can be one of the key technology of safe and reliable communication,in the field of civil and military communication communication all play an important role,can accurately identify the type of modulation method extracts the corresponding parameters,provide favorable basis for signal demodulation,according to the current 5G communication modulation signal oriented mature use,put forward to 5G communication commonly used way of modulation signal,the simulation research of main content has the following several parts:(1)Five 5G commonly used modulation signals,including?/2-BPSK,QPSK,16 QAM,64QAM and 256 QAM are used as the research object.Modeling and simulation are carried out on the signal style according to the modulation schematic diagram and formula of the signal,and the time-domain characteristics of the signal are analyzed.(2)To study the way of modulation signal recognition algorithm based on support vector machine(SVM),select the appropriate kernel function of the modulation signal small sample for testing,validation in a small sample of support vector machine(SVM)algorithm for the applicability of the communication signal modulation type recognition,using the constellation diagram feature extraction and support vector machine algorithm simulation,relative to the characteristics of using scatter diagram convolution neural network recognition algorithm,this algorithm has better accurate recognition rate.(3)Related parameters for support vector machine(SVM)choose difficult problem,selection of improved particle swarm algorithm for parameters optimization of support vector machines(SVM),so as to improve the PSO-SVM classifier is constructed,using wavelet transform feature extraction methods of communication signal modulation style feature extraction,the support vector machine(SVM)randomly selected parameters and the resulting from the improved PSO algorithm to optimize the parameters of the simulation and experiment show that by improving particle swarm optimized classifier performance improved.