Research On The Recognition Of Shortwave Digital Transmission Signal Modulation Mode

For communication security,regulators need to monitor and analyze signals.In the case of monitoring,if the communication parameters are not known,the signal needs to be blindly identified.The identification of the signal mainly includes carrier estimation and modulation identification,wherein the identification of the modulation mode is the most basic and most important.The short-wave digital transmission channel has the characteristics of signal fading,multipath effect,Doppler shift,etc.,and the signal is difficult to directly recognize after passing through the short-wave number transmission channel.Therefore,this paper studies how to effectively identify the digital modulation mode in the short-wave digital channel environment.This paper first studies the short-wave digital transmission channel,builds a short-wave channel model,and experimentally analyzes the characteristics of the short-wave channel.Many changes have taken place after the discovery signal passes through the short-wave number transmission channel,which is not recognized at all from the basic definition.Then,in this paper,2ASK(Amplitude Shift Keying),4ASK(Quadrature Amplitude Shift Keying),2FSK(Frequency Shift Keying),4FSK(Quadrature Frequency Shift Keying),2PSK(Phase Shift Keying),4PSK(Quadrature Phase Shift Keying)are generated.Six kinds of digital modulated signals are extracted through the channel,and the statistical characteristics(instantaneous features,high-order cumulant characteristics)of the signals in the short-wave channel are extracted.Experiments have shown that these features are effective for the identification of digitally modulated signals.The experimental simulation of the statistical characteristic parameters with the change of signal-to-noise ratio curve,and then select the appropriate characteristic parameter threshold through the curve of the characteristic parameters in the short-wave channel environment.Then,the classification and recognition of the signal is realized by the method of the decision tree.The selection of feature and parameter value thresholds in the joint statistical statistic modulation method based on decision tree affects the recognition result,and the decision threshold result is related to the training set,which needs manual experiment to determine.Aiming at this problem,this paper proposes a modulationmode recognition model based on binary tree support vector machine to solve the problem of feature parameter selection in multi-classification problems.Finally,a comparative analysis of the two methods in the recognition rate,the complexity of the algorithm and the ability to adapt.The experiment validates the method based on binary tree support vector machine,which has good recognition performance and adaptability.It also verifies that the high-order cumulant has the characteristics of anti-noise and anti-channel fading.