Research On Digital Signal Modulation Recognition Technology Based On High Order Cumulants And Cyclic Spectral Parameters

Military powers around the world attach great importance to the research of communication signal countermeasures,and communication signal reconnaissance is a key focus.Any party that first detects the other party’s important communication information and then deciphers and interferes with it will bring huge information and battlefield advantages to the reconnaissance party.Digital signal modulation recognition technology is a key link and important component of communication signal reconnaissance,as well as the foundation and basis of signal demodulation.In recent years,modulation methods and communication channels have become increasingly complex,which has also brought new challenges to the modulation recognition technology of digital signals.Therefore,conducting research on digital signal modulation recognition technology has important research significance.The main research work and achievements of this article are as follows:1.Research has been conducted on endpoint detection technology.The general modulation recognition system analyzes a single signal,while the actual received signal is often a mixed signal.The endpoint detection technology not only intercepts the modulated signal segments from the mixed signal segments,but also performs denoising on the modulated signal,which is the technical foundation for subsequent enhancement and recognition of the modulated signal.In different signal types of environments,this article studied four endpoint detection algorithms,namely short-term energy to zero ratio method,short-term spectral entropy method,short-term energy to entropy ratio method,and short-term MFCC cepstrum method.The results show that for the seven modulation signals in the experiment,the short-term energy entropy ratio algorithm has the highest endpoint detection accuracy;In a Gaussian channel with a signal-to-noise ratio of 0dB,the endpoint detection accuracy for all seven modulation signals remains above 93%,which is superior to the other three algorithms.2.Three noise reduction algorithms based on short-term energy entropy ratio algorithm were studied.The three noise reduction algorithms are empirical mode decomposition(EMD),intrinsic time scale decomposition(ITD)and wavelet decomposition.The research results indicate that the wavelet decomposition denoising algorithm has the best effect and the shortest required time;In Gaussian and Rayleigh channels with a signal-to-noise ratio greater than-10 dB,this algorithm can maintain endpoint detection accuracy above 95%,which is superior to the other two denoising algorithms and the control group without denoising.3.In the Gaussian channel,modulation recognition techniques for digital signals were studied,and optimization schemes for 2FSK and 4FSK modulation signal recognition algorithms were proposed.This article analyzes the shortcomings of algorithms that use instantaneous parameters and the number of cyclic spectral peaks to distinguish 2FSK and 4FSK signals,and optimizes the related algorithms of cyclic spectra.The kurtosis coefficient values of the cyclic spectral parameter matrix at cyclic frequencies are used to distinguish these two signals.The results show that at a signal-to-noise ratio of 4dB,the modulation recognition algorithm proposed and designed in this paper can effectively distinguish these two types of signals,with a recognition accuracy of over 99%,which is superior to the recognition algorithm based on instantaneous parameters and the number of cyclic spectral peaks.