Signal Classification Algorithm In Spectrum Monitoring

With the development of wireless communication,the communication system has experienced from analog to digital,simple to complex development process,and has formed the situation in which a variety of communication systems coexist.So the wireless signal spectrum and the monitoring are becoming more and more important.Spectrum monitoring and management has become a measure of the strength of a country and also one of the key indicators.In the spectrum monitoring,signal recognition occupies a very important component.Signal recognition is mainly extraction and identification of some of the characteristics of the wireless signal through the wireless signal of unknown parameters.It was also because of this situation,the signal recognition is of great significance for spectrum monitoring.In general,signal recognition is considerable difficult.The difficulty is mainly derived from the parameters in the process of signal recognition and unpredictability,such as amplitude,frequency,phase of fanaticism and other unknown information.It is because of the signal recognition has great significance and has the certain difficulty,this paper mainly studies the single signal identification method based on high-order cumulants and mixed signals identification method based on natural gradient ICA algorithm.For single signal modulation mode recognition,this paper mainly studied the three types of digital signal modulation mode recognition method.First,this paper mainly introduces ASK,PSK and QAM signal generation and characteristics.Then we introduce the higher-order cumulants recognition feature extraction and classification recognition process,mainly including classification decision,tree recognition process and the recognition accuracy.Then the paper researches the influence of the gaussian white noise,the number of symbols,frequency offset and phase.Finally,we introduce how to use MATLAB,signal generator,and blackbird receiver actual measuring system about the signal recognition.Due to the complexity in practical system is stronger,mixed signals system is proposed in this paper and the algorithm based on natural gradient ICA.This paper first introduces the system model under block fading channel,and then illustrates how to separate the mixed signals.In the final separation step,using the number of symbols,separation accuracy and robustness are researched and vertified by simulation.Through simulation of the proposed method,the performance just reduces 1dB under the same SNR.When two kinds of BPSK and QPSK signal about the aliasing power ranges from-7 to 5 dB,we can achieve very high recognition accuracy,so the algorithm has good robustness.In this paper,many aspects need further research,such as channel problems,aliasing signal recognition measurement problem,and so on.