Reasearch On Technology For Wideband Spectrum Monitor

With the large-scale commercial use of the fifth-generation mobile communication(5G),its technology and business continue to mature,setting off a new wave of wireless communication industry,but also bringing new challenges.In face of the situation that the electromagnetic spectrum of the future wireless communication system will become more abundant and the electromagnetic environment will become more complex,the intelligent spectrum monitor is required.With the rapid development of Software Definition Radio(SDR)technology and machine learning algorithms,the integration of these two technologies with the wideband RF front-end can realize intelligent spectrum monitoring in the future.This paper focuses on the research on technology for wideband spectrum monitor.The main work of this paper is as follows:Firstly,a high-performance frequency synthesizer for spectrum monitor is developed.The frequency synthesizer has the advantages of low phase noise,high frequency resolution,large bandwidth and high stability.The output frequency can range from 45 MHz to22.4GHz,and the resolution is less than 1Hz.The results show that when the output frequency of the frequency synthesizer is 20 GHz,the phase noise is-106.52 d Bc/Hz@100k Hz.Secondly,a wideband RF receiver front-end for spectrum monitoring is proposed.The spectrum monitor is mainly composed of wideband RF receiver front-end and SDR processor.The wideband RF front-end module adopts a superheterodyne architecture with segmental filtering and parallel processing.The operating frequency is 6-18 GHz,which realizes the expansion of the SDR processor frequency band.The results show that the gain flatness of the wideband RF receiver front-end is less than 2.5d B,and dynamic range of input powert covers from-85 d Bm to-30 d Bm.When receiving QPSK,16 QAM,64QAM,and 256 QAM signals with bandwidth of 50 MHz,the EVM of the wideband receiver front-end is better than 3.52%rms,2.04%rms,1.73%rms,and 1.67%rms in the full frequency band respectively.When receiving 5G NR standard signals with bandwidth of400 MHz,the RF receiver front-end still achieves good performance,with the EVM of2.09%rms,1.87%rms and 1.56%rms at 8GHz,12 GHz and 16 GHz,respectively.Finally,an intelligent spectrum monitor based on a convolutional neural network(CNN)model is constructed by MATLAB and combined with the SDR module.After training and testing different simple modulation signals,the intelligent spectrum monitoring system realizes the recognition of RF modulation mode signal.