Research Of Broadband Spectrum Detection Technology And Its FPGA Implementation

In non-cooperative communication,the receiver is totally blind to the received signals.Besides,due to lacking accurate prior information,it is difficult for the receiver to design a reliable receiver structure and demodulation algorithm.Meanwhile,the generation and application of various modulation methods,as well as the variability of modulation parameters,make non-cooperative communication face great challenges.Therefore,the research on signal detection,modulation parameter estimation and modulation pattern recognition in non-cooperative communication is of great significance.Based on the background of non-cooperative communication,this thesis studies broadband spectrum detection technology,which includes broadband signal detection,parameter estimation and modulation mode identification.The parameters to be estimated are bandwidth,carrier frequency and symbol rate.In practical engineering application,according to the principle of spectrum detection in this thesis,FPGA is used to detect the spectrum of broadband signal with a bandwidth of 1.2GHz.Then,cooperate with the host computer to complete the design and verification of the whole broadband detection system.The wideband signal detection method is based on power spectrum detection.In this thesis,Welch method is used to estimate the signal power spectrum.Simulation result shows,compared with direct periodic graph method,the power spectrum estimated by Welch method has a smaller variance and more smoother.Similarly,the rough estimation of carrier frequency and bandwidth is also based on the signal power spectrum.The rough estimation results provide direction and basis for the subsequent signal processing.In order to obtain accurate carrier frequency,on the basis of rough estimation,for conventional modulation signals,the M-power spectrum of MPSK,the square spectrum of MSK and GMSK are used to make fine estimation of carrier frequency.At the same time,combined with CZT algorithm,the spectrum is refined and the estimation error further reduced.According to the M-power spectral characteristics of conventional modulation signals,combined with the relationship between spectral line information and symbol rate,the symbol rate of modulation signals can be estimated by the M-power spectral characteristics.The principle of modulation recognition is also based on the M-power spectrum characteristics of the signal.By extracting the number of spectral lines and calculating the flatness coefficient,the modulation recognition of BPSK,QPSK,OQPSK,8PSK,MSK and GMSK can be completed.During the implementation of the broadband spectrum detection system,in order to effectively process the 1.2GHz bandwidth signal,the analog front-end uses a four channel frequency converter to divide the broadband RF signal into four channels.After ADC sampling by FMC105 board card,the digital signal processing is realized on the platform of the large capacity signal processing card U7.The internal architecture of FPGA includes channelization,power spectrum estimation,signal detection,down conversion and feature extraction.Finally,in order to verify the system function,dual vector signal sources are used to simulate the single channel and dual channel input of the broadband signal respectively.Through the joint debugging with the upper computer,it is found that the real-time spectrum and detection results displayed normal in the upper computer,which shows that the function of the broadband spectrum detection system is correct.