Research On Characteristic Analysis And Feature Extraction Of Radar Signal Based On NYFR

Along with wide applications of new technologies in radar system development,the radar reconnaissance reciver faces the gradually complex electromagnetic environment.The features that complicated signal form,ultra-wide bandwidth of signal,making modulation recognition,parameter estimation,and multi-signal processing in the radar reconnaissance receiver be a tremendous challenge.Using higher-speed analog-to-digital converter(ADC)can not fundamentally solve the problem of ultra-wideband signal detection,sampling techniques must be improved from the radar reconnaissance receiving system.The Nyquist Folding Receiver(NYFR)is a new radar reconnaissance structure that utilizing modulated local oscillator(LOS)to fold ultra-wideband signals into a Nyquist zoon(NZ)to achieve full-band interception.NYFR has a clear physical meaning,which using low sampling rate ADC to uniformly sample,breaking the performance bottleneck of the ADC and improve the dynamic range of receiver.Based on four typical radar signals,this thesis deeply analyze the principle of interception of the ultra-wide band signal in NYFR prototype structure.Then,under the sampling architecture,this dissertation discusses the folding signal characteristics,feature extraction and modulation recognition in detail.The main contributions of this dissertation are as follows:1.Aiming at the problem of ultra-wide band signal interception,this dissertation researches the feature of NYFR prototype local oscillator,using nonuniform theory and Bessel function to explain the periodically nonuniformity and principle of spectral broadening of NYFR respectively.Demonstrate the problems that should be considered in structure parameter setting and give the setting rules qualitatively.2.When the input signals are the monopulse(MP),linear frequency modulation(LFM),binary phase shift keying(BPSK)signals and binary Frequency shift keying(BFSK),the mathematic models are studied and established based on NYFR.For the folding signals,we deeply analyze the signal characteristic from time domain,frequency domain and time-frequency domain,include instantaneous autocorrelation,Fourier transform(FT),short-time fourier transform(STFT)and so on,which lays the foundation for signal feature extraction and modulation recognition.3.We use the instantaneous autocorrelation and the instantaneous frequency(IF)extraction based on short-time Fourier transform(STFT)to realize the intra-pulse modulation recognition of several typical radar signals.The first algorithm is based on the periodicity of NYFR,which is divided into specific delay instantaneous autocorrelation function and non-specific delay nstantaneous autocorrelation function.This algorithm has the advantages of small computation and low complexity,but can not recognize the BFSK signal.In IF extraction algorithm,several characteristic parameters are extracted from IF curve,LFM and BFSK are identified well when the SNR(signal to noise ratio)is above-5dB.MP and BPSK are correctly identified at 90% probability with SNR greater than 5 d B.4.Based on the modulation recognition,the NZ information acquisition algorithms of folding signal are proposed,which are named Bessel function recursive matrix(BERM)algorithm and sliding windows combined with NZ matching algorithm.In the BERM,we use the principle that spectrum amplitude of folding signal is consistent with Bessel function value distribution to construct the amplitude matrix and then calculate the NZ.In the sliding window and matching algorithm,the modulation characteristic of LOS is used to construct a matching function,and then we odtain the NZ by modifying the sliding window.