Radar Signal Parameter Detection And In-pulse Identification

Both radar systems and radar signal modulation types are pursuing increasing complexity,especially the low probability of intercepting radar signals as a typical example,which puts forward higher requirements for the detection,reception and analysis of radar signals.This thesis explores the realization of channelized reception,detection,pulse parameter detection,intra-pulse modulation recognition and intra-pulse parameter estimation of enemy radar signals.The main research work and innovations of this thesis are summarized as follows:1.Thesis establishes a model for 12 types of radar signals,which are LFM,SFM,FSK,BPSK,QPSK,Frank,P1,P2,P3,P4,LFM-BPSK and FSK-BPSK.Then,the two time-frequency analysis methods of Wigner-Ville distribution(WVD)and Choi-Williams distribution(CWD)are deeply studied,and the advantages and disadvantages of the two time-frequency conversion methods are analyzed and summarized in combination with the WVD transformation and CWD transformation of 12 radar signals2.Aiming at the problem that the reconnaissance receiver intercepts multiple signals at the same time,a digital channelized and efficient receiving structure based on the polyphase filter structure is deduced and studied,and the advantages and disadvantages of several existing signal detection algorithms are discussed.Finally,the measurement method of pulse description words is given,and the relative error estimation performance curves of each parameter are obtained through simulation experiments.3.In this thesis,the intra-pulse modulation recognition of radar signal is studied from two aspects: the traditional method and the method based on neural network.First of all,the feasibility of using integral quadratic phase function(IQPF)and fractional Fourier transform(FRFT)to recognize LFM,DLFM,Frank,P1,P2,P3,P4 and LFMBPSK signals is studied,and a modulation recognition framework is designed.Simulation results show that the overall recognition of this method reaches 90% when SNR is-4d B.Then,combined with the CWD time-frequency analysis basis of 12 kinds of radar signals,such as LFM,DLFM,Frank,P1,P2,P3,P4,LFM-BPSK,NP,FSK,BPSK,QPSK and FSK-BPSK,the intra-pulse modulation recognition method of radar signal based on residual network is given.Simulation results show that,the overall recognition can reach92.13% when SNR is-4d B.4.In-depth study of the FRFT algorithm to achieve the parameter estimation of the LFM signal.For the parameter estimation of polyphase code based on IQPF and FRFT,the estimation error of FRFT spectral ridge interval is large at low signal-to-noise ratio,,an improved algorithm based on wavelet packet denoising is proposed.The simulation results show that the parameter estimation performance is significantly improved before and after denoising.For LFM-BPSK Hybrid Signal,if the BPSK is eliminated by square first,and then the modulation parameters of LFM are estimated,the squaring process will reduce the SNR and lead to the decline of parameter estimation performance.In this thesis,A New Method for Parameters Estimation of LFM-BPSK Hybrid Signal in Low SNR is proposed,Firstly,through IQPF estimating the frequency rate,and then,with the help of this parameter,the original LFM-BPSK hybrid signal is dechirped.Finally,the improved MAT algorithm is used to estimate the modulation parameters of BPSK.The parameter estimation performance curves are obtained by simulation,which proves that the proposed method performs well under low signal-to-noise ratio.