Research And Implementation Of Radar Signal Intra-pulse Modulation Feature Identification

Due to the rapid development of modern electronic technology,the traditional radar signal analysis and processing technologies have not been able to recognize the low probability of interception(LPI)radar.Based on this issue,it is necessary to analyze the feature of signal intra-pulse modulation.With the increase of signal density in the electromagnetic environment,multiple signals may exist in the same channel at the same time,which will affect the signal recognition accuracy.Therefore,how to effectively analyze and extract the intra-pulse modulation features is a key issue in current signal recognition.In this paper,we start with the recognition of intra-pulse modulation characteristics of radar signals,and study the identification method of six kinds of radar signals,including conventional radar signal,binary phase code signal,quardrature phase code signal,linear frequency modulation signal,nonlinear frequency modulation signal and binary frequency code signal.Therefore,the main content of this paper is:Firstly,based on the principle of more emphasis on hardware implementation,this paper studies two single signal recognition methods based on feature extraction.One of the proposed radar signal modulation identification methods is based on the characteristics of spectrum and instantaneous frequency.The method takes two steps.The first step is to divide the six kinds of signals into two categories: phase modulation signal and frequency modulation signal.The second step is to subdivide the phase modulation signal and the frequency modulation signal.The phase modulation signal is distinguished according to the characteristics of the discrete spectral lines which is obtained by the method of signal square,and the frequency modulation signal is distinguished according to the instantaneous frequency characteristics of the signal.The other identification method is based on spectral complexity of radar signal,the signal spectrum is quantified firstly,and then the Lempel-Ziv complexity of quantized spectral sequence is calculated.According to the different complexity of each signal,a tree classification process is designed.Both methods are simulated and complementary.Secondly,considering the problem of simultaneous existence of multiple signals in the same channel,this paper studies a multi-signal identification method based on signal separation.The key of this method is signal separation.The performance of signal separation is related to whether it is effective to extract signal modulation characteristics within the pulse.Therefore,joint approximation diagonalization algorithm is used as signal separation algorithm,which has stable performance of separation.Then,according to the separation signal,a method based on feature extraction is adopted to identify the multipie signals one by one.The recognition probability statistics of five groups of mixed signals under different SNRs are simulated.The simulation results show that the method has good separation effect and high recognition accuracy.Finally,based on the signal identification method of spectrum and instantaneous frequency characteristics,a signal segmentation recognition method based on DSP is proposed.A detailed introduction of the original data segmentation interception,segmentation signal digital orthogonal mixing transform,low-pass filtering and decimation,and segmentation signal spectrum and time-frequency features extraction are presented orderly.The flow of DSP is given,and the recognition probability under different SNRs is simulated by MATLAB.Several sets of parameters are tested based on signal source.Simulation and test results show that the method has good practicability.