Research On Key Technology Of Radar Emitter Intra-pulse Features Analysis And Recognition

Radar emitter recognition is a process of feature extraction of pre-sorted pulses and determination of radar emitter types by comparison and analysis with known signal feature parameters in the database.It is one of the core tasks of radar reconnaissance.However,with the rapid development of radar technology,a large number of new system radars have been put into equipment application,the signal environment is getting worse.The conventional feature parameters are far from meeting the needs of modern combat electronic intelligence.By acquiring more abundant and essential feature parameters,the analysis of radar intra-pulse features has become a research hotspot and development trend of radar emitter recognition,which is of great significance to radar emitter recognition.This paper mainly focuses on the key technology of radar emitter intra-pulse features analysis and recognition,and proposes corresponding algorithms for modulation recognition,modulation parameter estimation and individual idetification,and designs a verification system for radar emitter intra-pulse features analysis and recognition.The main work includes:1?Aiming at the problem of feature extraction difficulty and low recognition accuracy in the modulation recognition of complex multi-class radar emitter signals under low signal-to-noise ratio(SNR),a radar emitter modulation recognition algorithm based on time-frequency analysis and dilated residual network is proposed.The algorithm uses the time frequency transformation choi-williams distribution(CWD)to transform the radar emitter signal one-dimensional time domain waveform into a two-dimensional time frequency image,and after the time frequency image is preprocessed,the dilated residual network is constructed to automatically extract the signal time frequency image features for modulation recognition.Simulation results show that the proposed algorithm has good anti-noise performance,strong generalization ability,and improves the recognition accuracy of liner frequency modulation(LFM)signal with similar time-frequency image features.2?Aiming at the limitation of application of efficient fractional fourier transform(FRFT)to realize LFM signal parameter estimation,a fast LFM signal parameter estimation algorithm based on power spectrum analysis and efficient FRFT is proposed.Based on the rough estimation of the signal bandwidth and center frequency using the power spectrum,the algorithm uses dynamic selection of rotation angle,decentralized frequency and gaussian smoothing filter to improve the efficient FRFT to effectively estimate the normalized FRFT length,and then uses the geometric relationship of FRFT under different rotation angles to realize the fast estimation of LFM signal parameters.Simulation results show that the proposed algorithm is universal,has good anti-noise performance,high estimation accuracy and good real-time processing performance.3?Aiming at the problem of poor classification performance in radar emitter individual idetification based on unintentional phase modulation on pulse,a method of radar emitter individual idetification based on unintentional phase modulation on pulse analysis and long and short time memory plus full convolution network is proposed.In this algorithm,the observed phase is firstly de-modulated and bezier smoothed to accurately extract the unintentional phase modulation on pulse feature curve,and then constructs a long and short-term memory plus full convolution network to automatically extract the joint features of the unintentional phase modulation on pulse sequence to achieve individual identification of the radar emitter.Simulation results show that the proposed algorithm has good anti-noise performance,strong generalization ability,and good idetification effect in small samples.4?Aiming at the shortage that emphasizes theoretical research and light engineering application in the filed of radar emitter recognition,a radar emitter intra-pulse features analysis and recognition verification system is designed.The system uses different waveform generators to simulate the radar emitter to emit pulse signals,and uses real-time spectrum analyzer as a signal acquisition platform for data acquisition and storage.Finally,the measured data verifies that the algorithm in the previous chapter is feasible and effective,and has certain engineering application value.