Research On Analysis And Processing Of LPI Radar Signals Of LFM Class

In modern electronic warfare,the widespread use of low probability of intercept(LPI)radars has greatly increased the difficulty of acquiring electronic intelligence.It is especially important to analyze the intra-pulse part of linear frequency modulation(LFM)class signals,which are most used signal types of LPI radar.However,there are still great challenges: the time-frequency images(TFI)of LPI signals at low signal-to-noise ratios(SNR)are seriously disturbed by noise;the polyphase coded signals Frank codes,P1 codes,P2 codes,P3 codes and P4 codes are easily confused,which belong to LFM class signals;parameter estimation for LFM class signals based on TFI has high accuracy,but it requires a large amount of computation and has poor real-time performance.This thesis focuses on the intra-pulse intentional modulation of LFM class signals of LPI radars to address the above four aspects,the main work includes:1.In response to the existing algorithms cannot effectively distinguish noise and signal frequency when binarizing the grayscale TFI of LPI radar signals,K-means clustering algorithm is used to do binarization.The parameters of K-means algorithm are set to different values in different SNRs.F1-score is used to comprehensively consider the effect of binarization.The simulation experiments show that the binarization effect of the proposed algorithm is significantly better than that of the traditional Otsu method when SNR is lower than 2 dB.When SNR is higher than 2 dB,the F1-scores of both algorithms are close to 100%.The binarization effect of the proposed algorithm is excellent,which lays a solid foundation for the subsequent modulation type identification and parameter estimation of LPI radar signals based on binary images.2.Aiming at the LPI radar polyphase coded signals are easily confused,and the existed literatures rarely combined modulation type recognition and modulation parameter estimation.A feature extraction method based on time-frequency ridges is proposed.Based on the proposed features above,the modulation type identification is performed by the support vector machine classifier.At the same time,the modulation parameters including bandwidth,code length,carrier frequency and number of cycles of the carrier frequency per subcode(cpp)could be estimated from the extracted features.The simulation experiments show that when SNR is higher than 4 dB,the correct estimation rates of the coding length and cpp are both above 95%,and the average relative errors(RE)of the carrier frequency and bandwidth are stable around 2% and 4.3%respectively.Comparative experiments show that the proposed method outperforms the crosscorrelation method in the correct recognition rate of modulation types in the SNR range of 0 dB to 20 Db.Compared with the deep learning method in the case of small samples,the proposed method has smaller computation and higher recognition rate,which has certain application value.3.Aiming at the problem of large amount of computation when estimating the parameters of LFM class signals transmitted by LPI radars based on TFI,an iterative angle search(IAS)algorithm is proposed,which is used on Wigner Hough transform(WHT)and Radon Wigner transform(RWT)to reduce the computation.The IAS algorithm continuously narrows the angle search range and reduces the search step size through multiple cycles,and gradually approaches the real value of the parameter to be estimated.The SNR of simulation experiments is set from-4d B to 20 dB.Whether it is WHT or RWT,the IAS algorithm is combined with the same estimation accuracy as the original method when estimating LFM signal parameters,and the operation time is shortened to 3% of the original method.In addition,the IAS+WHT+RWT method is used to estimate the modulation parameters of the LFM signal through little amount of computation,the RE of the carrier frequency is reduced from 4%-4.5 to below 0.4% compared with RWT,and the RE of bandwidth is reduced from 0.7%-1.1% to 0.28%-0.7% compared with WHT.The modulation parameters of polyphase coded signals are effectively estimated,and the computing time is 5.14% of the traditional method.The IAS algorithm has the highest estimation accuracy and the shortest estimation time compared with other fast search methods and cyclostationary algorithm,which verifies that IAS algorithm has excellent performance when estimating modulation parameters for LPI radar LFM class signals.