Research On DRFM Active Deception Jamming Recognition Algorithms

With the continuous development of radar jamming technology,especially the emergence of Digital Radio Frequency Storage(DRFM),jammers have been able to intercept radar signals in a short time for modulating and transmitting jamming signals that are highly coherent with the corresponding radar signals,for which it is more difficult for radars to work efficiently.Therefore,the research on radar jamming recognition algorithm is an important prerequisite to counter active radar jamming.At present,the research on radar jamming recognition algorithm is still in its infancy and significant for further research.In this paper,combined with the structure of DRFM jammers,the DRFM active deception jamming recognition algorithm is studied.The main work of the research includes the following sections:The principle and characteristics of active deception jamming based on DRFM technology are studied in the first part.Firstly,the model of DRFM jammer is introduced,and the harmonic distortion characteristics of jamming are studied by taking Analog-to-Digital Converters(ADC)and Digital-to-Analog Converters(DAC)in DRFM jammers as examples.Secondly,to courter the ranging principle of pulse radar,the principle of range deception jamming is analyzed by simulating the range false target jamming.Finally,to courter the velocity measurement principle of Pulse Doppler(PD)radar,the principle of velocity deception jamming is analyzed by simulating repeater jamming with the frequency shift.Taking the harmonic characteristics of the digital controlled phase shifter of DRFM jammer into consideration,a velocity deception jamming recognition algorithm based on Singular Spectrum Analysis(SSA)is proposed in the second part.Firstly,the working model of digital controlled phase shifter of DRFM jammer is studied and its harmonic parasitic characteristics in the Doppler domain are simulated and analyzed.Secondly,the velocity deception jamming recognition algorithm of DFRM based on SSA is explained in detail.In this algorithm,SSA is used to decompose the Doppler spectrum of the signal in the same distance unit to obtain the singular value sequence.Then,the variance,kurtosis,skewness,energy and entropy of the statistical histogram of the singular value sequence are extracted to construct the eigenvector.Therefore,the Support Vector Machine(SVM)is used as a classifier to recognize the radar target and the DRFM deception jamming signal.The algorithm no longer relies on the existing harmonic model of ADC phase quantization and can effectively identify the jamming especially at a low jamming-to-signal ratio(JSR).The simulation verifies the effectiveness of the algorithm.Taking the Doppler characteristics of the delay-superimposed jamming generated by the analog frequency shifting method of DRFM jammer into consideration,a delay-superimposed jamming recognition algorithm based on Convolutional Neural Network(CNN)is proposed in the third part.Firstly,the signal model of delay-superimposed jamming is studied,and the difference between jamming and the real multi-target in the range-Doppler image that is processed by radar PD algorithm is simulated and analyzed.secondly,the main structure of CNN is studied,and the CNN structure of this algorithm is constructed according to the recognition task.Then,the training algorithm principle and derivation formula of CNN are studied.Finally,the delay-superimposed jamming recognition algorithm based on CNN is introduced.Specifically,the algorithm preprocesses the range-Doppler image of the received signal in the first stage.The constructed CNN is trained via the generated samples in the second stage.The jamming recognition rates at various signal-to-noise ratios(SNRs)are obtained by recognizing the test samples via the well-trained CNN in the last stage.In this algorithm,the deep learning theory is introduced into the field of jamming recognition and the jamming can be recognized effectively especially at low SNRs.The simulation verifies the effectiveness of the algorithm.