Research On Radar Active Jamming Signal Recognition Technology

At present,with the continuous development of jamming technology,radar has to confront more and more jamming styles and face more and more arduous challenges.Radar needs to take certain anti-jamming measures.Any anti-jamming method can not resist all jamming types.Anti-jamming measures need to be selected according to specific jamming types.Therefore,the first step of anti-interference is to identify the type of jamming signal.Among jamming signals,active jamming has a greater impact on radar than passive jamming.Therefore,this paper will study the identification of radar active jamming signal.From the perspective of feature extraction,this paper will solve the problem of low recognition accuracy of active jamming signals under low JNR,and solve the problem that the active jamming signals superimposed in time domain can not be recognized effectively from the perspective of no manual feature extraction.Firstly,the mathematical model of active jamming is established in this paper.This paper introduces the classification method of active jamming signals,establishes the mathematical models of spectrum dispersion jamming,intermittent sampling direct repeat forwarding jamming,intermittent sampling and forwarding jamming,RF noise jamming,noise amplitude modulation jamming,noise frequency modulation jamming,convolution smart noise jamming and product smart noise jamming,and obtains the time-domain and frequency-domain distribution of jamming signals through simulation.Then,the characteristics of active jamming signal are analyzed in this paper.The characteristics of different active jamming signals are analyzed from time domain,frequency domain and pulse compression domain respectively,and the characteristic parameters with discrimination are extracted under different JNR.On this basis,the difference of characteristic parameters of different interference signals is analyzed,and different jamming signals are distinguished by the difference of characteristic parameters.Secondly,based on the feature extraction of jamming signal,an jamming recognition algorithm based on SSA-SVM is proposed.Through the optimization ability of sparrow search algorithm,the kernel function parameters and penalty factors of support vector machine are optimized to improve the classification performance of support vector machine.Simulation experiments are carried out to test the recognition accuracy of the algorithm under different JNR.Experiments show that the radar active jamming signal recognition algorithm based on SSA-SVM has good classification effect.Finally,aiming at the low recognition accuracy of radar active jamming signal in complex background,a recognition method based on attention mechanism is proposed.Firstly,the jamming signal is transformed into time-frequency,and the obtained timefrequency diagram is input into the Faster R-CNN network.Then the location of the jamming signal is suggested by the region in the Faster R-CNN network,and the region of interest is obtained.The region is input into the classifier to get the category of the jamming.Experiments show that it can not only recognize a single jamming signal,but also recognize multiple signals.