Research On Radar Deception Jamming Signal Recognition Technology

Nowadays,the radar active deception jamming technology has been developed rapidly.In order to achieve the normal operation of radar system,it is necessary to add anti-jamming of active deception into radar system.Radar can have excellent anti-jamming ability,so as to take the initiative and gain the initiative in the war.Therefore,the research on the method of anti-jamming against active deception jamming has become a hot topic.Some suitable anti-jamming approaches must be adopted to minimize the jamming on our radars.Firstly,it's necessary to accurately identify the specific type of jamming and provide information for anti-jamming.In this paper,two kinds of typical deception jamming,such as conventional pulling off jamming and dense fake target jamming,are studied.The specific works of this paper are as follows:This paper expounds the background and significance of radar active deception jamming recognition,and summarizes the current situation of domestic and foreign development in this field.The structure and the working mechanism of DRFM are introduced.The generation process and principle of deception jamming based on DRFM are studied in detail,and their mathematic signal models are also established.The radar receiving signal model under pulling off jamming is analyzed,and a conventional pulling off jamming recognition algorithm based on time frequency image TFSD and Rényi entropy feature is proposed.The algorithm extracts two features of time-frequency image of SPWVD,and uses the difference of two characteristic values in different situations to complete the recognition of pulling off jamming.A recognition method of deception jamming based on image Zernike moment feature of time-frequency distribution is proposed.The algorithm uses a series of image processing methods,Zernike moments are calculated as the feature which constitute feature vector for signal recognition.The radar receiving signal model under SMSP with C&I jammings are analyzed,a jamming identification method based on fractal dimension of two-dimensional feature spectrum is proposed.Three-dimensional bispectrum information is obtained by performing bispectrum analysis on radar echoes and a two-dimensional feature curve is generated to reduce the calculation cost by using a dimension reducing method.The two features of the fractal dimension of the two-dimensional feature spectrum are extracted as the feature which constitute feature vector for signal recognition.A jamming identification method based on shape feature of the two-dimensional feature spectrum is proposed.The two-dimensional feature spectrum is transformed into a gray image,after a series of image processing,Zernike moment feature is taken as image shape feature to constitute feature vector for signal recognition.Theory of deep learning is applied to radar active deception jamming recognition field.A recognition algorithm of pulling off deception jamming is proposed based on stacked sparse autoencoder.In this method,SPWVD of received radar signal under jamming is given firstly,and dimensionality reduction is implemented with a series of image processing methods.In the phase of pre-training,stacked sparse autoencoder model is trained with unlabeled samples by greedy layer-wise training.On this basis,network parameters are fine-tuned with label information.Finally,the softmax classifier is used to recognize the active jamming.A recognition algorithm of SMSP and C&I jamming is proposed based on stacked sparse autoencoder.In this method,bispectrum analysis of received radar signal under three cases is given firstly,and after a series of dimensionality reduction,extracting bispectrum singular vector and bispectrum along Y axis projection,and use the stacked sparse autoencoder model to complete the identification of sparse dense false target jamming.