Research On Detection And Parameter Estimation Methods Against Interrupted-sampling Repeater Jamming Based On Deep Learning

Interrupted-Sampling Repeater Jamming(ISRJ)is a novel type of radar active coherent jamming based on digital radio frequency memory.With its versatility,superior jamming performance,realistic electronic dummy target construction,and ease of engineering application,ISRJ has emerged as a new challenge in the realm of current radar anti-jamming,becoming a threat for radar detection.Researchers and radar designers are working to develop more effective anti-ISRJ technologies.This thesis focuses on the two core issues of intelligent detection and parameter estimation in radar anti-ISRJ pilot work as the theoretical basis and technical support for the radar intelligent jamming countermeasures.The following are the key research findings of this thesis.1.The ISRJ mechanism and performance are investigated in detail.The signal of ISRJ is modeled,and the jamming effect is summarized.De-chirp processing is introduced under the Linear Frequency Modulation(LFM)signal radar,and the jamming characteristics of ISRJ de-chirped ISRJ signals,as well as the jamming differences under different ISRJ jammer parameters,are summarized.The implications of major parameters of ISRJ jammers,such as slice width,on the jamming performance is addressed to build a theoretical foundation for jamming detection and parameter estimation.2.The ISRJ intelligent detecting method is investigated.A stacked bidirectional gated recurrent unit network based on gated recurrent unit(GRU)is designed to yield the radar de-chirped echo signal’s signal component labels at each sample point in the time domain,count the number of jamming component signal labels among them,and determine the presence of ISRJ when it exceeds a particular threshold.This method overcomes the low accuracy of traditional jamming detecting algorithms in low signal-to-noise ratio scenarios,hence broadening the application range of ISRJ detection.Besides,the entire process runs more quickly,it has a good chance of being used in real-time applications.3.The ISRJ parameter estimation with great precision is investigated.To accomplish higher accuracy,a time-frequency enhancement network named TFCNN based on convolutional neural network(CNN)and time-frequency analysis(TFA)is built to enhance the time-frequency distribution of de-chirped echo signal,and the parameters of ISRJ are estimated using the enhanced time-frequency distribution.This method tackles the problem of traditional parameter estimation methods not being accurate enough to estimate the operational parameters of ISRJ directly from the original time-frequency distribution under low signal-to-noise ratio scenarios,considerably improving the performance of existing anti-ISRJ technologies...