Study On Theory And Method Of Radar Unconventional Active Barrage And Deception Jamming Intelligent Recognition

In the modern radar electronic warfare,various novel types of radar unconventional active jamming,which have the effects of suppression and deception,lead to the radar not being able to or erroneously detect the target and its parameter.In order to ensure that the radar can exert its operational effectiveness under the new electromagnetic environment,it is indispensable to carry out research on jamming recognition technology.Radar unconventional active barrage and deception jamming mainly include noise productive jamming,sparse noise convolutional jamming,interrupted-sampling repeater jamming,interrupted-sampling circularly jamming,and partial-pulse dense transmitted jamming.This thesis focus on the theory and the method of radar unconventional active barrage and deception jamming intelligent recognition.The main work of this thesis is summarized as follows:Firstly,aiming at current radar electronic warfare environment,the characteristics of radar unconventional active jamming signals are studied,and unconventional active barrage jamming and deception jamming are classified into slice retransmitted jamming and noise modulation jamming,which lays a foundation for feature extraction and jamming intelligent recognition.Secondly,the concept of the graph domain is introduced,and two methods(visibility graphs and recurrence plots)of converting the radar signal from the time domain to the graph domain are given.For the visibility algorithm,we theoretical prove that the degree distribution of the simple pulse signal satisfies the single-spike distribution,and the degree distribution of the chirp signal satisfies the multi-spike distribution.The degree distribution characteristics of radar received signals and unconventional active jamming signals are analysed,which provides a theoretical basis for the feasibility of extracting features from the graph domain.Thirdly,some features on graph domain such as degree,clustering coefficient,path length,correlation,centrality and entropy are introduced.Simulation experiments have analyzed the validity of average degree,average clustering coefficient,Newman assortativity coefficient,and normalized network-structure entropy for radar unconventional-active-jamming recognition.Moreover,the feasibility of using these graph domain features is verified for collected data.Fourthly,an active jamming detection method based on correlation domain is proposed.The method samples the non-zero position of the radar received signal and the transmitted signal,and designs the Wiener filter with the minimum mean square error(MMSE)criterion.The signal power loss ratio before and after the filter are used as the characteristic parameter.Setting an appropriate threshold to perform the radar active jamming detection,and this method improves the operational efficiency of the radar anti-jamming system.Fifthly,two kinds of classifiers,random forests and nuclear Fisher discriminator,are designed.Intelligent recognition of radar unconventional active jamming is accomplished by using four features on graph domain.Simulation experiments and performance analysis show that the former has a better recognition performance,but the computational cost is larger,and the latter is the opposite.