Radar Behavior Recognition Based On Predictive State Representation And Active Jamming

With the rapid development of electronics and information technology,electronic reconnaissance technology has become a key factor affecting the direction of war.Electronic reconnaissance undertakes very important tasks in the process of electronic warfare,such as obtaining enemy intelligence and judging threat levels.Among them,it is more important to recognize the different working states of more and more new system radars appearing on the battlefield.In order to characterize the dynamic game and interference confrontation between the two sides in the battlefield,this paper mainly studies the problem of identifying the target radar behavior under the condition that the scout party actively sends out interference.On the basis of studying radar behavior and radar interference,symbolize the radar pulse signal of each working state of the radar to obtain the radar pulse sequence.Through the Predictive State Representations(PSR)analysis of the radar pulse sequence,the analysis result of a s ingle working state is obtained.On this basis,the behavior identification of the radar test sequence in different scenarios is completed based on the Bayesian classifier.First,the radar behavior is classified based on the interference environment,and the radar behavior hierarchical model is established.Based on the research on the evaluation of the interference effect of radar radiation sources,the symbolic representation method of radar signals is studied,and the radar string sequence in the unit of radar word is obtained as a data model for subsequent PSR analysis and radar behavior identification.Secondly,using the Markov property of the radar pulse sequence,the PSR analysis method of the discrete-time dynamic system of the radar pulse sequence is studied.Due to the complex structure of the radar pulse symbol and the large amount of data,algorithms such as string processing,noise reduction and dimensionality compression are introduced in the analysis process to reduce the amount of calculation of the algorithm and enhance the robustness of the algorithm;based on the characteristics that the interference action is sent by oneself,Restricts some concepts when PSR is used in the radar field,and simplifies the calculation at the same time.Finally,the PSR processing results of the radar word sequence of the target radar in each working state when the reconnaissance party launches active interference are obtained,and the influence of different sequence lengths on the processing results of the algorithm is analyzed.Finally,the PSR analysis results are further studied.In this paper,the Bayesian maximum posterior probability classifier is used to realize radar behavior recognition based on PSR.Analyzing the recognition effect of radar sequences in different scenarios,the recognition accuracy rate is as high as 98% under ideal conditions,while the recognition accuracy rate of unknown states under active interference is close to 80%,indicating that this paper uses PSR and Bayesian classifiers to recognize radar behavior The advantages.Thus,a complete set of methods and procedures for identifying the working state of the radar sequence under the condition of active interference is established.