Research On Evaluation Method Of Radar Anti-jamming Effectiveness

With the development of modern electronic countermeasure technology,the fight between radar interference and anti-interference has become increasingly fierce.In the face of increasingly complex electronic interference environments,radars must improve their anti-jamming capabilities and adopt effective anti-jamming measures for radar systems.Only in this way can the radar be able to ensure the initiative of the battlefield in a complex electromagnetic environment,and thus exert real combat effectiveness and bring positive influence to the war situation.The established assessment indicators have good completeness and independence only by acquiring typical sample data,and when the selected assessment method is appropriate,the radar anti-jamming performance can be efffectively evaluated.Based on the above three requirements,this article proposes the following innovative methodsFirst,the radar interference test method.Based on the diversity of the values of the experimental variables of radar countermeasures and the complexity of constraints between the variables,the article proposes the concept of bi-partite covering array to construct the experimental scheme.The scheme requires that each pair of radar and jammer parameters occurs at lea-st once in the testing scheme.Second,the radar anti-jamming effectiveness measurement.Based on the analysis of the characteristics of a variety of jammers and radar performance factors,combined with the existing indicators,this article uses the principal component analysis method to obtain two comprehensive indicators.This method avoids the one sidedness of previous performance evaluation indicators.Third,radar anti-jamming effectiveness evaluation.It is different from the early eval-uation factor method and the later fuzzy comprehensive evaluation method.This paper proposes a machine learning algorithm,establishes a,radar anti-jamming effectiveness e-valuation model,and compares the model of decision tree,random forest,xgboost and SVM.At last,the paper selects the algorithm with the best generalization ability to build the final model.