Research On Radar Active Jamming Signal Disturbance Classifiction Base On Feature Extraction

With the increasing status of radar in modern warfare,as an important part of electronic warfare,radar confrontation is becoming increasingly fierce.At the same time,with the form of radar jamming,the various composite jamming is more and more complex,and the jamming effect is more and more remarkable.In order to achieve better anti-jamming purposes,radar active jamming as the main form of jamming,for its classification and identification of radar anti-jamming areas,need to study.Based on the development of radar anti-jamming technology,under the new situation,this thesis focuses on the research of radar active jamming and the identification of radar active composite jamming,which lays the foundation for improving the anti-jamming capability of radar.The main research contents of this thesis include the modeling of radar active jamming signal and composite jamming signal,and the classification and recognition of radar active composite jamming based on feature extraction.The main work is as follows:1.Based on the existing theoretical knowledge and related papers,the mechanism of active radar jamming is studied in conventional radar and a simulation analysis is made in this thesis.This thesis discusses the way of compound jamming generation,which is different from that of different forming methods and the jamming type are studied emphatically.It is found that the composite interference can be the best for the combined interference of velocity gate pull jamming and cover interference on composite interference effect.2.Based on the modeling of composite jamming,this thesis extracts the characteristics of the signal of the additive jamming of the velocity gate pull jamming and the radio frequency jamming,the additive jamming of the velocity gate pull jamming and the amplitude modulation jamming and the additive jamming of the velocity gate pull jamming and the frequency modulation jamming The characteristics of the extracted features are as follows: mean,variance,frequency domain kurtosis,frequency domain skewness,spectral symmetry,box dimension,approximate entropy,fractional low order moment and so on.In order to facilitate comparison,the target signal is also mentioned.Under different jamming to noise ratio conditions,these characteristic parameters are extracted and analyzed.Through the comparison and analysis of the simulation experiment,we can obtain the separability and anti-noise performance of different features.The separability and good anti-noise characteristic parameters,such as frequency domain moment and frequency domain skewness,through vector construction,are good for improving the recognition rate and recognition quality.3.Two algorithms for composite jamming classification recognition are introduced in this thesis: random forest algorithm and softmax classification algorithm,respectively,and bases on these two classification algorithm to build a composite jamming classification classification.The classification and recognition ability of the two kinds of classifiers is verified by radar simulation.When the jamming to noise ratio is relatively large,the two classifiers have a good ability to identify composite jamming.Between them,the classifier based on the algorithm of random forest has a better classification and recognition ability on the composite jamming.