Research On Radar Jamming Perception And Its GPU Realization

With the continuous development of radar technology,how to carry out efficient and accurate anti-jamming method has gradually become a hot topic for radar workers.Jamming perception technology distinguishes jamming types by different characteristics between different jamming to help radars adopt corresponding anti-jamming measures to suppress jamming signals,which can effectively improve the anti-jamming effect.Therefore,jamming perception technology has also been placed in an increasingly important position.In view of the above reasons,the jamming perception technology is introduced in detail in this thesis.Due to the large number of types of radar jamming,this thesis first summarizes and classifies all kinds of jamming according to different attributes.The theoretical formulas and simulation waveforms of the three types of repressive interference and three types of deceptive interference mainly studied in this thesis are given.Its statistical characteristics and time domain and frequency domain characteristics were also analyzed specifically.On this basis,feature extraction is performed on the time domain and frequency domain features of jamming,including time domain moment skew and time domain moment kurtosis.The entropy theory in thermodynamics was also extended to jamming perception technology.Based on four kinds of entropy theories,including shannon entropy and exponent entropy,were used to extract features,and they were simulated and analyzed in detail.The types of jamming that can be distinguished using specific feature extraction methods are analyzed based on the differences between different jamming after feature extraction.Then,using the feature extraction method based on entropy theory,three kinds of common classifiers,Naive Bayesian classifier,BP neural network classifier,and random forest classifier,were used to perceive three types of repressive jamming and three types of deceptive jamming.The recognition rate of each classifier is obtained,and the recognition rate and stability of the classifier are analyzed in detail.By comparing the classification effects of the three classifiers,the applicability of Na?ve Bayes classifier in this paper is obtained.Finally,the jamming perception algorithm based on naive Bayes classifier is transplanted to the GPU,and the GPU programming language CUDA C is used to realize the jamming perception algorithm,and it obtains satisfactory recognition rate,achieving the transition from theory to practice.