Research On Electronic Jamming Identification Method Based On Time Frequency Domain Analysis

New type of electronic interference in modern warfare severely affects radar detection and tracking performance,Accurate and efficient identification of various types of new electronic interference has always been the focus and difficulty of ECM research.This paper aims at a new type of electronic interference identification classification problem,and proposes to use a decision tree classification method and a BP neural network recognition algorithm to identify and classify new types of electronic interference.It also focuses on the electronic interference recognition method for deep learning based on the time-frequency domain characteristics analysis.This article focuses on electronic interference feature extraction and identification and classification issues.The main research contents include the following:1?The mathematical models of eight typical electronic disturbances are analyzed and their characteristics in time domain,frequency domain and time-frequency domain are studied.The principle of various types of interference is studied and the corresponding simulation analysis is given to provide a theoretical basis for interference identification.2?This paper studies the feature extraction methods of eight kinds of electronic interference principles in different signal space such as time domain,frequency domain,wavelet domain and waveform domain.Study the classification and classification of various types of interference based on the method of decision tree and BP neural network,and give a detailed analysis of the simulation results.3?Aiming at the problem of interference identification,the deep learning method based on time-frequency domain analysis is further studied to identify and classify 8 types of electronic interference.Firstly,we analyze the time-frequency domain of eight kinds of electronic interferences and obtain the time-frequency domain image of each type of interference.The obtained images were input into the convolutional neural network as training and test samples,then the recognition probability of each type of jamming is obtained and a detailed simulation analysis was given.Finally,comprehensively compare the characteristics of the three methods.The results show that the convolutional neural network can extract the interference features automatically and efficiently,and the recognition probability is obviously higher than the former two methods.