Modulation Type Recognition Of Radar Signals Based On Cyclic Spectrum And Support Vector Machine

The electromagnetic environment of modern battlefield is complicated,and the means of electronic confrontation emerge in an endless stream.As an important part of electronic countermeasures,modulation type recognition of radar signals plays an important role in modern battlefields.Extracting radar signal features and selecting classifiers are two important links in modulation type recognition of radar signals,which are of great significance for modulation type recognition of radar signals.Therefore,the above two links are the focus of the research on modulation type recognition of radar signals.Firstly,the recognition rate of the single radar signals is verified by using the spectrum bandwidth and phase jump parameters of radar signals in this paper,but this method can not help to identify the hybrid radar modulation signals.For this reason,the cyclic spectrum method is used to extract the modulation type characteristics of radar signals,and the distance discriminant method is adopted to further optimize the extracted cyclic spectrum characteristics,and only one row of the most suitable cyclic spectral information is selected.The support vector machine is used as a classifier to identify the signals based on the cyclic spectral information after dimensionality reduction.The results show that the proposed method has a high recognition rate for radar modulation signals in low SNR environment.The main contents and innovative points of this paper are as follows:(1)The PCA(Principal Component Analysis)method is used as denoising methods for the radar signals received by the receiver in order to improve the recognition rate of the radar modulation signals.Based on the spectral smoothing method,we calculate the cyclic spectrum of radar signals.Mahalanobis distance is used to distinguish every row of all kinds of cyclic spectrum signals,and the row with the largest Mahalanobis distance is selected.(2)The LDA(Linear Discriminant Analysis)method is used to descend dimension for cyclic spectral features of multiple samples in order to reduce the computational complexity.The cyclic spectrum data after dimensionality reduction are divided into training set and test set.The data of the training set are used to train support vector machines,and the trained support vector machine model is used to classify the data of the test set into several known categories.5 kinds of single radar signals and 6 kinds of mixed radar signals are identified byusing support vector machine.Besides,the running time and recognition rate of the improved algorithm and the complete cyclic spectrum algorithm are compared.In this paper,MATLAB is used as a simulation tool for the improved algorithm of cyclic spectrum and support vector machine,the simulation results prove the feasibility of the improved method in theory.