X-ray Pulsar Signal Recognition Based On Support Vector Machine

X-ray pulsar signal recognition is a significant problem to be solved in pulsar navigation.Power spectrum and bispectrum methods are traditional ways in signal recognition which are used widely.These methods take advantage of the statistics of the signal,and get good results.However,calculating Higher-order spectrum needs much time and memory,which restricts the practical application of these methods.What is more,cumulative profile should be calculated firstly.For the problem of pulsar signal recognition,this paper introduces a methodSupport Vector Machine(SVM).It can recognize the pulsar signal photos directly without calculating the cumulative profile,which simplifies the calculate process and system design.After analyzing the frame of support vector machine,this paper designs the recognition algorithm based on pulsar radiation and C-SVM model.The solution based on Sequential Minimal Optimization(SMO)method subjects to KKT factor.Simulation experiments show that SVM method has high recognition accuracy.Compared with one-dimension selected line spectra algorithm,SVM algorithm has a higher application value in practical.