Blind Modulation Classification Based On Symbol Rate Estimation

With the increment of wireless communication users number,the communication technology rapidly and revolutionary develop.To meet the communication demand of users,various of modulation modes and signal format has proposed in the field of communication,however,which bings about some problems.The spectrum resources become more and more scarce,and more than one kind of modulation signal are used at the same time,which affects the communication quality badly.To impove the situation,an effective means of spectrum supervision is essential.So,in this paper,we research some related algorithm and propose the algorithm of blind modulation classification based on symbol rate estimation.The contribution of this paper can be summarized as follows.1)For signal symbol rate estimation,a method based on matched filter is proposed and another method use the signal cyclostationary features,which has a better performance than the former.2)Based on the survey and analysis of existing signal modulation classification algorithm's advantages and disadvantages,a algorithm of blind modulation classification based on symbol rate estimation is proposed,which uses the decision tree algorithm to select the category features and the neural network tool to train the signal classifier.Most of test signals reach a classification rate of 95%under a SNR of 5 dB.A singal classification process based on symbol rate estimation and modulation classification is proposed.An actual signal test on the process is performed,the Agilent E4438C is used to send signal and the N6841A to receive.4)We summary the problems of signal recognition using cyclostationary features in actual signal measurement.Based on the researches in this paper,some research topcis,including signal classification in complex electromagnetic environment and optimization of the process of signal measurement need further study.