Scrambled Type Identification Based on Convolutional Neural Network Encoding
Aiming at the problem of recognizing the scrambled type of linear blocked code and convolutional code,a scrambled type recognition method using the combination of correlation features and shallow neural network is proposed.Firstly,the mutual correlation features of scrambled symbols are derived,and the biased autocorrelation function is introduced,and the combination of the two is used as the input correlation features;then on the basis of analyzing the correlation of the scrambled sequences,a shallow neural network model with high real-time performance is constructed;finally,the scrambled dataset is input into the network model,and the training and the recognition test of the network are completed.The simulation results show that compared with the traditional algorithm based on multiple fractal spectra,the proposed algorithm can recognize multiple types of scrambling,and at the same time,the proposed algorithm has stronger anti-bit error performance,which lays the foundation for further scrambled parameter recognition.
万方数据
维普
