Modulation Recognition And Parameter Estimation Of Digital Signal Based On Convolutional Neural Network

In the non-cooperative condition,modulation identification and parameter estimation of communication signals are the preconditions for correct demodulation of signals.Nowadays,the communication environment is becoming more and more complex,and the modulation types of signals are gradually diversified,which will increase the difficulty of signal modulation identification and parameter estimation.With the improvement of computer computing power,deep learning algorithms gradually emerge in the field of artificial intelligence and begin to be applied in different fields.At present,applying deep learning technology to modulation recognition has become the mainstream research content in this field.Therefore,deep learning algorithm is used in this paper to conduct in-depth research on the recognition and parameter estimation of digital signal modulation mode.The main research contents are as follows:Deep learning can effectively deal with the current complex network environment through multi-level network structure and powerful feature learning capabilities.This paper analyzes intrusion detection methods and deep learning knowledge,and focuses on the research of intrusion detection methods based on convolutional neural networks.The main work of this paper is as follows:Aiming at the low recognition accuracy of digital Signal under low Signal Noise Ratio(SNR),modulation recognition algorithm based on convolutional neural network is improved in this paper,and CNN is combined with bidirectional long and short memory network and attention mechanism.First,the spatial features of the signal are directly extracted by CNN,and then the features are sent to BILSTM to further extract the temporal features of the signal.Finally,the attention distribution of the feature information is calculated by using the attention mechanism to distinguish the importance of the feature information and identify the modulation modes of 11 common digital signals.Simulation results show that the improved algorithm can improve the recognition accuracy of digital signal modulation under low SNR.In order to further verify the performance of the algorithm,the paper also makes a comparative analysis with classical machine learning algorithm and convolutional neural network algorithm,and finds that the algorithm proposed in this paper has a better overall recognition effect on the modulation mode of digital signal than the comparison algorithm.To solve the problem of large parameter estimation error of digital signals at low SNR,the existing methods are improved by combining CNN,BILSTM and attention mechanism to construct a regression algorithm model in this paper.Firstly,the spatial features of the signal are extracted by CNN,and then the temporal features of the signal are extracted by BILSTM.Finally,the carrier frequency and symbol rate of the above digital signals are estimated by using the regression model combined with the attention mechanism.Classical machine learning algorithm and convolutional neural network algorithm are also selected for comparative analysis.The results show that the estimation error of the proposed algorithm is smaller than the other algorithm at low SNR.