Research On Rapid Recognition Method Of Radar Jamming Signal Based On Deep Complex Network

With the continuous development of anti-jamming technology,the combat effectiveness of radar will face serious threats.As an important equipment for obtaining information in a war,the combat effectiveness of radar is a key factor affecting the outcome of a war.Therefore,in order to protect the combat effectiveness of our radar,there are higher requirements for the radar's anti-jamming capability.Accurate and rapid identification of radar jamming signals is a prerequisite for effective anti-jamming measures,which will directly affect the operational effectiveness of radars.In this paper,based on the shortcomings of the radar jamming signal recognition method based on the real number convolutional neural network,a radar jamming signal recognition model based on the complex convolutional neural network is constructed,and all the information of the complex radar jamming signal is fully extracted through operations such as complex convolution.The main content of the paper includes:First,research the basic theory of real number convolutional neural network,deeply understand the working principle of its key network structure,and be familiar with the training process of the network;for the real number convolutional neural network cannot fully extract all the information of complex data,lead to complex convolution Neural network theory,and in-depth understanding of the working principle of its important network structure.At the same time,in view of the overparameter problem of the deep learning model,the network pruning is modeled and analyzed,which lays the foundation for subsequent research.Secondly,combined with the characteristics of the radar interference signal used in the subject,design single-channel and dual-channel real number convolutional neural network models,and use real number convolutional neural networks to extract the information of the complex radar interference signal;in order to further extract all the information of the complex radar interference signal,The real number convolutional neural network is extended to the complex number domain,and a radar jamming signal recognition model based on the complex number convolutional neural network is proposed,and compared with the real number convolutional neural network and other traditional methods to verify the effectiveness of the proposed model.Finally,in order to meet the real-time requirements of radar jamming signal recognition,the layer-by-layer pruning and cross-layer pruning algorithms are proposed,and the radar jamming signal recognition model based on complex convolutional neural network is network pruned to reduce the amount of parameters and calculations.Then the radar jamming signal recognition model based on deep complex network is deployed on the Jetson Nano deep learning development board to verify the actual acceleration effect of the model and meet the real-time requirements of the radar jamming signal recognition task.