Radar Signal Recognition Based On Improved Quantum Particle Swarm Optimization Convolution Neural Network

Because of the increasingly dense and complex modern electromagnetic environment,the rapid development of electronic information technology and the emergence of new system radars,Traditional radar signal recognition methods have been unable to meet the needs of modern electronic warfare.Therefore,it is necessary to study and explore more advanced radar signal recognition technology,so that it can extract more refined,more representative and universal radar signal characteristics to adapt to the development of electronic warfare in modern warfare.Convolutional neural network models are widely used in the field of deep learning,due to its strong robustness and fault tolerance,parallel processing capability,self-learning,self-organization,self-adaptation,strong fitting,efficient feature extraction convolution,the computational and dimensionality reduction pooling operations,as well as the powerful information synthesis capabilities,have gradually become a new direction in radar signal recognition technology research.However,the traditional convolutional neural network training algorithm often adopts the error back propagation algorithm,which makes the network fall into local optimum,and the convergence speed is slow,the generalization ability cannot be guaranteed,and the calculation model is more complex.Therefore,this paper proposes an autonomously improved quantum particle swarm optimization algorithm to overcome the shortcomings of the backpropagation algorithm,Combined with the convolutional neural network model,it could be used in the field of radar signal identification.The main work is as follows:(1)The structure of basic neural network and Convolutional Neural Network(CNN)is analyzed.The advantages and disadvantages of Standard Particle Swarm Optimization(PSO)and Quantum-behaved Particle Swarm Optimization(QPSO)are discussed.(2)Based on the QPSO algorithm,a differential evolution operator and adaptive factor are introduced,and an improved Quantum-behaved Particle Swarm Optimization Algorithm(IQPSO)is presented.The improvement of the IQPSO algorithm is verified by the test function comparison experiment.Advantage.(3)Combining IQPSO algorithm with CNN,the specific coding strategy,parameter design method and algorithm flow chart are given,and the performance of IQPSO-CNN is tested by referring to the data in UCL database.(4)The simulation of radar signals of different systems is studied.The simulation data of radar signals of different systems are obtained through simulation experiments.A certain amount of data is selected as the training sample of IQPSO-CNN.The remaining data is used as the test sample of IQPSO-CNN to verify the network identification.Correct rate.