Radar Signal Modulation Recognition Based On Machine Learning

The modulation recognition of radar signals plays a vital role in electronic reconnaissance,and the accurate modulation recognition is the powerful guarantee for the implementation of radar countermeasures.With the increasing complexity of battlefield environment,the modulation types becomes changeable and the number of radar signals is limited under the condition of non-cooperation,which leads to the low recognition rate at low signal-to-noise ratio.At the same time,with the continuous emergence of new radar systems,the modulation recognition of unknown signals has become an urgent problem to be solved.This paper studies the recognition of known radar signal,the discrimination of unknown radar signal and the recognition of unknown radar signal with time-frequency analysis and machine learning.Firstly,the basic theories of radar signal recognition are introduced.Nine common radar signal models are constructed,and the feature extraction method of time-frequency analysis is studied.On this basis,the swarm intelligence optimization algorithm is improved,and the performance of single classifier and multi classifiers is discussed.Secondly,the recognition of known radar signal is studied.The theory of transfer learning is analyzed,and a new eigensmap network is designed based on the mechanism of filtering transformation.In order to improve the recognition rate under low signal-to-noise ratio,this paper proposes the known radar signal recognition algorithm based on stacked autoencoder feature fusion.The simulation results show that the performance of the proposed algorithm is better than the other existing algorithms.Thirdly,the discrimination of unknown radar signal is studied.The model of convolutional autoencoder is analyzed,which is trained by time-frequency images.Combined with mean squared error,an unknown radar signal discrimination algorithm based on reconstruction error is proposed,and the reliability of the algorithm is verified.Finally,the recognition of unknown radar signal is studied.Combined with transfer learning,the time-frequency feature of unknown radar signal is extracted.The kernel principal component analysis and affinity propagation are introduced,and the improved swarm intelligence optimization algorithm is applied to optimize the cluster centers.The recognition algorithm of unknown radar signal based on unsupervised learning is proposed,and the simulation verifies the feasibility of the proposed algorithm.At the same time,this paper gives out the scheme of radar signal modulation recognition system,which can realize the recognition of known radar signal,the discrimination of unknown radar signal and the recognition of unknown signal at low signal-to-noise ratio.The simulation results verify the performance of the proposed system.The research explores the combination of machine learning and modulation recognition.