Research On Classification And Recognition Algorithm Of Radar Signal Based On RNN

The classification and identification of radar emitter signals is an important part of electronic countermeasures.With the development and advancement of science and technology,new complex radars have gradually joined the battlefield,and the electromagnetic environment has become more complicated,which has brought unprecedented challenges to electronic countermeasures.Traditional radar emitter signal recognition technology can no longer meet the needs of modern battlefields.Therefore,more advanced classification algorithms must be sought to solve the problem of radar emitter signal identification.In recent years,deep learning has been widely used in all walks of life.Artificial intelligence has gradually played an important role in human society.Therefore,the introduction of deep learning into radar radiation source signal recognition complies with the requirements of the development of the times and also makes this problem get an effective solution.In this paper,a radar radiation source classification and recognition algorithm based on recurrent neural network is studied.The algorithm does not need to perform time-frequency transform on the signal,but directly applies the original radar radiation source signal,which simplifies the signal recognition process,greatly improves the signal recognition efficiency,and has broad application prospects.The main work of this paper is as follows:1.By studying the modeling method of recurrent neural network and the commonly used derivative network,a method of applying sequence model to radar signal recognition task is found.A radar radiation source classification and recognition algorithm based on Bi-directional Long Short-Term Memory network is proposed.The method effectively solves the shortcomings of the traditional recurrent neural network that cannot memorize long-term information,and by using the bidirectional structure,fully extracts all the information of the input sequence,and obtains a good classification recognition effect.In this paper,different modulation methods and the same modulation methods of radar radiation source signals are simulated,and the proposed algorithm is compared with other deep learning algorithms and traditional signal classification methods.The results show that the proposed method can obtain good recognition results with low signal-to-noise ratio and significantly improve the robustness of radar signal classification.2.Aiming at the complexity of the Long Short-Term Memory network,a radar radiation source classification and recognition algorithm model based on Bi-directional Gated Recurrent Unit is proposed.The construction adopts the hierarchical structure extraction feature built by the two-way gated loop unit,and finally obtains the classification result.The model replaces the forgetting gate and input gate in the Long Short-Term Memory network with an update gate,and combines the hidden state and the memory cell state value,which simplifies the network structure,reduces the model parameters,and reduces the memory footprint of the model and over-fitting risk.In this paper,the simulation experiment of the same radar emitter signal data set is carried out,and the proposed algorithm is compared with other deep learning algorithms and traditional signal classification methods.The results show that the method can obtain almost the same classification effect as the Bi-directional Long Short-Term Memory network under the same conditions.The method inherits the advantages of the Bi-directional Long Short-Term Memory network,effectively reduces the complexity of the model,accelerates the convergence speed.