基于LSTM和残差网络的雷达有源干扰识别

针对目前雷达干扰识别方法存在人工特征提取难、强噪声环境下识别率不高的问题,提出了一种基于长短时记忆(long short-term memory, LSTM)网络和残差网络相结合的雷达有源干扰识别方法。输入有源压制干扰原始时域序列数据,搭建深度学习网络模型对不同干噪比下的干扰信号进行分类识别。仿真结果表明：在干噪比为0 dB的情况下,该方法对4类雷达有源干扰信号的识别准确率均高于98.3%,与单纯的残差网络和卷积神经网络(convolutional neural networks, CNN)等其他深度学习算法相比,具有更佳的网络性能,验证了该算法的有效性。

Radar active jamming recognition based on LSTM and residual network

Aiming at the problem that the current radar jamming recognition method is difficult to extract artificial features and the recognition rate is not high in a strong noise environment, a radar active jamming recognition method combining long short-term memory (LSTM) network and residual network is proposed. The original time domain data of active suppression jamming is inputt, and a deep learning network model to is built classify and identify jamming signals under different jamming-to-noise tatio (JNR). The simulation results show that when the JNR is 0 dB, the recognition accuracy of the method for four types of radar active jamming signals is higher than 98.3%, which is compared with other deep learning algorithms such as pure residual network and CNN-LSTM, it has better network performance, which verifies the effectiveness of the algorithm.