By the use of the extended homogenous balance method,the B(?)cklund transformation for a (2+1)- dimensional integrable model,the(2+1)-dimensional Nizhnik-Novikov-Veselov (NNV) equation,is obtained,and then the NNV equ...By the use of the extended homogenous balance method,the B(?)cklund transformation for a (2+1)- dimensional integrable model,the(2+1)-dimensional Nizhnik-Novikov-Veselov (NNV) equation,is obtained,and then the NNV equation is transformed into three equations of linear,bilinear,and tri-linear forms,respectively.From the above three equations,a rather general variable separation solution of the model is obtained.Three novel class localized structures of the model are founded by the entrance of two variable-separated arbitrary functions.展开更多
针对低信噪比(signal to noise ratio,SNR)低截获概率(low probability of intercept,LPI)雷达脉内波形识别准确率低的问题,提出一种基于时频分析、压缩激励(squeeze excitation,SE)和ResNeXt网络的雷达辐射源信号识别方法。首先通过Cho...针对低信噪比(signal to noise ratio,SNR)低截获概率(low probability of intercept,LPI)雷达脉内波形识别准确率低的问题,提出一种基于时频分析、压缩激励(squeeze excitation,SE)和ResNeXt网络的雷达辐射源信号识别方法。首先通过Choi-Williams分布(Choi-Williams distribution,CWD)获得雷达时域信号的二维时频图像(time-frequency image,TFI);然后进行TFI预处理降低噪声干扰和频率维的位置分布差异,以适应深度学习网络输入;最后在ResNeXt基础上加入扩张卷积和SE结构提取TFI特征,实现雷达辐射源分类。实验结果表明,SNR低至-8 dB时,该方法对12类常见LPI雷达波形的整体识别准确率依然能达到98.08%。展开更多
文摘By the use of the extended homogenous balance method,the B(?)cklund transformation for a (2+1)- dimensional integrable model,the(2+1)-dimensional Nizhnik-Novikov-Veselov (NNV) equation,is obtained,and then the NNV equation is transformed into three equations of linear,bilinear,and tri-linear forms,respectively.From the above three equations,a rather general variable separation solution of the model is obtained.Three novel class localized structures of the model are founded by the entrance of two variable-separated arbitrary functions.
文摘针对低信噪比(signal to noise ratio,SNR)低截获概率(low probability of intercept,LPI)雷达脉内波形识别准确率低的问题,提出一种基于时频分析、压缩激励(squeeze excitation,SE)和ResNeXt网络的雷达辐射源信号识别方法。首先通过Choi-Williams分布(Choi-Williams distribution,CWD)获得雷达时域信号的二维时频图像(time-frequency image,TFI);然后进行TFI预处理降低噪声干扰和频率维的位置分布差异,以适应深度学习网络输入;最后在ResNeXt基础上加入扩张卷积和SE结构提取TFI特征,实现雷达辐射源分类。实验结果表明,SNR低至-8 dB时,该方法对12类常见LPI雷达波形的整体识别准确率依然能达到98.08%。
