基于相位统计图的调相信号的智能调制识别

Intelligent Modulation Recognition of Phase Adjusting Signal Based on Phase Chart

在通信对抗中,调制样式识别为后续信号进行盲解调提供保障和条件。针对通信中的调相有关信号的调制样式识别问题,提出一种基于相位统计图结合卷积神经网络(CNN)的调制智能识别方法。相位统计图抓住调相的本质特征,并且结合卷积神经网络的特征提取和自学习能力,因此无需人工提取特征,能够在低信噪比时提高调制样式识别的精确率。在信噪比-20dB到20dB的范围,仿真生成五种常用相位调制有关的信号,并以相位统计图的形式保存。将图片输入到卷积神经网络模型中,能够快速进行调制分类,分类精确率高、效果好。仿真实验结果证明,在信噪比等于-6dB时,调制分类识别率能够达到95%以上。

In communication confrontation,modulation style recognition provides guarantee and condition for blind demodulation of subsequent sig⁃nals.Aiming at the modulation style recognition problem of phase modulation related signals in communication,a modulation intelligent recognition method based on phase statistics combined with convolution neural network(CNN)is proposed.the phase graph captures the es⁃sential features of the phase modulation and combines the feature extraction and self-learning capabilities of the convolutional neural net⁃work.therefore,without the need for manual feature extraction,the accuracy rate of modulation style recognition can be improved at low sig⁃nal-to-noise ratio.in the range of signal-to-noise ratio-20 dB to 20 dB,the simulation generates five signals related to the common phase modulation and is saved in the form of phase statistics.the images are input into the convolutional neural network model,which can quickly perform modulation classification,and the classification accuracy is high and the effect is good.Simulation results show that the recognition rate of modulation classification can reach more than 95%when the SNR is equal to-6 dB.