一种基于复数残差网络的通信辐射源个体识别方法

A Method of Personal Identification of Communication Radiation Source Based on Complex-valued Residual Network

在复杂电磁环境的通信辐射源个体识别任务中,针对传统特征提取识别方法分类效果不佳和低信噪比环境下基于实数神经网络的方法识别准确率不高的问题,本文提出了一种基于复数残差网络的通信辐射源个体识别方法。将实际采集的I路和Q路电台数据组合成复数作为输入,根据电台数据集特点选取复数初始化方法、复数激活函数,以改进的复数残差块为基础构建复数残差网络,进一步调整和优化网络结构并运用到辐射源个体识别任务中。通过实验证明,相比于实数残差网络和人工特征提取方法,复数残差网络的性能更优,并且在低信噪比的条件下,基于复数残差网络的方法鲁棒性更强。

In complex electromagnetic environment,a method of individual identification of communication emitter based on complex residual network was proposed in this paper.The aim was to solve the problems of poor classification effect of traditional feature extraction and low identification accuracy of real neural network in low SNR environment.In this paper,we first combined the data collected from the I-channel and Q-channel into a complex number as input,and selected the complex initialization method and activation function according to the characteristics of the radio data set.Then we constructed the complex-valued residual network based on the improved complex-valued residual block,and finally applied the further optimized network structure to the task of individual identification of communication emitter.Experimental results show that,the complex-valued residual network has better performance,compared with the real residual neural network and the artificial feature extraction method.Moreover,the method based on complex-valued residual network is more robust under the condition of low SNR.