一种基于IQ数据时域功率图的通信辐射源识别方法

A Method for Specific Emitter Identification of Communication Radiation Sources based on Time-Domain Power Graph of IQ Data

通信辐射源个体识别在民用和军事中应用广泛,主要涉及特征参数提取和识别分类方法两方面问题。随着深度学习技术的发展,卷积神经网络(Convolutional Neural Network,CNN)在图片分类识别上已经具有强大能力。为了发挥CNN对图像的优越识别分类特性,提出了一种利用IQ数据得出时域功率图,并对时域功率图进行识别的方法。时域功率图包含不同辐射源个体的IQ不平衡特征,具有个体差异性,能达到辐射源识别的效果。通过实验,该方法在普通电台上可达到93%的识别率。对比双谱特征,该方法有更好的识别能力。实验结果表明,该方法在手持机识别上具有较强的泛化性。

Individual identification of communication radiation sources is widely used in civil and military applications.It mainly consists of two aspects:Extraction of characteristic parameters and identification method.With the development of deep learning,CNN(Convolutional Neural Network)has a strong ability in image classification and recognition.In order to take advantage of the superior characteristics of CNN in recognition and classification,a method to obtain and identify the time-domain power graph by using IQ data is proposed.Because the time-domain power diagram can contain the IQ imbalance characteristics of individuals with different radiation sources,which has individual differences and can achieve the effect of radiation source identification.Through experiment,the recognition rate of this method can reach 93%on ordinary radio station.Compared with bispectral features,this method has better recognition ability.Experimental results indicate that this method has strong generalization in mobile phone recognition.