摘要
自动调制识别是认知无线电、智能解调器、电子侦察等各种民用及军事应用的基本需求。使用USRP B210采集8种调制类别空中接口IQ数据,训练分类模型实现USRP B210接收空中接口IQ数据实时输出调制类型。为了提高自动调制识别的准确度,优化ResNet、GoogLenet、SENet用于自动调制识别。实验结果显示在低信噪比条件下自动调制识别分类准确度有了较大提升。
Automatic modulation recognition is a basic requirement for various civilian and military applications such as cognitive radios,intelligent demodulators,and electronic reconnaissance.The USRP B210 was used to collect 8 kinds of modulation class air interface IQ data,and the training classification model was trained to realize real-time output modulation type of USRP B210 receiving air interface IQ data.In order to improve the accuracy of automatic modulation recognition,ResNet,GoogLenet and SENet were optimized for automatic modulation recognition.The experimental results show that the accuracy of automatic modulation recognition and classification is greatly improved under the condition of low SNR.
作者
刘桥平
高兴宇
邱昕
郭瑞
Liu Qiaoping;Gao Xingyu;Qiu Xin;Guo Rui(Institute of Microelectronics of the Chinses Academy of Sciences,Beijing 100085,China;University of Chinese Academy of Sciences,Beijing 100049,China)
出处
《计算机应用与软件》
北大核心
2020年第8期110-114,121,共6页
Computer Applications and Software
基金
国家自然科学基金项目(61702491)
装备预研教育部联合基金项目(6141A020223)。
