摘要
针对有标签样本较少条件下的通信信号调制识别问题,提出了一种基于伪标签半监督学习技术的小样本调制方式分类算法,通过优选人工特征集、设计高性能分类器以及基于输出概率的伪标签数据选择方法,构建高效的伪标签标注系统,然后通过该伪标签标注系统与基于深度学习的信号分类方法配合,实现在少量有标签样本和大量无标签样本条件下的调制方式分类。仿真结果表明,对6种数字信号进行调制识别,在信噪比大于5 dB时,伪标签算法可将模型识别性能提高5%~10%,该算法设计简单,具有较大应用价值。
In order to solve the problem of insufficient labeled samples in modulation recognition,this paper propo⁃ses a few⁃shot modulation recognition algorithm based on pseudo⁃label semi⁃supervised learning(pseudo⁃label algo⁃rithm).First of all,high quality artificial feature,excellent classifier and data⁃labeling method are used to build ef⁃ficient pseudo label system,and then the pseudo label system is combined with signal classification method based on the deep learning to realize the modulation classification under the condition of a small number of labeled sam⁃ples and a large number of unlabeled samples.The simulation results show that the pseudo⁃label algorithm can im⁃prove the model recognition performance by 5%⁃10%when the six kinds of digital signals are classified and identi⁃fied and its SNR is greater than 5 dB.At the same time,the algorithm has a simple network design and is of great application value.
作者
史蕴豪
许华
刘英辉
SHI Yunhao;XU Hua;LIU Yinghui(Institute of Information&Navigation,Air Force Engineering University,Xi′an 710077,China)
出处
《西北工业大学学报》
EI
CAS
CSCD
北大核心
2020年第5期1074-1083,共10页
Journal of Northwestern Polytechnical University
基金
国家自然科学基金(61601500)资助。
关键词
调制识别
伪标签
半监督学习
modulation recognition
pseudo⁃label algorithm
semi⁃supervised learning
simulation
