王梓楠,
东南大学信息科学与工程学院 南京 211189
详细信息
作者简介:冯熳：女，1979年生，副教授，主要研究方向为超窄带通信、无线携能通信、军事抗干扰通信、智能通信信号处理等
王梓楠：男，1993年生，硕士生，主要研究方向为通信信号处理、机器学习
通讯作者:王梓楠 183742104@qq.com
中图分类号:TN911.72计量
文章访问数:1430
HTML全文浏览量:954
PDF下载量:111
被引次数:0
出版历程
收稿日期:2019-04-08
修回日期:2020-03-03
网络出版日期:2020-04-15
刊出日期:2020-11-16
Interference Recognition Based on Singular Value Decomposition and Neural Network
Man FENG,Zinan WANG,
College of Information Science and Engineering, Southeast University, Nanjing 211189, China
摘要
摘要:无线通信中的抗干扰技术对通信的稳定性和安全性都具有重要意义，干扰识别作为抗干扰技术的重要环节一直是研究的热点。该文提出一种基于奇异值分解与神经网络的干扰识别方法，该方法只计算信号矩阵的奇异值即完成特征提取，与传统方法相比节省了多个谱特性的计算量。仿真结果表明：基于奇异值分解与神经网络的干扰识别方法与传统方法相比在干信比为0 dB左右的条件下识别准确率有10%～25%的提高。
关键词:干扰识别/
神经网络/
奇异值分解
Abstract:The anti-interference technology in wireless communication is great significance to the stability and security of communication. As an important part of anti-interference technology, interference recognition is a research hotspot. An interference recognition method based on singular value decomposition and neural network is proposed. This method only calculates the singular value of the signal matrix as the feature. Compared with the traditional method, it saves the computational complexity of multiple spectral features. The simulation results show that the recognition accuracy based on singular value decomposition and neural network is 10%~25% higher than the traditional method under the condition of jamming-signal ratio at 0 dB.
Key words:Jamming recognition/
Neural network/
Singular value decomposition
PDF全文下载地址:
https://jeit.ac.cn/article/exportPdf?id=d384cef1-9886-473b-ac4d-c1f32accc8fe
