| 引用本文： |
-
谢葳,王红军,马春来,等.基于高斯原型网络的物联网设备物理层认证策略[J].电讯技术，2026，66（8）： - . [点击复制]
- XIE Wei,WANG Hongjun,MA Chunlai,et al.Physical Layer Authentication Scheme of Internet of Things Devices Based on Gaussian Prototype Network[J].，2026，66（8）： - . [点击复制]
|
|
| 摘要: |
| 随着5G、6G等先进通信技术的提出和演进，物联网技术得到了飞速的发展和应用，同时也面临着安全风险的挑战。物理层认证技术作为面向未来通信场景的关键物理层安全技术，可用于对无线物联网设备进行身份认证，确保网络的安全和隐私的保护。针对目前传统策略需要大量合法设备信道状态信息（Channel State Information，CSI）样本的问题，提出了基于高斯原型网络（Gaussian Prototype Network，GPN）的物理层认证策略。该策略首先基于元学习使得GPN具备特征提取能力，进而基于GPN利用合法设备的CSI幅度构建合法设备的高斯原型，最终根据待测设备利用CSI幅度提取的特征向量与合法设备的高斯原型之间的马氏距离实现物理层认证。通过真实采集数据实验表明，所提策略在1-shot的情况下达到了0.98的准确率。 |
| 关键词: 物联网设备 物理层认证 物理层安全 高斯原型网络（GPN） 马氏距离 |
| DOI：10.20079/j.issn.1001-893x.250313002 |
|
| 基金项目:国家自然科学基金面上项目（61971473，62071234） |
|
| Physical Layer Authentication Scheme of Internet of Things Devices Based on Gaussian Prototype Network |
| XIE Wei,WANG Hongjun,MA Chunlai,LI Xinhao |
| (College of Electronic Countermeasure，National University of Defense Technology，Hefei 230037，China) |
| Abstract: |
| With the proposal and evolution of advanced communication technologies such as 5G and 6G，the Internet of Things(IoT) technology has been widely developed and applied，while it also faces the challenge of security risks.Physical layer authentication，a key physical layer security technology for future communication scenarios，can be used to authenticate wireless IoT devices to ensure network security and privacy.In order to solve the limitation that the traditional schemes require extensive channel state information(CSI) samples from legitimate devices，a physical layer authentication scheme based on Gaussian Prototype Network(GPN) is proposed.The scheme firstly enables GPN to extract feature by meta-learning，and then constructs Gaussian prototype of the legitimate device based on the GPN from CSI amplitude.Finally，physical layer authentication is performed according to the Mahalanobis distance between the feature vector extracted from the testing device’s CSI amplitude and the Gaussian prototype of the legitimate devices.Experiments with real collected data show that the proposed scheme achieves an accuracy of 0.98 in the case of 1-shot. |
| Key words: Internet of Things device physical layer authentication physical layer security Gaussian prototype network(GPN) Mahalanobis distance |
《电讯技术》编辑部 