A Comprehensive Survey on Physical Layer Authentication Techniques: Categorization and Analysis of Model-Driven and Data-Driven Approaches

作者：Zhifan Lai, Zikai Chang, Mingrui Sha, Qihong Zhang, Ning Xie, Changsheng Chen, Dusit Niyato · 发表于：ACM Computing Surveys · 年份：2024 · DOI：10.1145/3708496 · 被引用次数：11 · 研究领域：Physical Unclonable Functions (PUFs) and Hardware Security、Electrostatic Discharge in Electronics、Cellular Automata and Applications

The open and broadcast nature of wireless mediums introduces significant security vulnerabilities, making authentication a critical concern in wireless networks. In recent years, Physical-Layer Authentication (PLA) techniques have garnered considerable research interest due to their advantages over Upper-Layer Authentication (ULA) methods, such as lower complexity, enhanced security, and greater compatibility. The application of signal processing techniques in PLA serves as a crucial link between the extraction of Physical-Layer Features (PLFs) and the authentication of received signals. Different signal processing approaches, even with the same PLF, can result in varying authentication performances and computational demands. Despite this, there remains a shortage of comprehensive overviews on state-of-the-art PLA schemes with a focus on signal processing approaches. This article presents the first thorough survey of signal processing in various PLA schemes, categorizing existing approaches into model-based and Machine Learning (ML)-based schemes. We discuss motivation and address key issues in signal processing for PLA schemes. The applications, challenges, and future research directions of PLA are discussed in Part 3 of the Appendix, which can be found in supplementary materials online.