Model-Driven Learning-Based Physical Layer Authentication for Mobile Wi-Fi Devices

Yijia Guo; Junqing Zhang; Yao-Win Peter Hong; Stefano Tomasin

cs.LGMar 20, 2026

Model-Driven Learning-Based Physical Layer Authentication for Mobile Wi-Fi Devices

Authors: Yijia Guo, Junqing Zhang, Yao-Win Peter Hong, Stefano Tomasin

Organizations: School of Computer Science and Informatics, University of Liverpool, Liverpool, L69 3DR, United Kingdom · Institute of Communications Engineering, National Tsing Hua University, Hsinchu, Taiwan 300044 · Department of Information Engineering, University of Padova, Padova, Italy

Abstract

The rise of wireless technologies has made the Internet of Things (IoT) ubiquitous, but the broadcast nature of wireless communications exposes IoT to authentication risks. Physical layer authentication (PLA) offers a promising solution by leveraging unique characteristics of wireless channels. As a common approach in PLA, hypothesis testing yields a theoretically optimal Neyman-Pearson (NP) detector, but its reliance on channel statistics limits its practicality in real-world scenarios. In contrast, deep learning-based PLA approaches are practical but tend to be not optimal. To address these challenges, we proposed a learning-based PLA scheme driven by hypothesis testing and conducted extensive simulations and experimental evaluations using Wi-Fi. Specifically, we incorporated conditional statistical models into the hypothesis testing framework to derive a theoretically optimal NP detector. Building on this, we developed LiteNP-Net, a lightweight neural network driven by the NP detector. Simulation results demonstrated that LiteNP-Net could approach the performance of the NP detector even without prior knowledge of the channel statistics. To further assess its effectiveness in practical environments, we deployed an experimental testbed using Wi-Fi IoT development kits in various real-world scenarios. Experimental results demonstrated that the LiteNP-Net outperformed the conventional correlation-based method as well as state-of-the-art Siamese-based methods.

Figures & tables

Fig. 1: The system model for PLA in a mobile scenario. Multipath varies due to Bob’s movement.

Fig. 2: The structure of LiteNP-Net.

Panel of Figure 3 — Fig. 3: (a) The embedding network $\Psi_{\rm A}$ . (b) The embedding networks $\Psi_{\rm B}$ and $\Psi_{\rm C}$ .

Fig. 4: The proposed LiteNP-Net PLA system.

Model	RMS Delay (ns)	No. Clusters	Mapped Environment
Model B	15	2	Residential apartment
Model C	30	2	Small office
Model D	50	3	Typical office
Model E	100	4	Large office
Model F	150	6	Large space

TABLE I: WLAN TGn Channel Models [ 39 ] .

Fig. 5: The AUC versus

\Delta t_{k}

on the simulation test dataset with WLAN TGn channel model F,

\text{SNR}=6

dB,

d_{\rm bm}/\lambda=0.25

and

v_{0}=1

m/s.

Fig. 6: The ablation comparison on the simulation test dataset with WLAN TGn channel model F,

\text{SNR}=6

dB,

d_{\rm bm}/\lambda=0.25

and

v_{0}=1

m/s.

Fig. 7: ROC curves on simulation test dataset with WLAN TGn channel model B, $\text{SNR}=12$ dB, $\Delta t_{k}=20$ ms, $d_{\rm bm}/\lambda=1$ and $v_{0}=1$ m/s.

Fig. 8: The AUC versus SNR on the simulation test dataset with WLAN TGn channel model B,

d_{\rm bm}/\lambda=1

\Delta t_{k}=20

ms and

v_{0}=1

m/s.

Fig. 9: The AUC versus

d_{\rm bm}/\lambda

on the simulation test dataset with WLAN TGn channel model B,

\text{SNR}=6

dB,

\Delta t_{k}=20

ms and

v_{0}=1

m/s.

	Parameter	FLOPs	Inference Time
LiteNP-Net	25,258	433,502	1.016 ms
CNN-based Siamese in [ 27 ]	57,108	2,141,480	1.390 ms
FCN-based Siamese in [ 28 ]	296,209	1,558,516	1.076 ms

TABLE II: Comparison of Computational Overhead.

Fig. 10: Experiment devices [ 27 ] . (i) ESP32 in the Wi-Fi AP mode was used as Alice, and (ii) two LoPy4 boards in the Wi-Fi station mode were used as Bob and Mallory.

Panel of Figure 13 — Fig. 11: Experiment scenarios [ 27 ] . (a) Scenario I: corridor A. (b) Scenario II: office (Moving route 1), and Scenario III: corridor B (Moving route 2).

Fig. 12: ROC curves on the test dataset collected in the scenario I (corridor A) with

d_{\rm bm}/\lambda=0.25

Panel of Figure 15 — Fig. 13: The AUC versus the distance between Bob and Mallory normalized by wavelength $d_{\rm bm}/\lambda$ in various experimental scenarios.

Yijia Guo (Member, IEEE) received the M.Eng. degree in information and communication engineering from Sun Yat-sen University, Guangzhou, China, in 2021. She is currently pursuing the Ph.D. degree with the School of Computer Science and Informatics, University of Liverpool, Liverpool, UK. Her current research interests include the Internet of Things, deep learning and physical layer authentication.

Table 16

Junqing Zhang (Senior Member, IEEE) received B.Eng and M.Eng degrees in Electrical Engineering from Tianjin University, China in 2009 and 2012, respectively, and a Ph.D. degree in Electronics and Electrical Engineering from Queen’s University Belfast, UK in 2016. From Feb. 2016 to Jan. 2018, he was a Postdoctoral Research Fellow at Queen’s University Belfast. From Feb. 2018 to Oct. 2022, he was a Tenure Track Fellow and then a Lecturer (Assistant Professor) at the University of Liverpool, UK. Since Oct. 2022, he has been a Senior Lecturer (Associate Professor) at the University of Liverpool. His research interests include the Internet of Things, wireless security, physical layer security, key generation, radio frequency fingerprint identification, and wireless sensing. Dr. Zhang is a co-recipient of the IEEE WCNC 2025 Best Workshop Paper Award. He is a Senior Area Editor of IEEE Transactions on Information Forensics and Security and an Associate Editor of IEEE Transactions on Mobile Computing.

Table 17

Y.-W. Peter Hong (S’01 - M’05 - SM’13) received the B.S. degree in electrical engineering from National Taiwan University, Taipei, Taiwan, in 1999, and the Ph.D. degree in electrical engineering from Cornell University, Ithaca, NY, USA, in 2005. He is currently a Distinguished Professor with the Institute of Communications Engineering and the Department of Electrical Engineering at National Tsing Hua University (NTHU), Hsinchu, Taiwan. His research interests include AI/ML for wireless communications, UAV and satellite communications, distributed signal processing for the IoT and sensor networks, and physical layer security. Dr. Hong received the IEEE ComSoc Asia-Pacific Outstanding Young Researcher Award in 2010, the Y. Z. Hsu Scientific Paper Award in 2011, the National Science Council Wu Ta-You Memorial Award in 2011, the Chinese Institute of Electrical Engineering (CIEE) Outstanding Young Electrical Engineer Award in 2012, and the National Science and Technology Council (NSTC) Outstanding Research Award in 2018 and 2022. He was the Chair of the IEEE ComSoc Taipei Chapter from 2017 to 2018. He was the Co-Chair of the Technical Affairs Committee, the Information Services Committee, and the Chapter Coordination Committee of the IEEE ComSoc Asia-Pacific Board, from 2014 to 2015, from 2016 to 2019, and from 2020 to 2021, respectively. He is now the Vice Director of the IEEE ComSoc Asia-Pacific Board from 2022 to 2027. In the past, he also served as Senior Area Editor and Associate Editor, respectively, for IEEE Transactions on Signal Processing, Associate Editor for IEEE Transactions on Information Forensics and Security and an Editor for IEEE Transactions on Communications. He was also a Distinguished Lecturer of the IEEE Communications Society from 2022 to 2023.

Table 18

Stefano Tomasin (Senior Member, IEEE) received the Ph.D. degree from the University of Padova, Italy, in 2003. During his studies, he did internships with IBM Research (Switzerland) and Philips Research (Netherlands). He joined the University of Padova, where he has been Assistant Professor (2005-2015), Associate Professor (2016-2022), and Full Professor (since 2022). He was visiting faculty at Qualcomm, San Diego (CA) in 2004, the Polytechnic University in Brooklyn (NY) in 2007, and the Mathematical and Algorithmic Sciences Laboratory of Huawei in Paris (France) in 2015. His current research interests include physical layer security, security of global navigation satellite systems, signal processing for wireless communications, synchronization, and scheduling of communication resources. He has been a senior member of IEEE since 2011 (member since 1999) and a member of EURASIP since 2011. He is or has been an Editor of the IEEE Transactions on Vehicular Technologies (2011-2016), of the IEEE Transactions on Signal Processing (2017-2020), of the EURASIP Journal of Wireless Communications and Networking (since 2011), and of the IEEE Transactions on Information Forensics and Security (since 2020). He also serves as a Deputy Editor-in-Chief of the IEEE Transactions on Information Forensics and Security since January 2023.

Table 19

Wireless Communications

Wireless communications studies how information is sent over radio links and how networks manage signals, spectrum, and resources reliably.

See all 391 papers

Channel Prediction-Based Physical Layer Authentication under Consecutive Spoofing Attacks
Mar 20, 2026Yijia Guo, Junqing Zhang, Yao-Win Peter HongChannel State Prediction Wireless Communications
School of Computer Science and Informatics, University of Liverpool, Liverpool, L69 3DR, United Kingdom · Institute of Communications Engineering, National Tsing Hua University, Hsinchu, Taiwan 300044
Adaptive Distributed Physical-Layer Authentication and Attack Detection in 6G Non-Terrestrial Networks via Causal Meta-Learning
Sep 8, 2026Parsa Rajabi, Mohammad Reza Abedi, Nader Mokari +26G Networks Meta-Learning
Department of Electrical and Computer Engineering, Tarbiat Modares University, Tehran, 14115-111, Iran · Department of Systems and Computer Engineering, Carleton University, Ottawa, K1S 5B6, Canada
Joint Deployment and Resource Management for Group-Tag-Based Physical-Layer Authentication and Collaborative Communication in NTN
Sep 8, 2026Parsa Rajabi, Mohammad Reza Abedi, Paeiz Azmi +1Wireless Communications Wireless Resource Allocation
the Department of Electrical and Computer Engineering, Tarbiat Modares University, Tehran, 14115-111, Iran

Mar 20, 2026J/K move · Enter open · S save

Channel Prediction-Based Physical Layer Authentication under Consecutive Spoofing Attacks

Yijia Guo, Junqing Zhang, Yao-Win Peter Hong

School of Computer Science and Informatics, University of Liverpool, Liverpool, L69 3DR, United Kingdom · Institute of Communications Engineering, National Tsing Hua University, Hsinchu, Taiwan 300044

Channel State Prediction Wireless Communications

Open paper AskPDF arXiv

Wireless networks are highly vulnerable to spoofing attacks, especially when attackers transmit consecutive spoofing packets. Conventional physical layer authentication (PLA) methods have mostly focused on single-packet spoofing attack. However, under consecutive spoofing attacks, they become ineffective due to channel evolution caused by device mobility and channel fading. To address this challenge, we propose a channel prediction-based PLA framework. Specifically, a Transformer-based channel prediction module is employed to predict legitimate CSI measurements during spoofing interval, and the input of channel prediction module is adaptively updated with predicted or observed CSI measurements based on the authentication decision to ensure robustness against sustained spoofing. Simulation results under Rayleigh fading channels demonstrate that the proposed approach achieves low prediction error and significantly higher authentication accuracy than conventional benchmark, maintaining robustness even under extended spoofing attacks.

Abstract

Wireless Communications

Explore similar work

Channel Prediction-Based Physical Layer Authentication under Consecutive Spoofing Attacks

Adaptive Distributed Physical-Layer Authentication and Attack Detection in 6G Non-Terrestrial Networks via Causal Meta-Learning

Joint Deployment and Resource Management for Group-Tag-Based Physical-Layer Authentication and Collaborative Communication in NTN