EVALUATING PRIVACY-PRESERVING TECHNIQUES AGAINST RADIO FREQUENCY FINGERPRINTING

Abstract

Radio Frequency Fingerprinting (RFF) has emerged as a promising physical-layer technique for identifying wireless devices based on hardware-induced signal variations. Unlike traditional authentication mechanisms, RFF does not rely on cryptographic keys and is particularly suitable for resource-constrained environments such as Internet of Things (IoT) systems and wireless sensor networks. However, the same characteristics that enable accurate device identification also introduce significant privacy risks, as RF emissions can be passively captured and used for persistent device tracking. This thesis investigates both the effectiveness and the privacy implications of RFF by analyzing device-specific signal patterns under controlled experimental conditions. A dataset consisting of multiple transmitters was collected using Software Defined Radio (SDR) platforms in both cable-based and over-the-air environments. To isolate hardware-induced features, different types of noise—including Gaussian, Uniform, Laplacian, and Impulse noise—were systematically introduced at varying levels. An autoencoder-based approach was employed to model the intrinsic characteristics of RF signals and evaluate reconstruction error across different scenarios, including clean signals, noisy signals, and signals from unseen devices. Experimental results show that while RFF can reliably distinguish between devices, it is also highly sensitive to noise and averaging effects. Furthermore, the findings demonstrate that carefully designed noise injection can reduce fingerprint separability, highlighting its potential as a privacy-preserving mechanism. Overall, this work provides a balanced analysis of RFF as both a security tool and a potential privacy threat, and explores noise-based obfuscation as a practical approach to mitigate unauthorized device tracking in wireless systems.

Date of Award	2026
Original language	American English
Awarding Institution	HBKU College of Science and Engineering