Funder: EPSRC Federated Telecoms Hub 6G Research Partnership Funds
PI: Dr Yuan Ding
Period: 1st Jul. 2024 ~ 31st Mar. 2025
Research Scope: Wireless communications have brought revolutionary transformation to our everyday lives, industry, etc, achieved by ubiquitous connectivity of massive wireless devices (21.7 billion in 2024) and exciting applications, e.g., smart homes, connected healthcare, autonomous driving, etc. The wireless infrastructure is traditionally protected by cryptography, which, however, faces difficulties to be applied to devices where cost and accessibility are limiting factors.

Radio frequency fingerprinting identification (RFFI) is an emerging non-cryptographic and physical-layer security technique to secure existing and future telecommunication infrastructure. It exploits unique and intrinsic hardware fingerprints of RF components for device identification, which does not rely on cryptography. Its implementation does not require any change to the transmitters and can be readily applied to existing and future radio communication systems. RFFI, however, is still in its nascent stage and there are fundamental research questions that remain unanswered, including
i) RF impairments sources are unclear and the methods for accurate modelling are yet available.
ii) The existing studies of deep learning-based RFFI are mainly empirical, and an optimised design is still missing.
iii) Existing RFFI evaluation is over-idealised and RFFI performance in practical scenarios has not been experimentally demonstrated.

This project will carry out a comprehensive investigation of RFFI to address the above remaining challenges. A synergistic approach will be adopted involving RFF modelling, deep learning-based protocol design and practical system-level validation. Specifically, we will dive into the RFF sources at device-/circuit-/signal-levels and then accordingly tailor the design of signal representations and deep learning models to fully expose and exploit dominant RFF features. This holistic approach will make a major difference from the empirical exploration in the existing works and allow us to achieve optimal performance. This research programme will apply RFFI to two use cases, i.e., WiFi and satellite communications, for demonstrating the disruptive potential of RFFI.