DETECTION OF REPLAY/RELAY ATTACKS IN WIRELESS SYSTEMS USING RF-FINGERPRINTING AND SDR

Abstract

This paper systematizes methods for detecting Replay and Relay attacks in wireless systems based on radio-frequency fingerprinting (RF-fingerprinting) of the signal using Software-Defined Radio (SDR) receivers. It is shown that both attack types preserve data correctness at the protocol level, so their detection is only possible through analysis of physical hardware artifacts of the transmitter and propagation channel rather than cryptographic or timing protocol features. A comparative analysis of SDR platforms (RTL-SDR V3, HackRF One, ADALM-PLUTO, USRP B210) relevant to Replay/Relay detection is performed. A generalized, protocol-independent detection architecture is proposed, combining an SDR receiver, an IQ preprocessing module, an RF-fingerprint extractor, and a channel-analysis module that evaluates propagation delay and relay-induced distortions in parallel. The detection task is formalized as metric classification in embedding space combined with two additional criteria — payload correlation for Replay and a channel-anomaly indicator for Relay. A decision algorithm is developed that classifies a signal as original, Replay, Relay, or belonging to an unknown device. Analysis of available experimental results confirms the practical feasibility of the proposed approach, with reported accuracy of 95–99% for original/attack discrimination on both budget and high-end SDR hardware.