Network-Centric Anomaly Filtering and Spoofer localization for 5G-NR Localization in LAWNs

Abstract

This paper investigates security vulnerabilities and countermeasures for the 3rd Generation Partnership Project (3GPP) Fifth Generation New Radio (5G-NR) Time Difference of Arrival (TDoA)-based unmanned aerial vehicle (UAV) localization in low-altitude urban environments. We first optimize node selection strategies under Air to Ground (A2G) channel conditions, proving that optimal selection depends on UAV altitude and deployment density, and propose lightweight User Equipment (UE)-assisted approaches that reduce overhead while enhancing accuracy. Next, we then expose critical security vulnerabilities by introducing merged-peak spoofing attacks where rogue UAVs transmit multiple 5G-NR Positioning Reference Signalss (PRSs) that merge with legitimate signals, bypassing existing detection methods. Through theoretical modeling and sensitivity analysis, we quantify how synchronization quality and geometric factors determine spoofing success probability, thereby revealing fundamental weaknesses in current 3GPP positioning frameworks. To address these vulnerabilities, we design a network-centric anomaly detection framework at the Localization Management Function (LMF) using 3GPP-specified parameters, coupled with recursive gradient descent-based robust localization that filters anomalous data while estimating UAV position. Our unified framework simultaneously provides robust victim localization and spoofer localization, enabling active attacker attribution beyond passive defense. Extensive simulations validate the effectiveness of our optimization and security mechanisms for 3GPP-compliant UAV positioning.