Resilient Maritime Localization Using 5G Non-Terrestrial Networks Against GNSS Spoofing and Jamming

Abstract

Autonomous and smart vessels require accurate, reliable, and secure positioning to ensure safe navigation, collision avoidance, and autonomous berthing. However, Global Navigation Satellite Systems (GNSS) are vulnerable to spoofing and jamming, which can degrade positioning accuracy or generate misleading location information, posing significant risks to maritime operations. As a complementary and resilient alternative, 5G Non-Terrestrial Networks (NTNs) integrate low-Earth-orbit (LEO) satellites into the 5G ecosystem, enabling global connectivity and an independent positioning capability. Although recent 3GPP standardization efforts have enabled single-satellite localization using Round-Trip Time (RTT) and Doppler measurements, current approaches typically achieve only kilometer-level accuracy, limiting their applicability for maritime navigation. This paper proposes a deep-learning-enhanced single-satellite localization framework to significantly improve 5G NTN positioning performance. The method leverages RTT, Doppler (range-rate), and satellite beamforming features defined in the 3GPP NTN standard. A deep learning super-resolution (DLSR) model is trained to compensate the sampling uncertainty so as to reach higher measurement resolution and accuracy. Simulation results demonstrate that the proposed scheme achieves sub-29 m error for 90% of cases, with a root-mean-square error of 12 m using a single LEO satellite at 600 km altitude. The results indicate that decameter-level positioning is feasible with single-satellite 5G NTN, offering a viable GNSS-resilient solution for port and near-shore maritime operations.