In-Tunnel Single-Anchor Localization Exploiting Near-Field and Radio-Reflective Road Markings

Accurate vehicular localization in Global Navigation Satellite System (GNSS)-denied environments, such as road tunnels, remains a key challenge for cooperative intelligent transport systems (C-ITS). This paper investigates single-anchor positioning by exploiting near-field (NF) propagation and passive radio-reflective structures. We first derive a geometric validity condition for the single-reflector NF (SR-NF) channel model, establishing a bound on the array size under which multipath can be consistently modeled by a single reflector, and linking it to Fresnel-region scaling. Building on this result, we propose JAVELIN, a single-anchor localization framework combining tensor-based NF parameter estimation, adaptive NF/far-field (FF) processing, and recursive Bayesian tracking. The method integrates angle, delay difference, and curvature measurements into a variable-dimension extended Kalman filter with gated nearest-neighbor (NN) association, enabling operation without prior environmental knowledge. Radio-reflective road markings (RRMs) are further introduced to enhance geometric diversity. Simulation results in realistic tunnel scenarios demonstrate accurate and robust localization under different line-of-sight (LoS) conditions, outperforming state-of-the-art single-anchor approaches and benefiting from passive reflector deployment.