Enhancing 5G-based Localization in Dynamic Environments through Network Digital Twins

The increasing demand for reliable Vehicle-to-Everything (V2X) communications and autonomous mobility necessitates sophisticated simulation frameworks and intelligent optimization strategies. This paper presents a Network Digital Twin (NDT) that integrates high-fidelity ray-tracing with real vehicular traffic data to model wireless propagation in dynamic urban environments and derive theoretical localization bounds. By explicitly exploiting multipath reflections, both line-of-sight (LOS) and non-line-of-sight (NLOS), from static and mobile reflectors such as vehicles, the framework supports the design of an optimized precoding scheme for enhanced user equipment (UE) positioning. Numerical results indicate that the proposed NDT-guided method reduces the Position Error Bound (PEB) by approximately 35%, underscoring NDT benefits and the utility of NLOS exploitation for high-accuracy localization in dense urban scenarios.