DT-LAWN: Incrementally Synchronized Digital Twins for Real-Time NextG Wireless Navigation

Abstract

Connectivity-aware navigation in NextG wireless networks requires a digital twin (DT) that remains consistent with the physical environment as channel conditions evolve. Existing DT-based navigation systems rely on static wireless maps computed offline, causing routing decisions to degrade as vehicles and temporary obstructions alter propagation conditions. This paper presents DT-LAWN (Local-Aware Wireless Navigation), which enables real-time connectivity-aware navigation through incremental DT state synchronization, whereby only the grid cells within the spatial influence region of moved blockers are recomputed at each time step. Coupled with event-driven replanning, the framework adapts routes across pedestrian XR, connected vehicle, and UAV corridor scenarios, with a fine-tuned LLM providing a natural-language planning interface. Evaluated on received signal strength (RSS) maps generated from real OpenStreetMap geometry using the 3GPP TR 38.901 UMa-based propagation model, DT-LAWN reduces outage probability by up to 85% in dense urban deployments while achieving estimated full-pipeline median replanning latencies of 373–427 ms across the evaluated mobility scenarios.