Comparative study of distance metrics for indoor radio propagation modeling in complex environments

This paper investigates experimentally and theoretically the impact of distance metric selection on path‑loss modeling in indoor corridor environments under non‑line‑of‑sight (NLoS) conditions. Measurements were conducted at five central frequencies between 2 and 7.4 GHz and complemented with deterministic ray‑tracing simulations in a geometry‑accurate environment. Five propagation models (FI, CI, CIF, ABG, Molisch) were evaluated using both Euclidean and Manhattan distance definitions, and their performance was assessed through the standard deviation error σ. Results show that the Manhattan metric consistently improves the accuracy of geometry‑sensitive models, particularly the Molisch formulation, by better approximating the effective propagation path in structured indoor layouts. The findings highlight the importance of incorporating environmental topology into distance definitions for reliable NLoS path-loss prediction in sub-7 GHz and lower mid-band frequencies.