A Lightweight Interference Estimation Method for Sub-6GHz Private 5G Using 3D City Maps

In Japan’s private 5G, identical channels are often reused by different licensees, so adjacent cells are prone to mutual interference. The common remedy, large separation based on siteagnostic models, suppresses spatial reuse and wastes spectrum. We propose a lightweight estimator that fuses simple propagation with 3D city-model data to explicitly account for blockage at low cost. The method classifies LOS/NLOS from building footprints and heights, applies a two-ray/FSPL blend in LOS, and adds knife-edge diffraction plus proximity reflection in NLOS. Across outdoor trials, it reduces the measured-over-estimate ratio (OVRT) while keeping RMSE/MAE/STD comparable to conventional models, indicating safer coordination with smaller guard margins.