Empirical Path Loss Models for Long-Range Networks in Long Corridors of Multi-Floor Buildings

Long-range (LoRa) wireless networks in the UHF band are widely deployed in indoor sensing applications; however, standard propagation models such as ITU-R P.1238 and 3GPP do not accurately characterize LoRa performance because RSSI depends strongly on spreading factor (SF), bandwidth (BW), and signal-to-noise ratio (SNR), as well as building structures and inter-floor attenuation. This paper proposes new empirical path loss models for LoRa at 433 MHz, specifically developed for long-corridor, multi-floor buildings. The models incorporate a calibration factor for RSSI, an adaptive path loss exponent, and a floor attenuation factor (FAF) that increases by approximately 5 dB per floor. Measurements collected from four buildings demonstrate that the proposed models capture corridor wave-guiding behavior and extend the breakpoint distance to about 31.9 m. Validation results show high prediction accuracy, with RMSE values of 1.8-5 dB, outperforming ITU, LoRa, and 3GPP models. These models provide improved reliability for indoor LoRa network planning in complex multi-story environments.