A Large-Scale LoRa Measurement Campaign for Path-Loss Modeling in Urban and Suburban Environments

Abstract

Low-power wide-area networks, particularly Long-Range (LoRa), are important enablers of large-scale internet of things deployments in urban environments. However, inconsistencies in measurement methodologies have led to limitations in the accuracy and reproducibility of path loss models. This work presents a large-scale measurement campaign conducted in urban and suburban areas in the city of Funchal, Madeira Island, designed to characterize LoRa propagation in complex environments and to propose more reproducible alternatives for developing path loss models. The methodology uses fixed measurement positions, calibrated hardware, and controlled data acquisition to collect the received signal strength indicator and the signal-to-noise ratio to estimate effective signal power. A total of 1,203 locations were surveyed using three custom gateways, with multiple samples collected per position to mitigate small-scale fading. Results show fading-induced variations up to 35 dB, indicating the importance of acquiring multiple samples. Near the receiver sensitivity limit, path-loss estimation errors increase due to packet loss, reducing the available samples. An order-statistics-based compensation technique enabled accurate estimation with as few as two or three samples. The study also highlights the impact of interference on the path-loss measurement limit, and a similarity between uplink and downlink path loss behavior in the study area. Finally, the limitations of single-slope models are demonstrated, supporting the need for more refined modeling approaches in realistic urban scenarios.