Probabilistic Modeling of Dust-Induced FSO Attenuation for 5G/6G Backhaul in Arid Regions

Free-Space Optical (FSO) communication systems operating in arid regions, especially those envisioned for current and future 5G/6G networks, are significantly affected by dust storms, which cause signal attenuation and service disruptions. While previous studies have proposed deterministic models to characterize attenuation in both controlled and real environments, probabilistic modeling approaches remain largely unexplored, particularly for capturing the variability of FSO signal attenuation under dust conditions. This study proposes a probabilistic model for FSO signal attenuation developed from experiments conducted in a repeatable and well-characterized controlled dust chamber. The chamber-based setup allowed precise manipulation of dust visibility levels and consistent data collection, serving as a benchmark for statistical modeling. We analyzed the measurements to fit appropriate probability distributions for modeling the signal attenuation as a random variable. The empirical data were fitted to several candidate distributions, and the Johnson SB distribution consistently achieved superior performance with R2≥0.95 and RMSE and MAE values close to zero across all dust conditions. The results offer a foundational framework for modeling dust-induced attenuation as a random process, providing statistical bounds for FSO link planning in desert environments.