Signal Attenuation Prediction Model for a 22 GHz Terrestrial Communication Link in Sudan Due to Dust and Sand Storms Using Machine Learning

Signal attenuation due to dust and sand storms is one of the major problems in the utilization of microwave frequency bands for terrestrial and space communication especially in arid regions such as Northern Africa and Middle Eastern regions. In this paper, a machine learning (ML) model is developed to predict microwave signal attenuation due to atmospheric conditions recorded during dust and sand storms. The model utilizes recorded meteorological data, particularly optical visibility, temperature, relative humidity, atmospheric pressure, and wind speed conditions to predict the signal attenuation for a 22GHz operating frequency terrestrial link in Sudan. Compared to measured values, the predicted signal attenuation values show a relatively optimistic relation between meteorological data and microwave signal attenuation when utilizing all the meteorological features.