Effect of Environmental RF Propagation Impairments on Satellite Communication Link and Prediction of Cloud Attenuation using Machine Learning Algorithms

Abstract

Abstract

Satellite communication networks is advocating the usage of millimeter wave (mmWave) frequencies to cater to the growing demand for high data rates, broader bandwidths, and reliable communication links. However, these high-frequency bands are increasingly prone to propagation impairments, particularly those originating from atmospheric conditions such as rain, clouds, and ionospheric disturbances. Among these, cloud attenuation faces a notable challenge due to its extended duration and periodic occurrence, specifically in tropical regions like India. Low-level clouds, such as cumulonimbus, present significantly to signal degradation due to their dense water content. This research focuses on the importance of understanding and predicting cloud attenuation and its effect on satellite link reliability. To this end, machine learning algorithms—linear regression, decision trees, and multiple regression—are employed to estimate cloud-related parameters such as cloud cover, derived from atmospheric temperature, humidity, and pressure data. Furthermore, in the absence of fully developed cloud attenuation measurement systems in many tropical regions, predictive attenuation models serve as indispensable tools for system design and planning. This work reviews existing models, discusses their limitations, and explores ways to adapt them to local climatic conditions to ensure effective and flexible satellite communication.