Demand-Based Dynamic Bandwidth Allocation in Multi-Beam Satellites Using Machine Learning Concepts

In the realm of satellite communication, where the importance of efficient spectrum utilization is growing day by day due to the increasing significance of this technology, dynamic resource management has emerged as a pivotal consideration in the design of contemporary multi-beam satellites, facilitating the flexible allocation of resources based on user demand. This research paper delves into the pivotal role played by machine learning and artificial intelligence within the domain of satellite communication, particularly focusing on spot beam satellites. The study encompasses an evaluation of machine learning's application, whereby an extensive dataset capturing user demand across a specific geographical area is subjected to analysis. This analysis involves determining the optimal number of beams/clusters, achieved through the utilization of the knee-elbow method predicated on within-cluster sum of squares. Subsequently, the demand data are equitably segmented employing the weighted k-means clustering technique. The proposed solution introduces a straightforward yet efficient model for bandwidth allocation, contrasting with conventional fixed beam illumination models. This approach not only enhances spectrum utilization but also leads to noteworthy power savings, thereby addressing the growing importance of efficient resource management in satellite communication.