Ground User Clustering for Adaptive Multibeam GEO Satellite Networks

Area of Science:

• Satellite communication networks
• Wireless network optimization
• Geometric algorithms

Background:

• Adaptive multibeam satellites require efficient user clustering for beam pointing and scheduling.
• Maximizing satellite throughput necessitates grouping ground users effectively.
• Existing methods may not adequately balance SINR and overhead in GEO networks.

Purpose of the Study:

• To develop a geometry-aware user clustering and cluster center optimization method for adaptive multibeam GEO satellite networks.
• To introduce the GeoClust algorithm, balancing signal-to-interference-plus-noise ratio (SINR) and hopping overhead.
• To ensure feasibility and convergence in cluster center updates.

Main Methods:

• A polynomial-time geometric set-cover approach for user clustering.
• Boyle-Dykstra projection for cluster center updates within an alternating optimization framework.
• Nearest-center membership updates to enforce cluster-radius constraints.

Main Results:

• GeoClust achieves near-linear throughput scaling with the number of RF chains.
• Demonstrates more than double the zero outage and median user rates under heavy traffic.
• Validated using real-world population data.

Conclusions:

• The proposed GeoClust algorithm enhances performance in adaptive multibeam GEO satellite networks.
• The method provides a feasible and convergent approach to user clustering and optimization.
• Significant improvements in user rates and throughput are achievable compared to benchmarks.