Dynamic Resource Management for Next-Generation Satellite Networks

Abstract

This thesis presents novel beam management and radio resource allocation techniques for next-generation multibeam satellite networks equipped with phased array antennas. Advanced frequency reuse and dynamic beam steering capabilities of phased array antennas enable next-generation multibeam satellite systems to deliver adaptive coverage and high throughputs, thereby extending broadband connectivity to remote and disadvantaged areas. A fundamental challenge persists in efficiently managing scarce satellite resources to accommodate heterogeneous traffic profiles in such areas.The thesis specifically focuses on multibeam satellite networks where beam directions are dynamically steerable toward regions with active ground user presence, rather than being confined to predefined geographic areas. This adaptive beamforming allows efficient radio resource utilization, particularly within full-frequency reuse paradigms. It can, however, have an increased possibility of interbeam interference, necessitating effective beam coordination. In general, the resource optimization problem is combinatorial and non-deterministic polynomial-time hard (NP-hard), making exhaustive search methods computationally infeasible.This thesis presents resource optimization techniques for next-generation multibeam Geostationary Earth Orbit (GEO) satellite networks, including beam placement, beam hopping, and radio resource allocation. Firstly, ground user clustering for beam placement is addressed by presenting a novel polynomial-time geometric user clustering algorithm that balances two competing objectives: cluster compactness and beam switching overhead. User clustering for beam positioning aggregates heterogeneously distributed ground users into spatially coherent clusters so that each steerable beam serves a demand hotspot, concentrating the satellite’s limited resources and improving interference management. Secondly, a novel time-efficient user-cluster grouping algorithm for beam scheduling is proposed for adaptive multibeam hopping. The proposed algorithm integrates seamlessly with the satellite’s physical layer for beamforming and other radio resource allocation algorithms.The thesis also presents two novel radio resource allocation approaches. A joint beam direction and power optimization scheme for the downlink of multibeam GEO satellite networks is developed to attain target signal-to-interference-plus-noise ratio (SINR) requirements while minimizing the total power. An iterative feedback-based resource allocation algorithm is proposed to jointly optimize time and power resource allocation, maximizing total user demand satisfaction rate under minimum total power conditions.Numerical evaluations based on real-world population data show that the geometric user clustering algorithm more than doubles both the zero-outage probability and the median user rate compared to baseline methods under full network load. It also exhibits near-linear throughput scaling with the number of available radio-frequency chains. For adaptive multibeam hopping, the performance of the proposed user-cluster grouping algorithm is examined using extensive simulations, demonstrating that it produces near-optimum beam-hopping schedules with low outage probability compared to benchmark grouping approaches.For radio resource allocation, jointly optimizing beam direction and power is shown to achieve an SINR gain when compared to the power allocation without beam direction optimization in the case of GEO satellite networks with analog beamforming. It also offers a spatial multiplexing advantage by enabling the simultaneous provisioning of critically close user locations. The proposed iterative feedback-based resource allocation algorithm is shown to closely approximate the cumulative distribution function of ground user demand, while consistently outperforming existing benchmark schemes in terms of both fairness and efficiency.