Event Details
Utility-Based Beam and Resource Scheduling for Beam-Hopping LEO Satellite Networks with Interference-Aware Bandwidth Reuse
Presenter: Pooria Seyed Eftetahi
Supervisor:
Date: Thu, November 19, 2026
Time: 09:00:00 - 00:00:00
Place: EOW 430
ABSTRACT
Abstract:
The growing reliance on low Earth orbit (LEO) satellite constellations to provide broadband connectivity for diverse user and Internet of Things (IoT) applications has increased the need for intelligent resource management. In beam-hopping satellite networks, varying traffic demands, user priorities, and interference caused by frequency reuse introduce significant challenges in efficiently allocating beams and bandwidth. Existing approaches often optimize beam placement and bandwidth allocation independently, limiting their ability to meet heterogeneous service requirements.
This seminar presents UBER (Utility-Based Beam and Resource Scheduler), a two-stage framework that integrates reinforcement learning with interference-aware bandwidth allocation to improve resource utilization and quality of service (QoS). In the first stage, a reinforcement learning agent based on Proximal Policy Optimization (PPO) dynamically selects beam placements to maximize user utility while considering payload and interference constraints. In the second stage, an adaptive heuristic allocates bandwidth by intelligently reusing frequency resources among spatially separated beams, improving spectral efficiency while limiting interference.
By incorporating utility functions tailored to different traffic types and user priorities, UBER enables QoS-driven and fair resource allocation across heterogeneous users. Simulation results across multiple user distributions, including realistic spatial deployments, demonstrate that UBER significantly outperforms existing state-of-the-art approaches in terms of throughput, user utility, and fairness. These findings highlight the potential of combining reinforcement learning and interference-aware resource scheduling to improve the performance of next-generation LEO satellite networks.
