QoE-Oriented Streaming Power Allocation in LEO Satellite Networks

Abstract

Low Earth Orbit (LEO) satellite networks are expected to support global multimedia services in future sixth-generation (6G) systems, where streaming media constitutes a major traffic type with stringent Quality of Experience (QoE) requirements. However, existing satellite resource allocation schemes mainly optimize Quality of Service (QoS) metrics, overlooking streaming-specific characteristics such as playback buffering and user early departure, which may lead to severe QoE degradation and on-board energy waste. To address this issue, this paper proposes a QoE-oriented downlink power allocation scheme for on-demand streaming services in LEO satellite networks. We first formulate a long-term stochastic optimization problem that jointly minimizes playback stalling and energy waste caused by buffer overflow and user interruption, while guaranteeing long-term user QoE. By leveraging Lyapunov optimization, the original problem is transformed into deterministic per-slot subproblems. A low-complexity power allocation algorithm is then developed based on strong subgradient theory, with theoretical convergence guarantees established through the strong quasiconvexity analysis of the per-user objective function. Simulation results based on the practical OneWeb constellation demonstrate that the proposed algorithm substantially improves user Mean Opinion Score (MOS) while significantly reducing on-board energy waste. In particular, a microscopic analysis of per-user power dynamics reveals that the performance degradation of conventional QoS-centric schemes stems from their buffer-agnostic, channel-driven allocation, which causes severe resource imbalance and energy waste—an inherent limitation that the proposed scheme effectively circumvents through application-layer awareness.