Xiong, Ruoting
ORCID: https://orcid.org/0000-0003-0575-4218, Tong, Kit-Lun, Ren, Yi
ORCID: https://orcid.org/0000-0001-7423-6719 and Ren, Wei
(2026)
Mobility-Aware Satellite Backhaul Orchestration for 5G-Integrated Non-Terrestrial Networks.
IEEE Transactions on Vehicular Technology.
ISSN 0018-9545
Preview |
PDF (TVT_Main Manuscript_Camera Ready)
- Accepted Version
Available under License Creative Commons Attribution. Download (7MB) | Preview |
Abstract
Satellite-based non-terrestrial networks integrate satellites as backhauls to extend the coverage of terrestrial 5 G networks. With multiple types of satellites available as backhaul links, we consider the satellite selection problem for traffic orchestration, aiming to satisfy diverse quality-of-service (QoS) requirements while minimizing monetary costs. Nevertheless, satellite mobility introduces new challenges to traffic orchestration in 5 G satellite-terrestrial integrated networks. Existing solutions often fail to account for the time-varying latency, bandwidth constraints, and cost fluctuations of satellite backhaul links, which are inherently induced by satellite mobility. To address this gap, this paper proposes two satellite selection methods: a static approach based on predefined rules and a dynamic approach driven by reinforcement learning (RL). The static method first classifies incoming traffic into different application types and user behavior patterns using deep learning, then orchestrates the traffic to appropriate satellite links based on established rules. The dynamic method leverages multi-armed bandit (MAB) models within an RL framework to balance latency performance and monetary costs, thereby enabling adaptive satellite selection. Experimental results show that the static method satisfies 98% of the QoS requirements for various traffic types and reduces costs by a factor of 1.5-2. Moreover, the MAB-based models achieve a good balance between delay and price, yielding low average regret and reasonable traffic allocation even under price-surge scenarios.
| Item Type: | Article |
|---|---|
| Faculty \ School: | Faculty of Science > School of Computing Sciences Faculty of Science |
| UEA Research Groups: | Faculty of Science > Research Groups > Cyber Intelligence and Networks Faculty of Science > Research Groups > Data Science and AI Faculty of Science > Research Groups > Health Computing |
| Related URLs: | |
| Depositing User: | LivePure Connector |
| Date Deposited: | 19 Nov 2026 10:09 |
| Last Modified: | 19 Nov 2026 10:09 |
| URI: | https://ueaeprints.uea.ac.uk/id/eprint/105147 |
| DOI: | 10.1109/TVT.2026.3752419 |
Downloads
Downloads per month over past year
Actions (login required)
![]() |
View Item |
Tools
Tools