Breaking the Kbps Uplink Barrier: Semantic-Guided Generative Satellite Communications for Video Transmission
IEEE Transactions on Communications · 2026 · 74 · 11326-11339
DOI: 10.1109/tcomm.2026.3708606 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Communications, 2026. Verfasst von Zhihan Chen und 2 weiteren Autor:innen. Die Arbeit nennt 38 Literaturangaben. Gefördert von: “U35 Pei Miao Funding Program” of Changping District, Beijing, National Natural Science Foundation of China, NSF.
Abstract (Originalsprache)
Satellite communication systems are pivotal for global connectivity by ensuring information exchange when terrestrial connection fails. Traditional satellite uplink transmission suffers from several to tens of Kbps data rates due to limited user terminal (UT) power, thus hindering data-intensive video transmission. In this paper, we introduce large-scale generative model (LGM) enabled semantic satellite communication, which transmits semantic information (SI) rather than conventional bits, capable of supporting diverse and high-quality videos generation. However, the limited uplink rates makes the effective SI extraction and reliable transmission non-trivial for videos. Besides, the intensive computation and stochastic nature of LGMs may incur high latency and uncontrollable hallucinations. To address these issues, we propose a semantic satellite communication system featuring a tailored semantic encoder at UT transmitter, as well as an on-orbit hybrid video recovery scheme enabled by a semantic-guided LGM and a lightweight auxiliary model (LAM). At the uplink UT transmitter, key frames are strategically selected and mapped into compact SI, which is further encoded into redundancy-augmented symbols for transmission. On the satellite, the hybrid video recovery scheme enables low-latency, high-fidelity video reconstruction, where the LGM generates a temporally representative subset of intact SI and the LAM restores the remaining contents. Moreover, the uplink SI continually anchors the LGM’s generation path within the target sub-space and direction, thus suppressing generative hallucinations while preserving its creativity. Simulation results show that the proposed scheme improves over 32% perceptual fidelity than the state-of-the-art ones given uplink data rates of 5-100 Kbps.
Förderung
- “U35 Pei Miao Funding Program” of Changping District, Beijing
- National Natural Science Foundation of China
- NSF
- Japan Science and Technology Agency (JST) Adopting Sustainable Partnerships for Innovative Research Ecosystem
Zitieren
APA
Chen, Z., Di, B., & Han, Z. (2026). Breaking the Kbps Uplink Barrier: Semantic-Guided Generative Satellite Communications for Video Transmission. IEEE Transactions on Communications, 74, 11326-11339. https://doi.org/10.1109/tcomm.2026.3708606
Vancouver
Chen Z, Di B, Han Z. Breaking the Kbps Uplink Barrier: Semantic-Guided Generative Satellite Communications for Video Transmission. IEEE Transactions on Communications. 2026;74:11326-11339. doi:10.1109/tcomm.2026.3708606
BibTeX
@article{Chen2026,
author = {Chen, Zhihan and Di, Boya and Han, Zhu},
title = {Breaking the Kbps Uplink Barrier: Semantic-Guided Generative Satellite Communications for Video Transmission},
journal = {IEEE Transactions on Communications},
year = {2026},
volume = {74},
pages = {11326--11339},
doi = {10.1109/tcomm.2026.3708606}
}
EUVIMED