Deep Reinforcement Learning‐Enabled QoS‐Aware Adaptive Modulation for Visible Light‐Infrared Communication Systems
IET Communications · 2026-01 · 20 (1)
DOI: 10.1049/cmu2.70195 ↗Kurz gefasst
Fachartikel in der Zeitschrift IET Communications, online erschienen am 15.07.2026. Verfasst von Lei Xiao und 5 weiteren Autor:innen. Die Arbeit nennt 19 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China.
Abstract (Originalsprache)
Visible light communication has emerged as a promising technology for aircraft cabin communications, leveraging its abundant spectrum resources and inherent resistance to electromagnetic interference. However, its performance is frequently compromised by dynamic channel variations and unpredictable link interruptions. To address these challenges, a deep reinforcement learning (DRL)‐assisted visible light‐infrared communication (VLIRC) system with quality‐of‐service (QoS)‐aware adaptive modulation is proposed. Firstly, an improved pilot‐based linear minimum mean square error (LMMSE) algorithm is developed to estimate the downlink channel state information (CSI), significantly maintaining a balance between performance and hardware resource consumption. Then, a deep Q‐network (DQN) is introduced to learn an adaptive modulation policy over a multi‐dimensional state space comprising channel SNR, SNR variation trend and QoS class, enabling simultaneous optimisation for heterogeneous traffic (e.g., voice and video). Moreover, an infrared uplink is further introduced to feed back the CSI to the transmitter, and the impact of IR link packet error rate on system throughput is characterized. Hardware‐informed simulation results demonstrated that the proposed scheme consistently satisfies the BER requirements of both voice and video services while improving throughput robustness in the modulation transition region. Moreover, the hardware system analysis has validated its practical deployability.
Förderung
- National Natural Science Foundation of China
Zitieren
APA
Xiao, L., Cheng, Q., Lai, B., Liu, X., Liu, B., & Du, J. (2026). Deep Reinforcement Learning‐Enabled QoS‐Aware Adaptive Modulation for Visible Light‐Infrared Communication Systems. IET Communications, 20(1). https://doi.org/10.1049/cmu2.70195
Vancouver
Xiao L, Cheng Q, Lai B, Liu X, Liu B, Du J. Deep Reinforcement Learning‐Enabled QoS‐Aware Adaptive Modulation for Visible Light‐Infrared Communication Systems. IET Communications. 2026;20(1). doi:10.1049/cmu2.70195
BibTeX
@article{Xiao2026,
author = {Xiao, Lei and Cheng, Quanyi and Lai, Baolin and Liu, Xiaodong and Liu, BuGao and Du, Junyi},
title = {Deep Reinforcement Learning‐Enabled QoS‐Aware Adaptive Modulation for Visible Light‐Infrared Communication Systems},
journal = {IET Communications},
year = {2026},
volume = {20},
number = {1},
doi = {10.1049/cmu2.70195}
}
EUVIMED