Transformer-Based Cooperative UAV Encirclement Policies Under Uncertainty in Low-Altitude Wireless Networks
IEEE Transactions on Cognitive Communications and Networking · 2026 · 12 · 3525-3537
DOI: 10.1109/tccn.2025.3632389 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Cognitive Communications and Networking, 2026. Verfasst von Zhengzhi Yang und 2 weiteren Autor:innen. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 44 Literaturangaben. Gefördert von: Beijing Natural Science Foundation, National Natural Science Foundation of China.
Abstract (Originalsprache)
Integrated Sensing and Communication (ISAC) enables Uncrewed Aerial Vehicle (UAV) swarms to coordinate strategies and acquire environmental information via wireless networks. This capability is critical for cooperative multi-UAV encirclement, which serves as a precise and effective strategy for threat neutralization and is essential for the safe and orderly management of low-altitude airspace. However, ISAC services may fluctuate in some scenarios, making resilient and efficient encirclement strategies essential to ensure mission effectiveness under uncertain network conditions. We propose a multi-agent Transformer framework that models contextual dependencies within noisy or incomplete historical trajectories. By extracting latent representations of the joint state from partially observed data, the model captures temporal semantics and inter-agent correlations, enabling robust understanding of uncertain environments. To enhance resilience, a modular decision architecture is designed, consisting of networks for state prediction and action planning. This allows agents to infer global states and make reliable decisions, even in the presence of observation errors and communication dropouts. An offline-to-online training paradigm is further adopted, where the model is pretrained on clean offline data and then fine-tuned with real-time interactions, improving adaptability to dynamic and disturbed environments. Experimental results demonstrate that the proposed method achieves higher capture success rates than conventional DRL baselines. Notably, it maintains strong performance under observation and communication disturbances.
Förderung
- Beijing Natural Science Foundation
- National Natural Science Foundation of China
Zitieren
APA
Yang, Z., Cui, Y., & Li, Y. (2026). Transformer-Based Cooperative UAV Encirclement Policies Under Uncertainty in Low-Altitude Wireless Networks. IEEE Transactions on Cognitive Communications and Networking, 12, 3525-3537. https://doi.org/10.1109/tccn.2025.3632389
Vancouver
Yang Z, Cui Y, Li Y. Transformer-Based Cooperative UAV Encirclement Policies Under Uncertainty in Low-Altitude Wireless Networks. IEEE Transactions on Cognitive Communications and Networking. 2026;12:3525-3537. doi:10.1109/tccn.2025.3632389
BibTeX
@article{Yang2026,
author = {Yang, Zhengzhi and Cui, Yuanhao and Li, Yumeng},
title = {Transformer-Based Cooperative UAV Encirclement Policies Under Uncertainty in Low-Altitude Wireless Networks},
journal = {IEEE Transactions on Cognitive Communications and Networking},
year = {2026},
volume = {12},
pages = {3525--3537},
doi = {10.1109/tccn.2025.3632389}
}
EUVIMED