Maintaining Line-of-Sight Communications: A Vision-Aided Approach
IEEE Internet of Things Journal · 15.06.2026 · 13 (12) · 25655-25670
DOI: 10.1109/jiot.2026.3673966 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Internet of Things Journal, erschienen am 15.06.2026. Verfasst von You Li und 3 weiteren Autor:innen. Die Arbeit nennt 42 Literaturangaben. Gefördert von: Beijing Natural Science Foundation.
Abstract (Originalsprache)
Line-of-sight (LOS) communication plays a critical role in enhancing received signal quality in wireless communications. However, the presence of blockages located in the way between transmitters (Txs) and receivers (Rxs) can significantly degrade the performance of wireless transmission. To address this challenge, we propose a novel computer vision (CV)-aided approach for maintaining LOS communications. The proposed approach is designed with a two-stage framework and integrates lightweight models for image segmentation and amodal completion, which is specifically tailored for efficient deployment on resource-constrained edge devices in order to enable precise assessment for occlusion at low computational cost. Specifically, image segmentation model features improved network structures of the backbone and atrous spatial pyramid pooling (ASPP) module, as well as a specially designed loss function to alleviate the imbalance between background and foreground. The amodal completion model is trained via a self-supervised fashion in order for accurate mask completion without the need for manual annotations. In addition, based on the evaluated occlusion ratio, we present an efficient strategy for adjusting the Tx’s position to maintain a stable LOS link. A prototype verification system is developed to validate the proposed approach, and experimental results demonstrate its capability of highly efficient position adjustment for sustainable LOS links.
Förderung
- Beijing Natural Science Foundation
Zitieren
APA
Li, Y., Yin, S., Chen, J., & Deng, L. (2026). Maintaining Line-of-Sight Communications: A Vision-Aided Approach. IEEE Internet of Things Journal, 13(12), 25655-25670. https://doi.org/10.1109/jiot.2026.3673966
Vancouver
Li Y, Yin S, Chen J, Deng L. Maintaining Line-of-Sight Communications: A Vision-Aided Approach. IEEE Internet of Things Journal. 2026;13(12):25655-25670. doi:10.1109/jiot.2026.3673966
BibTeX
@article{Li2026,
author = {Li, You and Yin, Sixing and Chen, Jinzhen and Deng, Li},
title = {Maintaining Line-of-Sight Communications: A Vision-Aided Approach},
journal = {IEEE Internet of Things Journal},
year = {2026},
volume = {13},
number = {12},
pages = {25655--25670},
doi = {10.1109/jiot.2026.3673966}
}
EUVIMED