Semantic Sensing Across the Edge-Cloud Synergy: Resource-Efficient Monitoring for Industry 5.0
IEEE Access · 2026 · 14 · 87508-87525
DOI: 10.1109/access.2026.3701453 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Access, 2026. Verfasst von Nachuan Yang und 6 weiteren Autor:innen. Die Arbeit nennt 32 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Key Science and Technology Research Project of Henan Province.
Abstract (Originalsprache)
Semantic communication offers a promising paradigm for industrial monitoring by shifting the focus from bit-level transmission fidelity to task-oriented information utility. However, existing semantic models are often too computationally heavy for resource-constrained industrial devices. To bridge this critical gap, this paper proposes a lightweight Semantic-aware Edge-cloud collaborative Sensor Network (SESN) framework. Instead of relying on heavy neural inference, SESN derives semantic importance from low-dimensional statistical features, specifically local signal variation, abnormality probability, and equipment criticality. This design establishes a computationally lightweight and direct link between data value and network scheduling. The paper formulates a long-term stochastic optimization problem that couples admission control, resource allocation, and federated learning, and solves it through a drift-plus-penalty online policy that balances task utility against network congestion and synchronization cost. To limit the extra communication overhead introduced by federated edge-cloud collaboration, the framework further adopts a compressed model-update mechanism with error feedback. Theoretical analysis shows that the policy achieves near-optimal utility with a controllable optimality gap. Experimental results demonstrate that SESN reduces average delay by 28% to 35% and communication overhead by 30% to 40% compared with cloud-only baselines while preserving an anomaly-detection F1-score of 0.917. These results indicate that SESN provides an implementable cross-layer solution for low-overhead intelligent sensing and edge-cloud collaboration in industrial monitoring systems.
Förderung
- Key Science and Technology Research Project of Henan Province
Zitieren
APA
Yang, N., Wu, J., Cheng, L., Zhang, W., Xu, P., Ma, T., & Zhou, H. (2026). Semantic Sensing Across the Edge-Cloud Synergy: Resource-Efficient Monitoring for Industry 5.0. IEEE Access, 14, 87508-87525. https://doi.org/10.1109/access.2026.3701453
Vancouver
Yang N, Wu J, Cheng L, Zhang W, Xu P, Ma T, et al. Semantic Sensing Across the Edge-Cloud Synergy: Resource-Efficient Monitoring for Industry 5.0. IEEE Access. 2026;14:87508-87525. doi:10.1109/access.2026.3701453
BibTeX
@article{Yang2026,
author = {Yang, Nachuan and Wu, Junjie and Cheng, Long and Zhang, Wei and Xu, Peng and Ma, Ting and Zhou, Houchang},
title = {Semantic Sensing Across the Edge-Cloud Synergy: Resource-Efficient Monitoring for Industry 5.0},
journal = {IEEE Access},
year = {2026},
volume = {14},
pages = {87508--87525},
doi = {10.1109/access.2026.3701453}
}
EUVIMED