EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Thermal Aware Edge Inference Placement for Smart Factory Camera Networks Under Heat Constraints

Admin Admin, Rina Hoshino, Phoebe Fletcher

International Journal of Computational and Biological Sciences · 15.06.2026 · 3 (2) · 124-134

DOI: 10.66238/ijcbs91 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Computational and Biological Sciences, online erschienen am 15.06.2026. Verfasst von Admin Admin und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The integration of computer vision in smart manufacturing has driven the adoption of edge computing to process high-bandwidth video streams locally. However, deploying computation-intensive artificial intelligence models on edge devices within industrial environments introduces significant thermal management challenges. Factory floors often exhibit extreme and variable ambient temperatures, and edge devices frequently operate in passive cooling configurations to prevent dust accumulation. Consequently, conventional inference placement strategies that optimize solely for latency or energy efficiency frequently cause localized overheating, resulting in severe thermal throttling, unpredictable performance degradation, and reduced hardware lifespans. This paper presents a comprehensive investigation into thermal-aware edge inference placement specifically designed for smart factory camera networks. We introduce a novel optimization framework that continuously monitors the thermal states of heterogeneous edge nodes, models their heat dissipation characteristics, and dynamically redistributes neural network inference workloads to maintain optimal operating temperatures. By transforming the inference placement challenge into a multidimensional thermal-constrained scheduling problem, our approach ensures stable, real-time video analytics without violating hardware temperature thresholds. Through extensive evaluation using a simulated industrial environment, we demonstrate that the proposed thermal-aware strategy significantly reduces peak device temperatures while maintaining competitive end-to-end latency compared to traditional performance-centric algorithms. The findings highlight the critical necessity of incorporating thermal dynamics into edge resource management for robust and sustainable industrial automation.

Zitieren

APA

Admin, A., Hoshino, R., & Fletcher, P. (2026). Thermal Aware Edge Inference Placement for Smart Factory Camera Networks Under Heat Constraints. International Journal of Computational and Biological Sciences, 3(2), 124-134. https://doi.org/10.66238/ijcbs91

Vancouver

Admin A, Hoshino R, Fletcher P. Thermal Aware Edge Inference Placement for Smart Factory Camera Networks Under Heat Constraints. International Journal of Computational and Biological Sciences. 2026;3(2):124-134. doi:10.66238/ijcbs91

BibTeX
@article{Admin2026,
  author = {Admin, Admin and Hoshino, Rina and Fletcher, Phoebe},
  title = {Thermal Aware Edge Inference Placement for Smart Factory Camera Networks Under Heat Constraints},
  journal = {International Journal of Computational and Biological Sciences},
  year = {2026},
  volume = {3},
  number = {2},
  pages = {124--134},
  doi = {10.66238/ijcbs91}
}
Kostenlos starten →