EUVIMEDEUVIMED entdecken
Crossref

DNN Partitioning and Offloading in Vehicular Networks Based on RSMA

Zhijian Lin, Shaowei Yang, Celimuge Wu, Wenhao Wu, Zhizhao Lu, Jiguang He, Linhuang Wu

IEEE Transactions on Vehicular Technology · 2026-08 · 75 (8) · 18799-18804

DOI: 10.1109/tvt.2026.3673413 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Vehicular Technology, 2026-08. Verfasst von Zhijian Lin und 6 weiteren Autor:innen. Die Arbeit nennt 15 Literaturangaben. Gefördert von: National Natural Science Foundation of China, NSF of Fujian Province, Japan Science and Technology Agency ASPIRE.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The proliferation of intelligent connected vehicles and related applications with deep neural network (DNN) inference tasks has boosted a pivotal requirement for the unprecedented low latency and low energy consumption of the Internet of Vehicles (IoV) networks. However, it is difficult to deploy large-scale and computation-intensive DNN on resource-constrained vehicles. Inspired by the idea of distributed computing, a DNN partitioning and offloading strategy based on RSMA is designed, where various computational auxiliary nodes are introduced in a stochastic geometry approach to alleviate the computation dilemma of the task vehicle. This strategy selects the optimal DNN model partition points based on the computing capability of the auxiliary nodes, and then develops the optimal task offloading strategy to realize the effective distribution and execution of tasks between the edge server and the service vehicle. To minimize the average cost problem, we propose a joint optimization algorithm combining heterogeneous-earliest-finish-time algorithm and successive convex approximation algorithm (HEFT-SCA) to achieve the optimal offloading strategy. Experimental results show that compared to the baselines, the HEFT-SCA can reduce the average cost.

Förderung

  • National Natural Science Foundation of China
  • NSF of Fujian Province
  • Japan Science and Technology Agency ASPIRE
  • Guangdong Research Team for Communication and Sensing Integrated with Intelligent Computing

Zitieren

APA

Lin, Z., Yang, S., Wu, C., Wu, W., Lu, Z., He, J., & Wu, L. (2026). DNN Partitioning and Offloading in Vehicular Networks Based on RSMA. IEEE Transactions on Vehicular Technology, 75(8), 18799-18804. https://doi.org/10.1109/tvt.2026.3673413

Vancouver

Lin Z, Yang S, Wu C, Wu W, Lu Z, He J, et al. DNN Partitioning and Offloading in Vehicular Networks Based on RSMA. IEEE Transactions on Vehicular Technology. 2026;75(8):18799-18804. doi:10.1109/tvt.2026.3673413

BibTeX
@article{Lin2026,
  author = {Lin, Zhijian and Yang, Shaowei and Wu, Celimuge and Wu, Wenhao and Lu, Zhizhao and He, Jiguang and Wu, Linhuang},
  title = {DNN Partitioning and Offloading in Vehicular Networks Based on RSMA},
  journal = {IEEE Transactions on Vehicular Technology},
  year = {2026},
  volume = {75},
  number = {8},
  pages = {18799--18804},
  doi = {10.1109/tvt.2026.3673413}
}
Kostenlos starten →