Communication-efficient federated learning for digital twin systems of industrial internet of things

Zhao, Yunming, Li, Li, Liu, Ying ORCID: https://orcid.org/0000-0001-9319-5940, Fan, Yuxi and Lin, Kuo-Yi 2022. Communication-efficient federated learning for digital twin systems of industrial internet of things. Presented at: 14th IFAC Workshop on Intelligent Manufacturing Systems (IMS 2022), Tel-Aviv, Israel, 28-30 March 2022. IFAC-PapersOnLine. IFAC-PapersOnLine. , vol.55(2) Elsevier, pp. 433-438. 10.1016/j.ifacol.2022.04.232

Preview

PDF - Published Version Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (754kB) | Preview

Abstract

With the rapid development and deployment of Industrial Internet of Things technology, it promotes interconnection and edge applications in smart manufacturing. However, challenges remain, such as yet-to-improve communication efficiency and trade-offs between computing power and energy consumption, which limits the application and further development of IIoT technology. This paper proposes the digital twin systems into the IIoT to build model between physical objects and digital virtual systems to optimize the structure of IIoT. And we further introduce federal learning to train the digital twins model and to improve the communication efficiency of IIoT. In this paper, we first establish the digital twins model of IIoT based on industrial scenario. Moreover, to optimize the communication overhead allocation problem, this paper proposes an improved communication-efficient distribution algorithm, which speeds up the training performance of federated model and ensures the performance of industrial system model by changing the update training mode of client and server and allowing some industrial equipment to participate in federated training. This paper simulates the real-word intelligent camera detection to validate the proposed method. Comparing our proposed method with the existing traditional methods, the results show the advantages of the proposed method can improve the communication performance of the training model.

Item Type:	Conference or Workshop Item (Paper)
Date Type:	Publication
Status:	Published
Schools:	Engineering
Publisher:	Elsevier
Date of First Compliant Deposit:	16 December 2021
Last Modified:	10 Nov 2022 10:15
URI:	https://orca.cardiff.ac.uk/id/eprint/146202

Actions (repository staff only)

Edit Item