Intelligent Digital Twin Communication Framework for Addressing Accuracy and Timeliness Tradeoff in Resource-Constrained Networks

Cakir, Lal Verda; Thomson, Craig J.; Özdem, Mehmet; Canberk, Berk; Nguyen, Van-Linh; Duong, Trung Q.

Abstract

The accuracy and timeliness tradeoff prevents Digital Twins (DTs) from realizing their full potential. High accuracy is crucial for decision-making, and timeliness is equally essential for responsiveness. Therefore, this tradeoff in DT communication must be addressed to achieve DT synchronization. Previous studies identified the issue but considered the problem as maximizing data transfer, which is infeasible due to resource constraints. To facilitate this, we quantify accuracy and timeliness as E and ϕ and define the problem as joint minimisation. We then introduce the Intelligent DT Communication (IDTC) Framework to solve the problem, which includes machine learning-based Predictive Synchronization (PS) and DT synchronization management (DTSYNC) protocol. Here, PS uses imputation and forecasting to generate future values, which are utilized to update DT at the projected time points. This mechanism of PS enables lowering E and ϕ of the communication. Subsequently, we utilize the DTSYNC to control synchronization and optimise the twining frequency ft. We evaluate the proposed framework using a public dataset and compare its performance with several state-of-the-art studies in a real-world scenario. Evaluation results indicate that IDTC outperforms the existing methods by 80% for E and 84% for ϕ while enabling ft adjustment, resulting in 3.8 times goodput improvement.

Citation

Cakir, L. V., Thomson, C. J., Özdem, M., Canberk, B., Nguyen, V.-L., & Duong, T. Q. (online). Intelligent Digital Twin Communication Framework for Addressing Accuracy and Timeliness Tradeoff in Resource-Constrained Networks. IEEE Transactions on Cognitive Communications and Networking, https://doi.org/10.1109/tccn.2024.3469234

• Journal Article Type: Article
• Online Publication Date: Sep 27, 2024
• Deposit Date: Oct 7, 2024
• Publicly Available Date: Oct 7, 2024
• Journal: IEEE Transactions on Cognitive Communications and Networking
• Publisher: Institute of Electrical and Electronics Engineers (IEEE)
• Peer Reviewed: Peer Reviewed
• DOI: https://doi.org/10.1109/tccn.2024.3469234
• Keywords: intelligent communication, machine learning, digital twin, synchronization, resource management