TWIST: Closed-Loop token Synchronization for Application-Aware Wireless Digital Twins

Problem
This paper addresses the challenge of synchronizing wireless digital twins—specifically, the need for efficient communication between a physical scene and its digital representation under constrained and variable communication resources. The authors highlight that existing methods often rely on pixel-domain transmission or uniformly protected bitstreams, which do not align well with the semantic requirements of applications utilizing these digital twins. This work is presented as a preprint and has not yet undergone peer review.

Method
The core technical contribution of this paper is the TWIST framework, which employs a closed-loop token synchronization approach. In TWIST, each physical observation is represented as a token, which is synchronized over a wireless link rather than focusing on visual fidelity. The framework categorizes token positions based on their relevance to specific tasks and employs mode-conditioned unequal error protection across three synchronization modes: low, medium, and high. This adaptive protection strategy allows for efficient resource allocation based on the current communication conditions. At the twin side, the framework utilizes decoding confidence to transform unreliable hard token decisions into erasures, which are subsequently restored using a completion model. This process updates the semantic state of the twin, enabling effective traffic-state inference. Additionally, TWIST generates compact feedback statistics, including channel quality, receiver uncertainty, semantic drift, and application priority, which inform subsequent mode adaptations.

Results
The experimental evaluation of TWIST was conducted in a dynamic road-scene digital twin scenario. The results demonstrate that TWIST significantly enhances traffic-state inference and semantic twin-state synchronization compared to baseline strategies, including fixed-mode and channel-only adaptation methods. Specifically, TWIST reduces the average synchronization cost relative to a strategy that always employs high transmission mode. While exact numerical improvements are not provided in the abstract, the qualitative improvements suggest a substantial effect size in terms of synchronization efficiency and application performance.

Limitations
The authors acknowledge that the performance of TWIST is contingent on the accuracy of the completion model used for restoring erasures, which may introduce additional complexity. They also note that the framework’s effectiveness may vary with different types of applications and communication environments, which could limit generalizability. An obvious limitation not discussed by the authors is the potential overhead introduced by the feedback mechanism, which may impact real-time performance in highly dynamic scenarios.

Why it matters
The implications of this work are significant for the development of application-aware wireless digital twins, particularly in scenarios where real-time synchronization is critical, such as autonomous driving and smart city applications. By optimizing the synchronization process based on task relevance and communication conditions, TWIST paves the way for more efficient and effective digital twin implementations. This research could influence future work in adaptive communication protocols and the design of intelligent systems that rely on real-time data synchronization.

Authors: Sige Liu, Kezhi Wang
Source: arXiv:2605.27205
URL: https://arxiv.org/abs/2605.27205v1