Faheem Quazi

Abstract - Digital Twins (DTs) provide a framework for building virtual representations of physically co-located or remote assets, and can be used by an end-user for command, control, and observation. These DT systems normally rely on high bit rate availability and minimal communication delay to accomplish their tasks. However, providing such resources may be challenging depending on the location of the asset, e.g., in a remote environment such as the lunar surface. In such cases, the ability to obtain information from the asset to maintain an up-to-date DT will be limited by the available data rate or by the propagation delay. This paper provides a novel scheduling paradigm for a remote asset to “synchronize” state with its DT over a tightly constrained network by combining the Age of Synchronization (AoS) metric with the Multi-Armed Bandit (MAB) mathematical decision framework. This provides a general weighted solution for scheduling data to minimize the Age of Information (AoI). Preliminary simulations were developed to validate this MAB- based AoS-Driven weighted scheduler against other scheduling paradigms. Experimental results show the MAB-based scheduler trades off priorities and AoS to create a delay-tolerant DT.