PolyTwin: Edge Blockchain-empowered Trustworthy Digital Twin Network for Metaverse

Yinfeng Cao, Jiannong Cao, Zeyang Cui, Dongbin Bai, Mingjin Zhang, Long Wen

Abstract: Digital Twin (DT) is increasingly being adopted into the metaverse to enhance its immersiveness by creating virtual representations of physical objects. A major unsolved challenge of this adoption is ensuring the trustworthiness of DT, which requires that the mapped virtual object is strictly consistent with its corresponding physical object. However, existing metaverse/DT platforms either rely on specialized equipment or use a centralized solution, which is infeasible to be applied on a large scale. In this paper, we study the trustworthy issue of DT for the metaverse. We propose PolyTwin, a novel edge blockchain-enabled trustworthy DT network for the metaverse. PolyTwin deploys sensors and edge devices with AI models in the physical world to collect multiple attributes of physical objects for efficiently generating digital twins. To further enhance the trustworthiness of the generated DT, we introduce a Proof of Consistency (PoC) mechanism. PoC iteratively verifies whether edge devices produce consistent inference results, thus avoiding inconsistent results caused by model inaccuracy or malicious edge devices. Lastly, we evaluate PolyTwin through its two prototypes, PolyCampus and PolyExchange, which both enable immersive physical-to-virtual interaction in the metaverse. Experimental results show that PolyTwin is practical in terms of latency and security.