8 October 2025 in IEEE Transactions on Network Science and Engineering
Digital Twin (DT) creates virtual replicas of physical objects and systems for cost-efficient simulation and prediction applications in virtual space. In DT systems, ensuring consistency between physical objects and corresponding digital twins during both generation and adoption in applications is an essential requirement for high-fidelity modeling. However, existing solutions typically rely on centralized parties, which lack verification mechanisms or suffer from trust issues, thus being vulnerable to manipulation attacks in practice. Decentralized blockchain-based DT approaches often fail to verify the authenticity of physical object states, which likewise limits their practical effectiveness. To solve this digital twin consistency problem, this paper presents ChainTwin, a trustless blockchain-based framework for consistent DT generation and cross-domain adoption. With ChainTwin, a novel Digital Twin Generation Machine (DTVM) is designed to produce verifiable cryptographic proofs, and a two-dimensional selective verification mechanism is employed to efficiently cross-validate generation results. Additionally, ChainTwin achieves end-to-end consistency guarantees across the entire DT lifecycle through a lightweight cross-domain verification protocol. Experimental results from our PolyWorld prototype demonstrate ChainTwin's practicality, indicating a low-latency verification process in bandwidth-constrained environments. Security analysis confirms that ChainTwin maintains consistency guarantees under Byzantine adversarial assumptions.
17 October 2025 in Cluster Computing
Blockchain wallet manages decentralized identities in Web3, which allows users to identify themselves in a verifiable yet anonymous manner by digital signatures. However, native wallet scheme has been proven vulnerable to Sybil attacks in practice, where adversaries can easily create large numbers of controlled wallets at a low cost to undermine Web3 applications. To this end, many solutions are proposed for mitigating Sybil attacks. Nevertheless, they typically rely on either costly specialized biometric hardware or behavior-based heuristics that can be easily bypassed. In this paper, we introduce EdenDID, the first trinity-bound identity management system that uniquely binds human user, wallet address, and physical device into a unified framework. EdenDID combines edge-based video recognition, on-chain transactional activity analysis, and device computational power verification to establish a decentralized Proof-of-Trustworthiness consensus. The system provides compliant users with a trusted, verifiable credential to prove their unique identity. We prototype low-cost Eden Devices and deploy them on a Layer-2 blockchain network. Our experiments demonstrate the low end-to-end latency and robust resistance performance to Sybil attacks. Moreover, we present a case study to show how EdenDID can mitigate virtual-hardware fraud in DePIN, in which EdenDID successfully binds the user, wallet, and device uniquely, thus preventing users from binding multiple wallets or devices.