Context‐Aware Synchronisation of Digital Twins in the Metaverse: Balancing Freshness, Energy and Network Efficiency

The convergence of digital twin (DT) technology with metaverse ecosystems introduces unprecedented challenges in maintaining real‐time synchronisation across heterogeneous, large‐scale and dynamic virtual‐physical systems. Existing synchronisation approaches—whether cloud‐centric, edge‐based, or predictive—fail to simultaneously ensure data freshness, energy efficiency and network scalability. This paper presents a context‐aware adaptive synchronisation framework that integrates hierarchical federated learning, transfer learning‐enhanced double deep Q‐networks (TL‐DDQN) and semantic knowledge transfer to optimise DT interactions in the metaverse. The proposed cloud‐centric architecture dynamically balances Age of Information (AoI), bandwidth usage and sensor energy through event‐driven decision‐making and R‐tree spatial indexing. Experimental results across four metaverse domains (gaming, industrial, educational and social) demonstrate a 62% reduction in energy consumption, a 45% decrease in network traffic and a 34.7% latency improvement compared to baseline methods. Statistical validation via repeated‐measures ANOVA confirms the significance of these gains ( p < 0.001). The proposed framework advances resource‐efficient synchronisation and establishes a scalable foundation for sustainable, intelligent and context‐aware metaverse infrastructures.