Abstract
The emergence of immersive metaverse applications in vehicular environments poses unprecedented challenges for wireless networks due to their stringent quality of service (QoS) requirements, including ultra-high data rates and low latency. Single mobile network operators (MNOs) often struggle to meet these demands, particularly in dynamic vehicular scenarios where network conditions rapidly fluctuate. This paper proposes a novel framework that combines digital twin (DT) technology with multi-agent reinforcement learning to enable intelligent collaboration among multiple MNOs. This collaborative approach optimizes resource allocation, thereby enhancing overall network performance. To support this framework, real-world measurement data from major United Kingdom operators (EE, Vodafone, O2, and Three) are used both to analyze the limitations of individual MNO performance and to construct digital replicas of the network environment. These DTs enable accurate prediction of achievable data rates and support data-driven decision-making. Experimental results demonstrate that collaboration among multiple MNOs is essential for ensuring robust coverage and meeting the demanding QoS requirements of vehicular metaverse applications. These findings highlight the critical role of multi-operator cooperation and artificial intelligence-driven orchestration as foundational components for future 6G networks supporting the metaverse.