Efficient Twin Migration in Vehicular Metaverses: Multi-Agent Split Deep Reinforcement Learning with Spatio-Temporal Trajectory Generation

TL;DR

This work tackles VT migration in vehicular metaverses under rapid mobility and heterogeneous RSU resources. It introduces a novel MSRL framework with split learning and an EST-TG trajectory generator to enable efficient pre-migration decisions while preserving privacy and reducing computation. The EST-TG algorithm produces realistic spatio-temporal trajectories for robust MADRL training, and the MSRL method—including dynamic switching and a switching buffer—improves QoE with roughly 29% gains and 25% fewer parameters than baselines. Collectively, the approach enhances immersive VT experiences and demonstrates strong generalization across dynamic traffic conditions and resource scenarios, offering practical gains for real-world vehicular metaverse deployments.

Abstract

Vehicle Twins (VTs) as digital representations of vehicles can provide users with immersive experiences in vehicular metaverse applications, e.g., Augmented Reality (AR) navigation and embodied intelligence. VT migration is an effective way that migrates the VT when the locations of physical entities keep changing to maintain seamless immersive VT services. However, an efficient VT migration is challenging due to the rapid movement of vehicles, dynamic workloads of Roadside Units (RSUs), and heterogeneous resources of the RSUs. To achieve efficient migration decisions and a minimum latency for the VT migration, we propose a multi-agent split Deep Reinforcement Learning (DRL) framework combined with spatio-temporal trajectory generation. In this framework, multiple split DRL agents utilize split architecture to efficiently determine VT migration decisions. Furthermore, we propose a spatio-temporal trajectory generation algorithm based on trajectory datasets and road network data to simulate vehicle trajectories, enhancing the generalization of the proposed scheme for managing VT migration in dynamic network environments. Finally, experimental results demonstrate that the proposed scheme not only enhances the Quality of Experience (QoE) by 29% but also reduces the computational parameter count by approximately 25% while maintaining similar performances, enhancing users' immersive experiences in vehicular metaverses.

Figures (11)

• Figure 1: The migration of VTs tasks with generated trajectories in the vehicular metaverse.
• Figure 2: Architecture of the EST-TG algorithm for spatio-temporal trajectory generation.
• Figure 3: Design of the MSRL Algorithm for Pre-migration of VT Tasks.
• Figure 4: Real vehicle trajectory spatial distribution data v.s. generated vehicle trajectory spatial distribution data.
• Figure 5: Comparison of real vehicle time distribution in a day against the generated vehicle time distribution.