Resource Allocation for Platoon Digital Twin Networks

Vehicle platooning involves a group of vehicles driving in close coordination, maintaining short inter-vehicle distances to improve road capacity, reduce fuel consumption for following vehicles, and enhance overall driving safety. Achieving this coordination requires continuous exchange and processing of environmental sensor data. To further enhance control and service performance without overloading individual vehicles, maintaining digital twins (DTs) of platooning vehicles has emerged as a promising approach. A platoon digital twin (PDT), which integrates the DTs of all platoon members, can serve as a unified interface for coordinated traffic management. However, the effectiveness of PDT-based applications relies heavily on the quality of the PDT, which in turn depends on timely and accurate synchronization with the physical platoon that requires robust communication and computation resources. In this talk, we present our recent work on joint communication and computation resource allocation to support high-quality PDTs under highly dynamic vehicle mobility. We model the problem as an M-th order Markov Decision Process (MDP) to better capture the temporal dynamics of the system. Our solution leverages a multi-agent Deep Deterministic Policy Gradient (DDPG) framework, enhanced with temporal feature extraction, to adapt to rapidly changing network conditions and improve resource allocation <a href="http://decisions.

Speaker(s):” target=”_blank” title=”decisions.

Speaker(s):”>decisions.

Speaker(s): Dongmei Zhao,

Virtual: https://events.vtools.ieee.org/m/481625