Resource Scheduling in Vehicular Networks with Age of Information and Channel Awareness

Abstract

Vehicle-to-everything connectivity (V2X) has the potential to revolutionize the mobile experience, which has been widely concerned. Since Age of Information(AoI) is particularly impressive in characterizing information freshness, designing a resource scheduling policy to optimize the AoI in vehicular networks can satisfy the demand for real-time and reliable communications. However, optimizing AoI is challenging due to limited communication resources and unpredictability of time-varying channels. In this paper, we are concerned with a joint-aware scheduling policy that minimizes AoI while satisfying power and bandwidth constraints. More specifically, the complexity of V2X is reduced by applying Mean Field Theory (MFT). The scheduling problem is then described as a Constrained Markov Decision Process (CMDP), and further decouples the state transfer consumption of multiple sensors into a single source-to-destination linear problem (LP). Besides, we conceive an iterative algorithm incorporating sub-gradient descent that provides us with the optimal freshness-aware scheduling policy. Finally, an approaching optimal threshold policy that optimizes the AoI is built on the basis of the optimal solution of each decoupled problem.