Abstract
Future vehicular applications, such as safety guarantee and autonomous driving, rely on vehicular-to-infrastructure (V2I) ultra-reliable low-latency communication (URLLC). This paper investigates the flow scheduling and power allocation mechanism to improve the transmission capacity of the downlink V2I orthogonal frequency division multiplexing (OFDM) URLLC network. Given the stringent latency requirements, short package transmission is adopted and the approximation of the finite blocklength codes capacity is introduced for the algorithm design. Also in the system design, we fully consider the effect of Doppler spread caused by high vehicular mobility. We formulate the problem of maximizing the number of flows that satisfy delay and reliability requirements while meeting the constrained radio and power resources. To solve this challenging non-convex problem, we propose a joint optimization framework for flow scheduling and power allocation. In the scheduling phase, we propose a deferred acceptance based flow scheduling algorithm by leveraging matching game. In the power allocation phase, we design a collection-reallocation algorithm for local power optimization while fully considering the dynamic characteristics of V2I scenarios. Numerical results show that the proposed scheme effectively enhances the system performance compared to other state-of-art mechanisms.