Joint Client Association and UAV Scheduling in Cache-Enabled UAV-Assisted Vehicular Networks

Abstract

In the case of explosive content requests being generated by numerous vehicles in rush hour, the cellular downlink resources become insufficient. This paper considers two supplementary schemes for exploiting the uplink resources: One is letting vehicles cache the browsed contents so that the vehicles can share contents with one another; and the other is dispatching a cache-enabled UAV to transmit contents to the vehicles nearby. We formulate a joint optimization problem for maximizing the average data rate of vehicles, and then decompose it into the subproblems of client association and UAV scheduling. The client association aims at matching the vehicles to the resources, and the UAV scheduling is to update the UAV’s caching and trajectory to adapt to the environment changing. The two subproblems focus respectively on the current condition and the long term profit, and are solved respectively by a matching-based algorithm and a deep reinforcement learning-based algorithm. Our simulation results based on a real-world traffic data set demonstrate the advantages of the proposed approaches.