Efficient Double-layer Coverage Scheme for Multiple UAVs Assist Intelligent Transportation Systems

Sudden traffic congestion on road leads to a huge increasing in the computing tasks of vehicle users in Internet of vehicles, it is a challenge to meet the emerging real-time requirements of users. Due to fast movement and line-of-sight communication links, unmanned aerial vehicles (UAVs) are integrated into intelligent transportation systems (ITS) to assist the roadside units (RSUs). However, for the limited capacity of a single UAV, multi-UAV cooperative networking can optimizes task completed delay in ITS. In this paper, we construct vehicular edge computing networks with double-layer UAVs. According to the on-predicted historical data, the first-layer UAVs are pre-deployed at user aggregation points to receive and process tasks from covered user equipments (UEs). The second-layer UAVs hovers at service points. A joint optimization problem is proposed to minimize the task completed delay and the energy consumption of UAVs. In order to handle the continuous action space and discrete state in this problem, a deep deterministic policy gradient (DDPG)-based algorithm is proposed. It enables the system to determine the optima vehicle and UAV action strategy in dynamic traffic environments. Simulation results indicate that the proposed algorithm achieves rapid convergence compared to the normal single-layer UAVs coverage model.