Joint smooth trajectory design and wireless communication control for mobile internet of vehicles assisted by a UAV and ground RISs

Low latency, reliable, and stable communication are essential for autonomous driving and mission accomplishment of Internet-of-Vehicles (IoVs) in smart cities. Therefore, future wireless networks will work based on quasi-optic millimeter wave (mmWave) signals for high-rate data transfer. However, given the mobility of vehicles, the mmWave links are prone to outages as they intrinsically rely on directional beamforming to line-of-sight (LoS) paths. Notably, fragile wireless links in dense urban canyons expose autonomous vehicles to safety risks. An unmanned aerial vehicle (UAV) equipped with a reconfigurable holographic surface (RHS) is navigated for establishing aerial LoS links for IoVs. RHS performs beamforming by adjusting the radiation patterns through the holographic surface, so it is energy efficient. The UAV-RHS is supported by terrestrial reconfigurable intelligent surfaces (RISs) installed on building facades, which are utilized to improve coverage and link reliability. The UAV’s navigation objectives are maintaining valid LoS links for IoVs, ensuring quality of service, and minimizing energy consumption. However, an obstacle-free kinematics-aware smooth trajectory, subject to motion constraints, is required for UAV navigation in dense urban environments. Satisfying these navigation objectives and constraints makes the trajectory design with valid LoS links a non-convex NP-hard optimization problem. To address this, we propose, for the first time, training generative adversarial networks (GANs) to generate valid paths in real time. State feedback control with quadratic optimization is proposed to smooth the trajectory. Simulation results are provided to evaluate the proposed method.