Joint Caching and Hovering Minimized Service Time for Cooperative Video Streaming in UAV-assisted VANETs

Abstract

The rapid growth of advanced applications and services (A\&Ss), particularly video streaming, imposes stringent latency and quality-of-experience requirements on vehicular ad hoc networks (VANETs). Although edge caching at roadside units (RSUs) can reduce service time, its effectiveness is limited in highly dynamic environments characterized by uneven vehicle user (VU) distributions, constrained caching resources, and sparse roadside infrastructure. To address these challenges, this paper proposes a caching and hovering (CAHO) optimization design for cooperative video streaming in VANETs assisted by unmanned aerial vehicles (UAVs). In the proposed CAHO design, VU distributions are modeled as independently thinned one-dimensional homogeneous Poisson point processes, and RSU-to-VU and UAV-to-VU communication models together with video popularity characteristics are incorporated. Based on this modeling, the CAHO problem jointly optimizes video caching at RSUs and UAVs as well as UAV hovering positions to minimize the service time under limited storage resources. The CAHO optimization problem is efficiently solved using genetic algorithms combined with a penalty function and a divide-and-conquer strategy. Simulation results demonstrate that the proposed CAHO scheme outperforms benchmark schemes under various system scenarios in terms of service time and storage resource utilization, highlighting the effectiveness and practical potential for video streaming A\&Ss in UAV-assisted VANETs.