Geolocation-based UAV-assisted Caching Strategies in Vehicular Networks

Abstract

In this paper, a geolocation-based UAV-assisted caching strategy is proposed to address the problem of large communication load of vehicular network in cities and extended time spent by vehicle users to request services. In Ad hoc Network (VANET), vehicle users may request content that is valid only in their location, such as maps, or content that is not related to geographic location, such as news. Based on this feature, we divide the content into local content and common content, let mobile vehicle users and UAVs cache only common content, and location-fixed roadside units (RSUs) cache local content and common content proportionally. Then we design a detailed content delivery method and analyze the content request latency of the vehicular network over a period of time, and establish an optimization problem for minimizing the average request latency. A joint optimization algorithm for UAV trajectory and node caching decision is designed to solve this problem. First, optimizing the proportion of common content cached by RSU through exhaustive method, then deciding the caching decision of each RSU and UAV through preference-based exchange collaborative caching algorithm, and finally using Greedy algorithm-based trajectory optimization algorithm to further optimize UAV trajectory. Simulation results show that this strategy outperforms existing common caching strategies.