A Novel Edge Computing Deployment Scheme Using Unmanned Aerial Vehicles and Buses for Internet of Vehicles

Deploying edge servers on Unmanned Aerial Vehicles (UAVs) has become a promising strategy to handle the spatiotemporally varying user demands on ground. However, a critical limitation of the UAVs is the limited flight endurance caused by the limited power supply. To this end, we propose in this paper a novel edge computing deployment scheme using UAVs and buses for Internet of Vehicles. The UAV-mounted edge servers (UAVESs) act as quick-response mobile servers to serve the workload hotspots, while buses offer energy-efficient transportation and charging service for these UAVs. Firstly, we use Temporal Mean Shift Clustering (TMSC) to identify the workload hotspots on the map of an IoV system at different time slots. Next, we design a Bus-Assisted UAV Scheduling (BA-UAVS) strategy to serve the user demand on the map. BA-UAVS selects a minimum number of bus routes to cover the maximum number of hotspots, and then schedules the UAVs along with the operation of the selected buses to reduce the fly distance, i.e., the cost. Finally, we conduct extensive trace-driven simulations on a real dataset, and the experimental results verify that our approach improves the service utilization and reduces the cost significantly compared with the approaches with no bus assistance.