Data Scheduling for Multi-item Requests in Vehicle-Roadside Data Access with Motion Prediction Based Workload Transfer

摘要

Road Side Unit (RSU) has been proposed by researchers to enhance data dissemination in Vehicular Ad-hoc Networks (VANETs). In particular, there has been an increasing interest for vehicles to access data items stored at the RSU. At the same time, the rapid development of mobile applications has resulted in a growing demand for vehicles to access multiple data items from RSU at a time. When scheduling multi-item requests in vehicle-roadside data access, the broadcast nature of RSU to vehicle communication can be used to achieve better performance since a single broadcast can serve multiple requests which contain the same data item simultaneously. Meanwhile, vehicle mobility pattern can be made use of to facilitate multi-RSU data dissemination. Specifically, overloaded RSU can transfer part of its requests to nearby non-overloaded RSUs based on the motion prediction of vehicles. In this paper, we propose a motion prediction based scheduling scheme which enables cooperative work among a set of RSUs for scheduling multi-item requests. Data requests can be transferred and thus balanced among RSUs. Scheduling performance is also improved by taking advantage of broadcast. Simulation results show that our scheme outperforms classical scheduling schemes under various circumstances.