Optimal Traffic Scheduling Between Roadside Units in Vehicular Delay-Tolerant Networks

This paper investigates the problem of traffic scheduling between roadside units (RSUs) in a vehicular delay-tolerant network. A source RSU needs the help of passing-by vehicles to forward its traffic to a destination RSU. Costs are associated with energy consumption and information loss. When a vehicle arrives, the source RSU needs to decide whether to stop, i.e., transmit its traffic to the vehicle, or skip this vehicle and continue to wait for other vehicles. The source RSU's objective is to achieve the minimal rate of cost. In this paper, we theoretically derive an optimal stopping strategy of the source RSU. Simulation results are presented to show the effectiveness of the derived strategy.

I. Introduction

TRAFFIC scheduling problems in vehicular delay-tolerant networks (VDTNs) have recently been attracting increasing research interests. One type of VDTN is installed in less-populated remote areas, which includes a number of roadside units (RSUs), and only a limited number of RSUs have connection to backbone networks [1]. The isolated RSUs (i.e., the RSUs without backbone network connection) are deployed to serve as gateways for sensor networks (for example, sensor networks for monitoring environment or wildlife [2]– [5]) in less-populated remote areas. Since it may be costly to set up direct communication connections from the isolated RSUs to backbone networks, passing-by vehicles may provide a solution: Passing-by vehicles can help forward traffic (e.g., sensed data) from isolated RSUs to RSUs that have connection to backbone networks [6]. When a vehicle arrives at a source RSU that is isolated, the source RSU may send its traffic to the vehicle, and then, the vehicle stores the traffic in its local buffer and forwards the traffic to a destination RSU with backbone network connection when the vehicle arrives at the destination RSU [7]. The destination RSU then forwards (through backbone networks) the traffic to a data center that processes the data.