VECADS: Vehicular Context-Aware Downstream Scheduling for Drive-Thru Internet

We study the downlink scheduling performance of an IEEE 802.11-based roadside Access Point(AP) serving a number of moving vehicles. For determining the scheduling order, throughput and Fairness are two important considerations. If system throughput Maximization is the sole consideration, some vehicles can always get resource from the AP whereas some are starved. On the other hand, if fairness is the sole consideration, resource should be allocated among all the contending vehicles regardless of individual vehicle's channel conditions and the system utilization should not be affected greatly. In this paper, we propose a novel scheduling algorithm called VEhicular Context-Aware Downstream Scheduling (VECADS). By exploiting the real-time vehicular context including the position, speed and cumulative data received by each vehicle, the scheduler can get up-to-date information to determine an appropriate scheduling order. The design objective of VECADS is to strike a better balance between system throughput and fairness among heterogeneous drive-thru vehicles. The performance of VECADS is evaluated via extensive ns2 simulations using real-world vehicular traffic traces. We show that VECADS outperforms MV-MAX, the state-of-the-art scheduling scheme for Drive-thru networks, in terms of system throughput, by 7 % while eliminating the bandwidth starvation problem of ``weak" vehicles under MV-MAX for a Jain's Index of 0.895 vs. 0.727.