Encounter Prediction-based Data Forwarding for High Reliability in Bus Networks
Lianhai Liu, Jianxin Wang, Jiawei Huang, Qilong Feng and Geyong Min
Data forwarding is a critical and challenging issue in Vehicular Ad Hoc Networks (VANETs) due to the rapid vehicle mobility and short-lived communication connectivity. Many data forwarding schemes have been developed by exploiting vehicular traffic statistics or predictable vehicle mobility. However, most of the existing studies have focused on the impact of encounter probability between vehicles on only packet delivery ratio or delivery delay but paid less attention to achieve higher packet delivery ratio under the delay constraint. To fill in this gap, this paper proposes an efficient data forwarding scheme for bus networks. Specifically, a new Predicted Encounter Graph (PEG) is constructed by exploiting the shared travel information and predictable mobility of buses in both the opposite direction and also the same direction to predict the potential encounter time and the encounter probability between any pair of buses. A novel Reverse Search (RS) algorithm based on PEG is then designed to improve the data forwarding performance in terms of packet delivery ratio and delay. Moreover, amendment measures are taken into account in order to weaken the impact of inaccurate prediction of encounter time. We performed the extensive simulation experiments based on real bus GPS traces. The results demonstrate that RS not only improves the delivery ratio by 20%, but also reduces the delivery delay compared to the existing data forwarding schemes.
Keywords: Vehicular Ad hoc networks, data forwarding, predicted encounter graph, delivery ratio, delivery delay.