VANET Modeling and Clustering Design Under Practical Traffic, Channel and Mobility Conditions

In Vehicular Ad Hoc Networks (VANETs), vehicles driving along highways can be grouped into clusters to facilitate communication. The design of the clusters, e.g., size and geographical span, has significant impacts on communication quality. Such design is affected by the Media Access Control (MAC) operations at the Data Link layer, the wireless channel conditions at the Physical layer, and the mobility of the vehicles. Previous works investigated these effects separately. In this paper, we present a comprehensive analysis that integrates the three important factors into one model. In particular, we model an unsaturated VANET cluster with a Markov chain by introducing an idle state. The wireless channel fading and vehicle mobility are integrated by explicitly deriving the joint distribution of inter-vehicle distances. Closed-form expressions of network performance measures, such as packet loss probability and system throughput, are derived. Our model, validated by extensive simulations, is able to accurately characterize VANET performance. Our analysis reveals intrinsic dependencies between cluster size, vehicle speed, traffic demand, and window size, as well as their impacts on the overall throughput and packet loss of the cluster. Performance evaluation results demonstrate the practical value of the proposed model in providing guidelines for VANET design and management.