A New Method for Predicting Cluster Stability in VANET Based on the Birth-Death Process

Abstract

One of the main characteristics of a Vehicular Ad-hoc Networks (VANET) is high mobility. Therefore, the high overhead and end-to-end delay are natural results. The prediction of future movements for vehicles provides traffic observations to improve and enhance the network performance and to serve the VANET applications. The clustering approach is a popular technique used in most studies to deal with the network's dynamic topology. In most studies, the clustering algorithms are tested and evaluated according to the cluster head (CH) selection approach. Selecting a suitable CH will lead to the prolonging of the cluster lifetime. A long life time represents an indication for cluster stability, also known as cluster maintenance. Cluster maintenance has got less consideration by the previous VANET studies. The stability prediction represents a challenging task in the clustering approach. This paper highlights the stability prediction in the VANET’s Highway scenario. It proposes a new approach to evaluate the vehicle’s cluster stability prediction by utilizing the stochastic process. The modeling of joining and leaving takes place using the birth-death process as a stochastic process for evaluating the stability of the clustering algorithm. The main advantage of this approach is to predict and evaluate the vehicle cluster stability with time, using a new mathematical model.