A Geographical Self-Organizing Approach for Vehicular Networks

Abstract—Cooperative vehicular networks have always been considered as the perfect way to bring more comfort to the passengers and more safety to the human life. Thus, research community and governmental organizations are interested to study and deploy these networks. The vehicular networks principle is connecting vehicles to each other and to existing infrastructure. However, their industrialization faces some challenges: (i) high mobility, (ii) frequently partitioned network, (iii) geographically constrained topology, and (iv) scalability. Therefore, in contrast to traditional networks, vehicular network protocols focus on both achieving adequate QoS level and reducing overhead. Achieving these two opposite requirements was the key driver of this work. The most promising way to do it is to self-organize the network. In a previous work, we introduced a proactive self-organizing architecture for vehicular networks called “CSP ” (Cluster-based Self-organizing Protocol). In this paper, besides detailing CSP function, we define a mathematical model to estimate CSP overhead and to show the effects of the different parameters on it. We also set up a developed simulation study to validate the mathematical model and to compare CSP to other self-organizing solutions. This study shows interesting results of CSP in terms of generated overhead, end-to-end delay and delivery ratio. Index Terms — vehicular networks, self-organization, clustering, virtual backbone, performance evaluation, analytical study. I