An intersection-based clustering algorithm for vehicular ad hoc networks

Recent years, Vehicular ad hoc Network (VANET) attracts tremendous attention due to its broad application prospects. However, channel fading, frangible link, unstable network topology caused by vehicles' high mobility and strict requirements to the quality of service (QoS) have become major challenges of VANET. Many algorithms have been proposed to ensure a stable network topology and it has been proved that cluster-based VANET is more stable and robust. But existing cluster algorithms cluster only on straight lane and are unable to maintain cluster stable as vehicles coming into intersection. Accounting for existing algorithms' limitations, this paper proposes a new clustering algorithm, in which an intersection situation is taken into consideration when clustering. Beyond the intersection, relative mobility concluding relative speed and distance is considered as Cluster Header (CH) selection metric. While around the intersection, turn direction is also considered as a metric. Simulation results show that the proposed algorithm has a better performance in cluster lifetime, cluster size and cluster stability.