An Improved Clustering Method Vanet Routing and Communication

Abstract

The rising popularity of network based technologies has witnessed growing interests in research of inter-vehicular communications. In this scenario, Vehicular Ad-hoc Network (VANET) has evolved as the largely admired network traffic routing and control system. The main idea of this work is to enhance quality of clustering approach implemented in VANET. To achieve this, authors proposed the addition of two new elements in the existing clustering architecture, namely, the number of cluster heads required in a specific simulation area and the selection of paramount candidate to be used as cluster head. Supervised learning technique is implemented in the employed methodology. The proposed architecture is evaluated in terms of Jitter and Packet Delivery Ratio (PDR). The simulation results demonstrated that the node variation with PDR shows relatively higher average PDR for polynomial kernel as compared to the average PDR for linear kernel.