Application de l’Approche Clustering dans le Contexte de l’IoV

Abstract

This PhD thesis deals mainly with the problem of optimizing networking overhead and improving network stability in the Internet of Vehicles (IoV) environment. In this context, we are interested in the clustering approach to meet the requirements of such environment. Therefore, the general goal of this research work focuses on the design, modeling and simulation of new clustering algorithms for the IoV network. In this respect, we start with an analysis, through a critical review of the literature, a number of the clustering algorithms in VANETs, to detect certain problems still open or whose proposed solutions are still to improve. Then, we propose three new clustering algorithms for IoV networks, which make it possible to improve their performances, while guaranteeing a high level of service required by the function performed by these networks. The objectives of the proposed algorithms is to minimize significantly the network overhead, decrease the network latency and increase the delivery ratio between a large number of connected vehicles traveling at high speed. The analytical study and the simulation for evaluating the proposed algorithms, carried out by a combination of the network simulator NS2 and mobility generator VanetMobiSim, have yielded convincing results, outperforming those exhibited by the basic referred algorithms.