Enhanced approach for performance analysis in Vehicular Communications leveraging Taguchi and ANOVA methods

Abstract

In Vehicular communications, routing the information efficiently and accurately is still a major concern for many researchers, owing to the high mobility of vehicles, the frequent changes in the network topology and the security vulnerabilities. In that regard, the present work aims at analyzing the influence of scalability, connectivity and routing attacks on routing protocols in vehicular communications, by applying two statistical approaches: The Taguchi Method and ANOVA analysis. The Taguchi Method is used to design the experiments by means of an L8 Orthogonal Array, the obtained experiment results were then analyzed using Taguchi Signal to Noise Ratio (SNR) analysis. Additionally, ANOVA method was implemented in order to consolidate the analysis and confirm the findings of the Taguchi Method. The current study considers investigating the effect of four factors on Throughput, Average delay, Packet Loss, and Overhead. We selected NS-3 as a Network Simulator and SUMO for generating a realistic vehicular mobility model. The implementation of Taguchi Method enabled to simultaneously discern the effect of factors and possible interactions on the considered responses, through a reduced number of experiments. Moreover, the ANOVA method helped in determing the contribution and the statistical significance of the different investigated factors.