Markov-reward based estimation of the idle-time in vehicular networks to improve multimetric routing protocols

Abstract

Analyzing vehicular ad hoc networks (VANETs) poses a considerable challenge due to their constantly changing network topology and scarce network resources. Furthermore, defining suitable routing metrics for adaptive algorithms is a particularly hard task since these adaptive decisions should be taken according to the current conditions of the VANET. The literature contains different approaches aimed at optimizing the usage of wireless network resources. In a previous study, we introduced an analytical model based on a straightforward Markov reward chain (MRC) to capture transient measurements of the idle time of the link formed between two VANET nodes, which we denote as $T_{idle}$. This current study focuses on modeling and analyzing the influence of $T_{idle}$ on adaptive decision mechanisms. Leveraging our MRC models, we have derived a concise equation to compute $T_{idle}$. This equation provides a quick evaluation of $T_{idle}$, facilitating quick adaptive routing decisions that align with the current VANET conditions. We have integrated our $T_{idle}$ evaluation into multihop routing protocols. We specifically compare performance results of the 3MRP protocol with an enhanced version, I3MRP, which incorporates our $T_{idle}$ metric. Simulation results demonstrate that integrating $T_{idle}$ as a decision metric in the routing protocol enhances the performance of VANETs in terms of packet losses, packet delay, and throughput. The findings consistently indicate that I3MRP outperforms 3MRP by up to 50% in various scenarios across high, medium, and low vehicular densities.