An Adaptive Multipath Geographic Routing for Video Transmission in Urban VANETs

Abstract

Vehicular ad hoc networks (VANETs) have attracted many researchers' attention in recent years. Due to the highly dynamic nature of these networks, providing guaranteed quality-of-service (QoS) video-on-demand (VOD) sessions is a challenging problem. In this paper, a new adaptive geographic routing scheme is proposed for establishing a simplex VOD transmission in urban environments. In this scheme, rather than one route, a number of independent routes are discovered between source and destination vehicles whose number of routes depends on the volume of the requested video and lifetime (span of time in which a route is almost fixed) for each route. A closed-form equation is derived for estimating the connectivity probability of a route, which is used to select best connected routes. Simulation results show the QoS parameters: packet loss ratio is decreased by 40.79% and freezing delay is significantly improved by 25 ms compared with those of junction-based multipath source routing at the cost of 2-ms degradation in the end-to-end delay.