A scenario based performance analysis of VANET protocols

Abstract

Inter-vehicle communication is a major part of the Intelligent Transportation Systems (ITS). Vehicular Ad-Hoc Network (VANET) was primarily developed to improve safety and comfort for vehicles, passengers and drivers. Because of high mobility and changeable infrastructure, there are many challenges in such networks one of which is routing. Existing routing algorithms for VANET are divided into five major classes: position based, cluster based, broadcast,geocast based, and topology based routing protocols.

Different researches for ITS compare existing routing protocols for VANET and evaluated their performance. Nevertheless, most of them are far from realistic conditions. Since these systems often consist of many nodes, a real-world test is very costly and time consuming. Therefore, most VANET researches use simulators which allow fast and cheap evaluation of protocols and applications. The simulation scenarios are controllable and reproducible. In simulation studies models are used to make a judgement on real-world viability. The model must reflect the real-world conditions to make the results reliable.

In this thesis we provide a realistic model for Auckland using Nakagami propagation model and evaluate the performance of several popular VANET protocols (AODV, DSR, OLSR,DSDV,GPSR, CBRP, and ZRP). We use Nakagami propagation model and investigate the impact of the shape factor of the Nakagami model on the performance of each protocol for two real scenarios: urban area and highway. For the urban area, we select Auckland CBD with maximum speed of 50 km/h. For the highway area, we select Auckland motorway with maximum speed of 100 km/h.

The simulations are carried out using OMNET++ and SUMO simulators, with scenarios configured to reflect real-world conditions. We compare the performance of the protocols using three metrics: packet error ratio, end-to-end delay and throughput.

The experimental results show that the performance of the protocols depend on several factors which include: number of vehicles, speed of vehicles, shape factor, etc. In general, DSR has the highest throughput for both scenarios in all conditions. In terms of packet error rate, for most of the cases in the CBD scenario, CBRP outperforms other protocols while for the motorway scenario, in most of the cases, the best protocol is ZRP. The lowest End to End Delay (EED) is achieved for both scenarios when using OLSR.