Utilising in-vehicle information to detect traffic conditions in vehicular ad-hoc networks

Tools

Abufanas, A.E., 2019. Utilising in-vehicle information to detect traffic conditions in vehicular ad-hoc networks. PhD, Nottingham Trent University.

Preview

Text
Thesis_Final_Version_AA.pdf - Published version
Download (6MB) | Preview

Abstract

In the last decade, cooperative vehicular network has been one of the most studied areas for developing the intelligent transportation systems (ITS). It is considered as an important approach to share the periodic traffic situations over vehicular ad hoc networks (VANETs) to improve efficiency and safety over the road. In addition to the uses of ITS, VANETs will contribute in service access, cooperative driving, entertainment and navigation for cars of the future. Vehicle to vehicle and vehicle to infrastructure communication are two distinct avenues that make possible efficient delivery of messages through direct wireless transmissions in traffic regions. Furthermore, promising quality of communication performance is desirable for a communication system composed mostly if roaming participants; such a system needs to be dynamic, flexible and infrastructure-less. Thus VANET architecture is a natural fit for ITS. However, there are a number of issues in exchanging traffic data over high mobility of VANET, such as broadcast storms, hidden nodes and network instability.

Therefore, vehicular traffic efficiency applications have been investigated recently using VANET. This aspect of research is primarily concerned with increasing the traffic awareness over roads. In this thesis, a novel model, Efficient Traffic Conditions Detection (ETraCD) is proposed to detect the traffic conditions utilising vehicles’ characteristics and in-vehicles sensors information to evaluate traffic situations that are gathered from the nodes (vehicles) in VANET.

The model revolves around the core idea to what extent we will be considering the traffic characteristics between groups of cars rather than individual cars. This does not concern the physical transmission of data but the data processing in the network. More precisely, vehicles are clustered into traffic groups based on the similarity of sensors’s data. ETraCD (a) divides the situations of vehicles into clusters, (b) designs a set of metrics to get the correlations among vehicles and (c) detects the traffic condition in certain areas. These approaches have been simulated in NS3 network simulator to investigate the performance of stability of the network, latency, and the accuracy of traffic situations detection.

The proposed model applies V2V clustering paradigm for detecting traffic conditions, it has been implemented and its features investigated through simulation runs. It shows the benefit of using the vehicular sensors informations such as ABS, windscreen lights and so on based on V2V communication to provide an efficient traffic conclusion in urban environment. Experiments also show improved overall performance when compared to previous protocols.

Item Type:	Thesis
Creators:	Abufanas, A.E.
Date:	June 2019
Rights:	© 2018 by Ayman Abufanas. I confirm that this is intellectual work and the use of all materials from other sources has been properly and fully acknowledged. You may use up to 5 percent of this study for personal research, or private, non-commercial study. Any re-use of the context contained with this thesis should be fully referenced.
Divisions:	Schools > School of Science and Technology
Record created by:	Linda Sullivan
Date Added:	31 Jul 2019 13:27
Last Modified:	31 Jul 2019 13:27
URI:	https://irep.ntu.ac.uk/id/eprint/37169