Review ArticleTraffic density estimation in vehicular ad hoc networks: A review
Introduction
Vehicular Ad hoc Network (VANET) is a promising Intelligent Transportation System (ITS) technology. Although safety is the main motivation for the development of VANETs, they also can be used in many applications such as managing traffic flow, monitoring road conditions, dissemination of data and many more [1], [2].
Based on traffic flow concepts, traffic density estimation is defined as the number of vehicles per unit length of a road (e.g. vehicles per km, or vehicles per km per lane) [3]. However, in some research papers vehicle density is defined as the number of vehicles in a certain area of a city, which is measured as vehicles/km2 in [4], [5], [6]. In addition, in some few cases vehicle density was defined as the number of vehicles per cell, whereas cells are areas defined based on the used clustering algorithm [7], [8]. Traffic flow theory depends mainly on the relationships between three quantities: density of vehicles on road, traffic flow, and vehicles speed [9].
Density of vehicles is an essential traffic condition metric used in many traffic information systems. For instance, the Infrastructure-Free Traffic Information System (IFTIS) [10] was designed to provide vehicles with an estimation of traffic density in city roads on a segment-to-segment basis. The application of such systems enables roads congestion prevention and control, as well as reductions in fuel consumption, gas emissions and travel time [11]. A substantial survey of both infrastructure-based and infrastructure-free cooperative traffic information systems, based on vehicular ad hoc networks, was presented in [2]. On the other hand, several density-aware routing protocols, such as Density Aware Routing using Road Hierarchy (DAR-RH) [12] and Road Vehicle Density based VANET Routing [13], depend mainly on measuring real-time road vehicle density. In order to provide fast and reliable communications, such protocols adapt to the variable density of vehicles in city environment. Moreover, various broadcasting schemes [14], [15], [16] rely on traffic density estimation to adapt their broadcast decision dynamically in order to mitigate broadcast storms in vehicular ad hoc networks.
Conventionally, most of the applied vehicle density estimation mechanisms were designed to be used in infrastructure-based traffic information systems, which require the deployment of vehicle detection devices such as inductive loop detectors, traffic surveillance cameras or wireless vehicle sensors. However, these mechanisms suffer from low reliability and limited coverage (i.e. they can only be aware of traffic density in the deployment coverage area). In addition, such mechanisms require high deployment and maintenance costs, also some of these density estimation mechanisms are not capable of calculating density estimations in real-time [5].
Nowadays, most newly manufactured vehicles are equipped with built-in wireless communication capabilities. For instance, Toyota and Microsoft announced a new partnership to connect Toyota vehicles to cloud computing. Thus, Toyota vehicles will be equipped with the latest technology to access telecommunications information and other online services while on the road. The idea is to connect vehicles to servers around the world to send and receive content and information [17]. In 2013, General Motors announced that all its new vehicles will be equipped with built-in Wi-Fi [18]. As the number of vehicles equipped with wireless communication devices upturns rapidly road traffic density can be measured in a more accurate and real-time mechanisms [11]. Such mechanisms are infrastructure-free as they require no extra deployment for special sensing or detection equipment. In other words, they depend only on the vehicle’s built-in wireless communication devices and Road Side Units (RSU). In addition, providing VANET users a variety of applications and services, which depend on vehicle density, will be the highest motivation for to employ vehicle density estimation mechanisms as part of software packages. Accordingly, the efficiency of such estimation mechanisms and VANET communication capabilities will increase, which facilitate early deployment and market penetration [19]. Thus, with the development of more density aware applications more and more vehicles will employ density measurement mechanisms.
This paper reviews the most recent techniques and algorithms used to measure and estimate on road vehicles density. The first section gives a brief introduction and highlights the importance of vehicle density estimation. Moreover, the concept of infrastructure-based and infrastructure-free density estimation was introduced. The second section points out the different mechanisms of infrastructure-based techniques used to measure vehicle density and explores requirements and main limitations of such techniques. The third section, describes the different classification ways to classify density estimation methods. In addition, it explains the classification used in this review paper and the reasons for using it. Afterwards, the fourth section emphasizes on investigating the three categories of infrastructure-free vehicle density estimation algorithms. Finally, section five is the conclusion of this work.
Section snippets
Infrastructure-based vehicular density estimation techniques
Traditionally road traffic density has been estimated using a number of techniques such as roadside magnetic loop detectors, surveillance cameras, wireless vehicle sensors, pressure pads, roadside radar and infra-red counters. However, these techniques require the detection devices to be installed in advance, at certain locations [11], [20]. Consequently, the infrastructure-based vehicular density estimation techniques are limited to regions where detection equipment are deployed.
In [21] video
Classification of infrastructure-free traffic density estimation methods
Infrastructure-free methods can be categorized into two main types, described as follows:
- •
Centralized vehicular density estimation methods: which adopt the principle of aggregating data at centralized locations for processing and then disseminating the results across the network. However, the centralized approach not only makes the system vulnerable to single point of failure, but also makes it harder to implement such system for distributed self-organized vehicular ad hoc networks. In addition,
Infrastructure-free vehicular density estimation algorithms
As infrastructure-based density estimation algorithms are not adequate for the highly dynamic and scalable environment of VANETs, many infrastructure-free approaches has been proposed in recent years. Some of these approaches formulate the problem of vehicle density estimation and solve it using statistical methods. Additionally, other schemes depend mainly on the communication and cooperation between vehicles to exchange traffic information, which can be used in density estimation.
Conclusions
In this paper the main approaches to road vehicle density estimations have been reviewed. Although the infrastructure-based density estimation techniques have been applied in many areas around the world, they still encounter many limitations such as limited coverage, deployment and maintenance costs as well as long time consumption in terms of collecting, processing and disseminating traffic related information (e.g. vehicle density estimations). Special attention has been devoted to
Tasneem Darwish is a Ph.D. student at the Faculty of Computing, University Technology Malaysia, Malaysia. She is a member of the Pervasive Computing research group. She received her B.Sc. degree in computer engineering from Islamic University of Gazah in 2005 and M.Sc. degree in electronics and electrical engineering from the University of Glasgow, United Kingdom, in 2007. From 2008 to 2013 she was working as a research and lecturer assistant at the Faculty of applied engineering, University of
References (60)
- et al.
Vehicular telematics over heterogeneous wireless networks: a survey
Comput. Commun.
(2010) - et al.
Road traffic density estimation using microscopic and macroscopic parameters
Image Vis. Comput.
(2013) - et al.
A traffic state detection tool for freeway video surveillance system
Procedia – Soc. Behav. Sci.
(2013) - et al.
Estimation of traffic density on urban freeways
J. Transport. Syst. Eng. Inform. Technol.
(2008) - et al.
Monitoring the evolution of traffic on main city roads
J. Transport. Syst. Eng. Inform. Technol.
(2013) - et al.
Traffic state estimation and uncertainty quantification based on heterogeneous data sources: a three detector approach
Transport. Res. Part B: Methodol.
(2013) - et al.
Traffic congestion detection in large-scale scenarios using vehicle-to-vehicle communications
J. Netw. Comput. Appl.
(2013) - et al.
Reducing emergency services arrival time by using vehicular communications and evolution strategies
Expert Syst. Appl.
(2014) - et al.
Improving path duration in high mobility vehicular ad hoc networks
Ad Hoc Netw.
(2013) - et al.
A survey of vehicular-based cooperative traffic information systems
Introduction to Modern Traffic Flow Theory and Control
A novel traffic information system for VANET based on location service
I-VDE: a novel approach to estimate vehicular density by using vehicular networks
An infrastructureless approach to estimate vehicular density in urban environments
Sensors
Towards efficient geographic routing in urban vehicular networks
IEEE Trans. Veh. Technol.
Enhanced junction selection mechanism for routing protocol in VANETs
Int. Arab J. Inf. Technol.
Store-carry-forward message dissemination in vehicular ad-hoc networks with local density estimation
An infrastructure-free traffic information system for vehicular networks
Density aware routing using road hierarchy for vehicular networks
A VANET routing based on the real-time road vehicle density in the city environment
DECA: density-aware reliable broadcasting in vehicular ad hoc networks
A dynamic broadcast scheme of emergency warning message in vehicular ad hoc network
IEICE Trans. Commun.
Traffic density-based broadcast scheme for vehicular ad hoc networks
IEICE Trans. Commun.
IPv6 address autoconfiguration in geonetworking-enabled VANETs: characterization and evaluation of the ETSI solution
EURASIP J. Wireless Commun. Netw.
Automatic traffic density estimation and vehicle classification for traffic surveillance systems using neural networks
Math. Comput. Appl.
Extracting vehicle density from background estimation using Kalman filter
Cited by (107)
A new approach to decentralized constrained nonlinear estimation over noisy communication links
2024, Journal of Mathematical Analysis and ApplicationsDistributed nonlinear estimation over time-varying directed networks
2023, Information SciencesTraffic density estimation using vehicle sensor data
2022, Journal of Intelligent Transportation Systems: Technology, Planning, and OperationsCenter clustering network improves piglet counting under occlusion
2021, Computers and Electronics in AgricultureCitation Excerpt :Counting CNN (CCNN) and Hydra CNN estimated the number of vehicles from image patches without a perspective map (Oñoro-Rubio and López-Sastre, 2016). However, almost all these studies focus only on pedestrians and vehicles (Darwish and Abu Bakar, 2015; Wang, 2021). Nonetheless, a few studies have investigated automatic counting methods for pigs.
Smart vehicular traffic management: An edge cloud centric IoT based framework
2021, Internet of Things (Netherlands)Citation Excerpt :Table 1 presents the comparative analysis of smart traffic management and other related works based on various classifications like the technologies it employed: cloud computing (CC), internet of things (IoT), and fog or edge computing (FC/EC). The study also reviewed the cited works’ approaches based on other parameters like real-time aspect (RTA) of the approach, service provisioning (SP) to people, area-wide analysis (AWA), and type of traffic estimation (ToTE) [15]. The analysis contrasts various works based on values: Y means yes (the work includes that parameter or technology); N means no (the work doesn’t include that parameter or technology); LW means lane-wise (this value signifies that the traffic analysis is based on lane-wise unlike of area-wise); IFB means infrastructure based traffic estimation (where, the traffic is analyzed on infrastructure devices like RSUs or centralized processing units at intersections); IFF means infrastructure free traffic estimation (where traffic parameters are sensed and relayed to some kind of server processing like in cloud computing), and HYB means the traffic estimation employs both IFB and IFF approaches.
A Novel Routing and Hybrid Based Clustering Scheme in vehicular adhoc networks
2021, International Journal of Intelligent NetworksCitation Excerpt :The paper concluded with section 7. The [3] proposed a clustering scheme termed as precedence-based clustering where every clustering framework is resolute by the physical position related data and the precedence is allocated to the vehicles. Every node transmits the data associated with them and adjacent nodes with elements such as the ID of the node, the ID of cluster head, position ID, the ID of preceding nodes along with the route to CH and node precedence.
Tasneem Darwish is a Ph.D. student at the Faculty of Computing, University Technology Malaysia, Malaysia. She is a member of the Pervasive Computing research group. She received her B.Sc. degree in computer engineering from Islamic University of Gazah in 2005 and M.Sc. degree in electronics and electrical engineering from the University of Glasgow, United Kingdom, in 2007. From 2008 to 2013 she was working as a research and lecturer assistant at the Faculty of applied engineering, University of Palestine. Her current research focuses on vehicular communications, wireless ad hoc networks, and mobile computing.
Kamalrulnizam Abu Baker is an Associate Professor in Computer Science at University Technology Malaysia, Malaysia, and member of the Pervasive Computing research group. He received his B.Sc. degree in computer science from the University of Technology Malaysia, Malaysia, in 1996, M.Sc. degree in computer communications and Networks, from Leeds Metropolitan University, United Kingdom, in 1998, and his Ph.D. degree in computer science from Aston University, United Kingdom, in 2004. His research interest includes mobile and wireless computing, ad hoc and sensor networks, information security and grid computing. He is a member of ACM and International Association of Engineering. He involves in many research projects and is a referee for several scientific journals and conferences.