Vehicular Social Networks

Anna Maria Vegni, Valeria Loscrì, Athanasios V. Vasilakos, Anna Maria Vegni, Valeria Loscrì, Athanasios V. Vasilakos

• Éditeur: CRC Press
• Année: 2017
• ISBN:
  9781315351452
• Édition: 1
• Sujet: Technologie et ingénierie
• Sous-sujet: Communications mobiles et sans fil

The book provides a comprehensive guide to vehicular social networks. The book focuses on a new class of mobile ad hoc networks that exploits social aspects applied to vehicular environments. Selected topics are related to social networking techniques, social-based routing techniques applied to vehicular networks, data dissemination in VSNs, architectures for VSNs, and novel trends and challenges in VSNs. It provides significant technical and practical insights in different aspects from a basic background on social networking, the inter-related technologies and applications to vehicular ad-hoc networks, the technical challenges, implementation and future trends.

INTRODUCTION TO VSNs

I

Chapter 1

Social Clustering of Vehicles

Leandros A. Maglaras and Helge Janicke

School of Computer Science and Informatics, De Montfort University, Leicester, United Kingdom

Pavlos Basaras and Dimitrios Katsaros

Department of Computer and Communication Engineering, University of Thessaly, Volos, Greece

Contents

1.1 Introduction
1.1.1 Motivation
1.2 Clustering of Vehicles and Vehicular Social Networks
1.2.1 Clustering Vehicular Networks
1.2.2 Social Vehicles
1.3 Social Clustering of Vehicles
1.3.1 Definition of the System
1.3.2 Subnetworks and the Roles of RSUs
1.3.3 Sociological Patterns
1.3.4 Sociological Pattern Clustering Method
1.4 Simulation and Performance Evaluation
1.4.1 Cluster Stability versus Communication Range
1.4.2 Cluster Stability versus Speed
1.5 Social Vehicle Clustering under Attack Scenarios: Countermeasures
1.6 Conclusions
1.7 Glossary
References

1.1 Introduction

The vision of the vehicular ad hoc network (VANET) [1] is now very close to becoming reality and entering our everyday lives. Vehicles will be equipped with onboard units (OBUs), that is, a communication device that allows short-ranged wireless transmissions and hence facilitates vehicle-to-vehicle (V2V) communication. Roadside units (RSUs) will further aid the vehicular environment by serving as gateways to the Internet or other networks [2], and support numerous applications.

Benefits born of the VANET are numerous, and find plausible applications in the generic environment of information exchange. Examples include messages regarding traffic or weather conditions, hazard areas or road conditions, that is, safety applications, or infotainment services [3], allowing users on board vehicles to receive information relevant to services available in certain areas, for example, live video streaming and file sharing. Nonetheless, benefits do not come free of hazard. For instance, flooding the network with all kinds of messages will most likely exhaust the wireless network’s resources, for example, cause severe contention and collisions, seize the very existence of the VANET, and thus abolish its benefits. Such challenges are addressed throughout the literature of vehicular networks in various ways [4]. Based on the fact that cars will be a major element of the expanding Internet of Things (IoT), with one in five vehicles having some sort of wireless network connection by 2020, the development of novel clustering mechanisms that exploit the augmented information gathered throughout the system is emerging.

The Social Internet of Things (SIoT) concept [5] is a network of intelligent objects that have social interactions. The Social Internet of Vehicles (SIoV) [6] is an example of an SIoT where the objects are smart vehicles. The authors in [7] describe a vehicular social network (VSN) as social interactions among cars that communicate autonomously to look for services (automaker patches or updates) and exchange information relevant to traffic. When moving on the social aspect of vehicular communications, new parameters like frequency of interactions between entities, historical data of driver behavior, and driver habits must be taken into account when creating groups of entities involved (e.g., cars, passengers, drivers, road and users).

This chapter discusses social aspects of vehicles and describes novel social clustering mechanisms. Although there is a lot of research dedicated to VSNs this work is mostly focused on the data sharing applications for infotainment. This chapter presents a novel vehicle clustering protocol that exploits the macroscopic social behavior of vehicles in order to create stable clusters for urban and high scenarios.

In particular, in this chapter we pursue two main goals: to analyze the different types of clustering methods and to present novel social clustering mechanisms for vehicles. Although many clustering methods for vehicular networks exist, they suffer from one or more of the following shortcomings: they are not generic in order to be applied in different environments, they are unpractical and difficult to be used in real-time situations, they do not exploit the road network topology, and they do not use the historic data in order to create stable clusters.

1.1.1 Motivation

Node clustering is proven to be an effective method to provide better data aggregation and scalability for wireless ad hoc networks. Clustering of nodes is also used, in order to cope with the interference caused by flooding of messages, since when the network is clustered, mainly the cluster head (CH) participates in the routing algorithm, which greatly reduces the number of necessary broadcasts. In vehicular networks, the social behavior of vehicles, that is, their tendency to share the same or similar routes, can be used in order to stabilize the clusters and increase their lifetime. The social characteristics of the drivers can help in the creation of more stable and robust clustering formations. The fact that each driver has some preferred routes that he or she tends to follow, depending on the time period, can be the basis for creating social profiles for each vehicle or driver based on historic trajectories of vehicles gathered by RSUs located throughout the road network. Combined communication capabilities along with social behavior of the vehicles can facilitate safety [8], eco-routing [9, 10], and infotainment [11], as well as dynamic charging of electric vehicle [12, 13] applications.

In this chapter, we present the steps that need to be taken in order to create social clusters of vehicles. The first step is the division of the road network into subnetworks in order to investigate each area in isolation. The partition is based on the connectivity among the road segments. The second step is the collection of the trajectories that the vehicles followed in these areas in order to create the social profiles (third step) for each of them, by using semi-Markov models. These profiles represent the behavior of each vehicle or driver in each area for each time period. The time periods are predawn (up until 8:00 a.m.), morning rush hour (8:00 a.m. to 10:00 a.m.), late morning (10:00 a.m. to noon), early afternoon (noon to 4:00 p.m.), evening rush hour (4:00 p.m. to 7:00 p.m.), and night time (after 7:00 p.m.). The last step is the exploitation of this enriched situation awareness, the social profiles of vehicles, in order to perform social clustering. A novel social clustering method for vehicles that follows these steps is presented, and its performance for different simulation settings, for example, communication range and velocity, is evaluated. The robustness of the method when some malicious vehicles tend to send bogus information due to infection is also investigated, and several defense mechanisms are proposed. The proposed social clustering of vehicles comes with low communication overhead and increased cluster stability.

1.2 Clustering of Vehicles and Vehicular Social Networks

Given the introduction of clustering in vehicular networks and its importance, this section briefly describes clustering techniques for vehicles [14] and the social aspect of vehicular communications [15, 16]. Note that we only discuss algorithms designed specifically for VANETS or techniques later improved to fit in the vehicular domain.

1.2.1 Clustering Vehicular Networks

In [17], the authors proposed a clustering technique based on a new aggregated local mobility (ALM) algorithm, with the objective to increase the stability of the devised clusters. The ratio of the received signal strength between two successive hello message exchanges, is used as a means to measure the relative variance in the mobility of the receiver’s vicinity. When clustering (re)configuration stimulates, the vehicle node with the lower ALM, that is, the most stable node, is elected as CH. Finally, to avoid frequent (or unnecessary) cluster reorganization, additional time (and message exchanges) is needed in order to initiate reclustering. By minimizing the relative mobility and distance between CHs and their corresponding CMs, the algorithm was found to create stable clusters with a low average CH change rate.

In [18], enhanced spring clustering (ES-Cl) was proposed for highway environments. ES-Cl, apart from favoring vehicles with relative constant velocity or vehicles with predefined routes for CH selection, further incorporates a vehicle’s physical dimensions, for example, its height. The intuition lies in the fact that communication can be much wider (less obstructed) when, for example, a tall vehicle acts as the transmitter. The results indicate that significant benefits can be obtained when the height of vehicles is taken into consideration for the CH selection process.

The authors in [19] utilize the affinity propagation algorithm [20] and propose a mobility-based clustering technique for VANETs. Their objective lies with low CH change rate and long duration of both CH and cluster member (CM) vehicle nodes. The algorithm is carried out with the use of paired exchanged messages, tha...