**2.2.1 Classification of applications for VANET**

Vehicular software applications may be categorized into four groups (Popescu-Zeletin et al., 2010):


Emerging Technologies for Urban Traffic Management 65

 *Geocast*. It is a multicast routing service that delivers messages to nodes located within a given geographical region. These routing protocols generally define a forwarding zone that limits flooding of messages. Using this routing scheme it is possible to, for instance, report an accident to vehicles located within a given region or alert a driver when

 *Forwarding*. The purpose of this routing scheme is to transport messages between two nodes via multiple hops. This mechanism is useful when the requested information is only of interest to a few nodes. For example, a node may request information to a nearby car parking about free car parking spaces and fees. When a node is requesting information, a unicast message is sent. To forward the message to its destination a route is reactively constructed, for example, by looking at local routing tables or by asking

 *Clustering*. The cluster-based approach consists on grouping nodes located within a given region (e.g. nodes with direct link with each other). For each cluster, a cluster head node is selected which is responsible for managing inter and intra-cluster communication. The cluster-based structure functions as a virtual network infrastructure whose scalability favors routing and media access protocols although an overhead cost is paid when forming clusters in highly mobile network environments

 *Beaconing*. This routing mechanism is suitable for applications that require sharing information with other vehicles periodically (e.g. exchange of local traffic information). In this routing scheme a node announces information periodically. Receiving nodes, do not re-broadcast the received message immediately, instead, they integrate and store received information on its local information cache. On the next beacon, a message is constructed using both local and the incoming information and broadcasted to

 *Position-based*. For this routing scheme to work, information on the location of each node is fundamental. To decide how to route messages, nodes utilize geographical location information obtained from sources such as street maps, traffic models and on-board navigational systems. Routing decisions at each node are done taking into consideration the position of the destination node and each node's location information. As routing tables are not required, no overhead is incurred on maintaining and establishing routes. *Delay-tolerant*. There exist scenarios where the density of vehicles is really low and consequently establishing end-to-end routes is not possible. For example at nights, traffic in cities can be really low and available vehicles may not be close enough to receive and forward messages. Also, in rural areas vehicles density may be low. In sparse networks like those, a delay-tolerant protocol can be utilized. This routing mechanism is based on the concept of carry and forward, where a node carries messages and these are only forwarded when another node moves into its vicinity,

increases.

driving on a motorway in the wrong-way.

nearby nodes whether they know about the destination node.

and network delays may occur on large networks.

neighboring nodes.

otherwise, they are simply carried.

broadcast messages only once, and a time to live parameter can be utilized to limit messages area of distribution. Using this routing scheme, delivery of messages to all nodes is guaranteed, however, a large amount of bandwidth is consumed, this is why this routing scheme only performs well when a small number of nodes is participating within the VANET and its performance drops quickly when the size of the network

Some other requirements must be considered for all the above applications, for instance whether they need sensors, human-machine interfaces, GPS, or maps in order to provide extra functional capabilities. Table 1 shows the requirements for types of applications for vehicular networks (CAMP Vehicle Safety Communications Consortium, 2005).


Table 1. Requirements for different types of vehicular networks applications

### **2.3 Data dissemination schemes**

Given the complexities of VANET in terms of their dynamic topology, mobility models, hard delay constrains, and the different system architectures utilized, transporting information from one vehicle to another or to all vehicles within a given region or area is a highly challenging task. A lot of research has been carried out to develop protocols and mechanisms that can provide network services (e.g. routing) to applications in a VANET environment. Next, a classification of the different protocols for transporting information that have been proposed is presented and briefly analyzed (Li & Wang, 2007; Maihofer, 2004; Nundloll et al., 2009; Zeadally et al., 2010; Mauve, 2010):

 *Broadcast*. This routing method is generally utilized for disseminating information such as traffic, weather, emergency, road conditions, among others, to other vehicles. This communication scheme sends packets to all nodes in the network using flooding. When messages need to be disseminated beyond the radio transmission range, a multi-hop mechanism is then utilized. Thus, in a native broadcast implementation, all receiving nodes simply rebroadcast the received messages. To limit message duplication, nodes

Some other requirements must be considered for all the above applications, for instance whether they need sensors, human-machine interfaces, GPS, or maps in order to provide extra functional capabilities. Table 1 shows the requirements for types of applications for

**latency Data transmitted Range** 

250 m

200 m

200 m

50 m

150 m

300 m

150 m

direction, road geometry.

slope, speed limit, surface.

velocity, acceleration, yaw rate.

velocity, acceleration, yaw rate.

acceleration, turn signal status.

acceleration.

direction

vehicular networks (CAMP Vehicle Safety Communications Consortium, 2005).

**Type Rate Maximum**

Traffic signal violation V2I 10 Hz 100 ms Signal phase, timing, position,

Curve speed warning V2I 1 Hz 1000 ms Curve location, curvature,

Emergency brake lights V2V 10 Hz 100 ms Position, heading, velocity,

Pre-crash sensing V2V 50 Hz 20 ms Vehicle type, position, heading,

Forward collision V2V 10 Hz 100 ms Vehicle type, position, heading,

Left turn assist V2I or V2V 10 Hz 100 ms Signal phase, timing, position,

Stop sign assist V2I or V2V 10 Hz 100 ms Position, velocity, heading. 300 m Electronic Toll Collection V2I 10 Hz 50 ms. 15 m Internet Access V2I 10 Hz 500 ms 300 m Automatic parking V2I 10 Hz 500 ms Position, distance 300 m Roadside service finder V2I or V2V 10 Hz 500 ms Position, velocity 300 m

Given the complexities of VANET in terms of their dynamic topology, mobility models, hard delay constrains, and the different system architectures utilized, transporting information from one vehicle to another or to all vehicles within a given region or area is a highly challenging task. A lot of research has been carried out to develop protocols and mechanisms that can provide network services (e.g. routing) to applications in a VANET environment. Next, a classification of the different protocols for transporting information that have been proposed is presented and briefly analyzed (Li & Wang, 2007; Maihofer,

 *Broadcast*. This routing method is generally utilized for disseminating information such as traffic, weather, emergency, road conditions, among others, to other vehicles. This communication scheme sends packets to all nodes in the network using flooding. When messages need to be disseminated beyond the radio transmission range, a multi-hop mechanism is then utilized. Thus, in a native broadcast implementation, all receiving nodes simply rebroadcast the received messages. To limit message duplication, nodes

Lane-change warning V2V 10 Hz 100 ms Position, heading, velocity,

Table 1. Requirements for different types of vehicular networks applications

2004; Nundloll et al., 2009; Zeadally et al., 2010; Mauve, 2010):

**Application Communication** 

**2.3 Data dissemination schemes** 

broadcast messages only once, and a time to live parameter can be utilized to limit messages area of distribution. Using this routing scheme, delivery of messages to all nodes is guaranteed, however, a large amount of bandwidth is consumed, this is why this routing scheme only performs well when a small number of nodes is participating within the VANET and its performance drops quickly when the size of the network increases.


Emerging Technologies for Urban Traffic Management 67

in vehicle-to-vehicle and vehicle-to-infrastructure communication environments (Papadimitratos et al., 2008). However, VANET applications will bring a series of challenges on the security area that help to solve several issues such as integrity, privacy and the non-

Integrity is related to honesty and verification of the information. For applications trustworthiness of data is more useful that trustworthiness of nodes communicating data. Data trust and verification ensures that, on the one hand, the exchanged information can be trusted, and on the other hand, the receiver nodes can verify the integrity of the received information in order to protect the vehicular network from attacks and impersonation security. In (Leinmuller et al., 2007) authors classify the trust and verification concepts into proactive security and reactive security. According to Leinmuller the former has been researched extensively and consists of digitally signed messages, a proprietary system design, and Tamper resistant hardware (Caladriello et al., 2007; Hu & Labertearx, 2006; Garfinkel et al., 2003). The latter consists of signature-based, anomaly-based and contextbased approaches. Their main characteristic is that they correlate the received information with information that is either already available into the system from observations on normal system operations or that is introduced additionally (Brutch & Ko, 2003; Zhang et

As mentioned before, security in vehicular networks must be designed to prevent potential attacks caused by drivers reacting dangerously as a result of receiving erroneous messages. Non-repudiation is related to define mechanisms, to prevent an entity from denying previous commitments or actions. Vehicular applications require a strong mutual authentication with non-repudiation because all safety-related messages may contain lifesaying information. For instance, the diffusion of fake safety messages by an attacker could

Privacy is related to protect user information, while at the same time authorities have to be able to reveal the identity of message senders in case of an eventuality (Raya et al., 2006). Therefore it is critical to develop mechanisms to preserve privacy in vehicular networks. Some of the proposed techniques to provide privacy are: anonymous certificates, group signatures and pseudonym certificates. The anonymous certificates technique is based on the usage of a list of anonymous certificates for message authentication, which is stored in a central repository (such as a transportation regulation center). The second technique is in charge of providing anonymity to a group of members. Any node of the group has the capacity of verifying whether a group member sent a certain message, however it is not necessary to know the real identify of the sender node. Finally, pseudonymous authentication is a technique widely accepted in vehicular networks. Its main use is

In (Rivas et al., 2011) authors analyse other important issue in the security area for vehicular networks, the detection and eviction of misbehaving and faulty nodes. Due to the attacker's ability or just to the devices aging process at some point in the time there will be

repudiation of messages and authentication.

**2.4.1 Integrity** 

al., 2003).

**2.4.2 Privacy and non-repudiation** 

anonymous authentication.

produce potentially dangerous situations on the road.

 *Ad-hoc (address-based/topology-based).* This category groups routing protocols initially designed to operate in *Mobile Ad-hoc Networks* (MANET) environments. VANET attempts to test these routing protocols in such new environments have been carried out. However, requirements on these address-based and topology-based mechanisms such as unique address identification among others make these protocols less suitable for VANETs.
