22-11-2012, 04:08 PM
COMMUNICATION PATTERNS IN VANETs SEMINAR REPORT
[attachment=40689]
ABSTRACT
Vehicular networks are a very promising technology to increase traffic safety and
efficiency, and to enable numerous other applications in the domain of vehicular
communication. Proposed applications for VANETs have very diverse properties and
often require nonstandard communication protocols. Moreover, the dynamics of the
network due to vehicle movement further complicates the design of an appropriate
comprehensive communication system. In this article we collect and categorize envisioned
applications from various sources and classify the unique network characteristics of
vehicular networks. Based on this analysis, we propose five distinct communication
patterns that form the basis of almost all VANET applications. Both the analysis and the
communication patterns shall deepen the understanding of VANETs and simplify further
development of VANET communication systems.
Research on vehicular ad hoc networks has focused mainly on efficient routing
protocol design under conditions where there are relatively large numbers of closely
spaced vehicles. These routing protocols are designed principally for urban areas with
high node density and fully connected networks and are not suitable for packet delivery in
a sparse, partially connected VANET. In this article, we examine the challenges of
VANETs in sparse network conditions, review alternatives including epidemic routing,
and propose a border node-based routing protocol for partially connected VANETs. The
BBR protocol can tolerate network partition due to low node density and high node
mobility. The performance of epidemic routing and BBR are evaluated with a geographic
and traffic information-based mobility model that captures typical highway conditions.
The simulation results show that under rural network conditions, a limited flooding
protocol such as BBR performs well and offers the advantage of not relying on a location
service required by other protocols pro- posed for VANETs.
INTRODUCTION
Up to now, a number of research projects have been carried out to investigate the
vision of communicating vehicles. In projects like FleetNet researchers designed
protocols, algorithms, and systems to test the general feasibility of wireless
communication between vehicles. As the first-generation projects in both the United
States and Europe finished their work some time ago, vehicular networks are now being
both extended and consolidated at the same time. In a number of more recent research
projects such as Network-on Wheels have started the process of standardizing systems
and protocols. For instance, as a basis for wireless communication, the U.S. FCC has
allocated 75 MHz of bandwidth for dedicated short-range communication (DSRC) [1],
which is now used by the emerging IEEE 802.11p standard.
Starting with the idea of making driving safer by inter vehicle communication, the
concept of vehicular networks or vehicular ad hoc networks (VANETs) has been
extended to a large collection of various applications that can profit from wireless
communication between vehicles. Nowadays, vehicles are not only envisioned to
communicate between each other, but also to get information from and send data to
infrastructural units. These stationary parts of the vehicular network range from traffic
lights and dynamic traffic signs to access points at home, gas stations, and elsewhere. In
addition, although active safety applications still represent the central idea, traffic
efficiency applications as well as entertainment and business applications have also
been proposed. In summary, the diverging requirements of all these applications make
the design of a comprehensive communication system a very complex topic.
APPLICATIONS
Applications based on vehicular communication range from simple exchange of
vehicle status data to highly complex large-scale traffic management including
infrastructure integration. As a start to analyzing applications, this section gives an
overview of envisioned application categories for vehicular networks. Although exact
operation details are not yet standardized for most applications, and in spite the fact
that such a collection can never be completely finished, the overview delivers basic
mechanisms, components, and constraints involved in the system. This provides an
initial understanding of the properties of VANET communication and leads to a more
detailed analysis of network characteristics in the next section.
The applications presented in Table 1 are compiled from several sources. A large
collection of applications was gathered in a report [2] by the Vehicle Safety
Communications (VSC) project. Concentrating on active safety, Dötzer et al.
categorized a number of applications in [3]. In the FleetNet project a categorization of
applications was presented in [4], which has some similarities to our classification but is
less detailed. A deeper description of virtual warning sign applications is given by
Maihöfer et al. in [5]. Additionally, most publications on VANETs also contain
examples of applications. The chosen classification scheme groups applications by their
purpose, which leads to groups of logically similar applications.
ACTIVE SAFETY
Active safety applications are considered as the typical and most desirable group
of applications for VANETs with direct impact on road safety. The basic intention is to
make driving safer by communication, which can mean that drivers are warned about a
dangerous situation or even that the vehicle can try to avoid an accident or react
appropriately if an accident cannot be avoided anymore.
In Table 1 we categorize active safety applications according to the danger level,
which can be seen as a compilation of criteria elaborated in [3]. Dangerous road features
like curves are static and thus foreseeable; thus, danger is low. Abnormal traffic and
road conditions are still almost static, but have a dynamic notion (i.e., differ from the
expectation of drivers that regularly pass the event location). In these cases danger is
elevated. Danger is high when applications try to prevent collisions (e.g., if a vehicle
brakes heavily in dense traffic). If this does not help anymore (i.e., in case of imminent
danger when a collision cannot be avoided anymore), precrash sensing will prepare the
vehicle in order to minimize the impact of the impending crash (e.g., by closing windows
or raising dampers). Finally, when danger has turned into an incident, it is important
to warn approaching vehicles or call for help.
PUBLIC SERVICE
Vehicular networks are also intended to support the work of public service
such as police or emergency recovery units. Prominent examples of this category are
the support of emergency vehicles by virtual sirens or signal preemption capabilities.
Using these applications, emergency vehicles may be able to reach their destination
much faster than today. In addition, traffic surveillance could be simplified by
applications such as an electronic license plate. However, such an application must
not be abused by any- one, which clearly underlines security requirements and the
need for a discussion of legal aspects of vehicular communication.
NETWORK CHARACTERISTICS
Besides the application requirements another major set of constraints to the
development of applications, respective message dissemination methods and security
mechanisms is given by the network characteristics, which make VANETs a very
distinct category of ad hoc networks. In this section, we categorize these characteristics
and elaborate on consequences for development.
NODE VELOCITY
One of the most important aspects of mobility in VANETs is the potential node
velocity. Nodes either denote vehicles or roadside units (RSUs) in this case. Node
velocity may range from zero for stationary RSUs or when vehicles are stuck in a traffic
jam to over 200 km/h on highways.
In particular, these two extremes each pose a special challenge to the
communication system. In case of very high node velocities, the mutual wireless
communication window is very short due to a relatively small transmission range of
several hundred meters. For example, if two cars are driving in opposite directions at 90
km/h each, and if we assume a theoretical wireless transmission range of 300 m,
communication is only possible for 12 s. At the same time, moderate relative velocities
of vehicles driving in the same direction lead to less topology dynamics among those
vehicles. These node velocity characteristics have implications on all communication
layers. In case of high relative velocity, the transceivers have to cope with physical
phenomena like the Doppler effect. Because the link layer cannot predict when a
connection will be disrupted, link failures will occur frequently. For routing or
multihop message dissemination, the short encounters between vehicles and general
movement lead to a highly unstable topology, rendering topology-based routing
practically use- less. In their review of challenges of inter vehicle communication [6],
Blum et al. show by means of simulations that routes discovered by classic topologybased
routing protocols get invalid even before they are fully established. For
applications, high node velocities have the effect that e.g., context awareness gets
difficult because the immediate context changes very fast.
NODE DENSITY
Apart from speed and movement pattern, node density is the third key property of
vehicular mobility. As already briefly discussed, it is not hard to imagine that the number
of other vehicles in mutual radio range may vary from zero to dozens or even hundreds.
If we assume a traffic jam on a highway with four lanes, one vehicle every 20 m and a
radio range of 300 m, every node theoretically has 120 vehicles in its transmission range.
In case of very low density, immediate message forwarding becomes impossible.
In this case more sophisticated information dissemination is necessary, which can store
and forward selected information when vehicles encounter each other. In this case the
same message may be repeated by the same vehicle multiple times. In high-density
situations the opposite must be achieved. Here, a message should be repeated only by
selected nodes, because otherwise this may lead to an overloaded channel.
NODE HETEROGENEITY
Among the nodes participating in the envisioned applications, we find numerous
different kinds and types. A basic distinction can be made between vehicles and
infrastructural units (RSUs). Vehicles can be further categorized as private vehicles,
authority vehicles, road construction and maintenance vehicles, and so on. Certainly, not
all applications will be installed in all vehicles; for example, only an emergency vehicle
should be able to issue warnings about its approach. The situation is similar for RSUs.
Depending on the capabilities of the units, infrastructural nodes may simply emit data to
the network or have complete ad hoc functionality, and thus may be used for forwarding
like other vehicles. Moreover, infrastructural nodes may provide access to background
networks (e.g., to inform a traffic operation center about road conditions). In contrast to
vehicles, RSUs have widely different capabilities.
[attachment=40689]
ABSTRACT
Vehicular networks are a very promising technology to increase traffic safety and
efficiency, and to enable numerous other applications in the domain of vehicular
communication. Proposed applications for VANETs have very diverse properties and
often require nonstandard communication protocols. Moreover, the dynamics of the
network due to vehicle movement further complicates the design of an appropriate
comprehensive communication system. In this article we collect and categorize envisioned
applications from various sources and classify the unique network characteristics of
vehicular networks. Based on this analysis, we propose five distinct communication
patterns that form the basis of almost all VANET applications. Both the analysis and the
communication patterns shall deepen the understanding of VANETs and simplify further
development of VANET communication systems.
Research on vehicular ad hoc networks has focused mainly on efficient routing
protocol design under conditions where there are relatively large numbers of closely
spaced vehicles. These routing protocols are designed principally for urban areas with
high node density and fully connected networks and are not suitable for packet delivery in
a sparse, partially connected VANET. In this article, we examine the challenges of
VANETs in sparse network conditions, review alternatives including epidemic routing,
and propose a border node-based routing protocol for partially connected VANETs. The
BBR protocol can tolerate network partition due to low node density and high node
mobility. The performance of epidemic routing and BBR are evaluated with a geographic
and traffic information-based mobility model that captures typical highway conditions.
The simulation results show that under rural network conditions, a limited flooding
protocol such as BBR performs well and offers the advantage of not relying on a location
service required by other protocols pro- posed for VANETs.
INTRODUCTION
Up to now, a number of research projects have been carried out to investigate the
vision of communicating vehicles. In projects like FleetNet researchers designed
protocols, algorithms, and systems to test the general feasibility of wireless
communication between vehicles. As the first-generation projects in both the United
States and Europe finished their work some time ago, vehicular networks are now being
both extended and consolidated at the same time. In a number of more recent research
projects such as Network-on Wheels have started the process of standardizing systems
and protocols. For instance, as a basis for wireless communication, the U.S. FCC has
allocated 75 MHz of bandwidth for dedicated short-range communication (DSRC) [1],
which is now used by the emerging IEEE 802.11p standard.
Starting with the idea of making driving safer by inter vehicle communication, the
concept of vehicular networks or vehicular ad hoc networks (VANETs) has been
extended to a large collection of various applications that can profit from wireless
communication between vehicles. Nowadays, vehicles are not only envisioned to
communicate between each other, but also to get information from and send data to
infrastructural units. These stationary parts of the vehicular network range from traffic
lights and dynamic traffic signs to access points at home, gas stations, and elsewhere. In
addition, although active safety applications still represent the central idea, traffic
efficiency applications as well as entertainment and business applications have also
been proposed. In summary, the diverging requirements of all these applications make
the design of a comprehensive communication system a very complex topic.
APPLICATIONS
Applications based on vehicular communication range from simple exchange of
vehicle status data to highly complex large-scale traffic management including
infrastructure integration. As a start to analyzing applications, this section gives an
overview of envisioned application categories for vehicular networks. Although exact
operation details are not yet standardized for most applications, and in spite the fact
that such a collection can never be completely finished, the overview delivers basic
mechanisms, components, and constraints involved in the system. This provides an
initial understanding of the properties of VANET communication and leads to a more
detailed analysis of network characteristics in the next section.
The applications presented in Table 1 are compiled from several sources. A large
collection of applications was gathered in a report [2] by the Vehicle Safety
Communications (VSC) project. Concentrating on active safety, Dötzer et al.
categorized a number of applications in [3]. In the FleetNet project a categorization of
applications was presented in [4], which has some similarities to our classification but is
less detailed. A deeper description of virtual warning sign applications is given by
Maihöfer et al. in [5]. Additionally, most publications on VANETs also contain
examples of applications. The chosen classification scheme groups applications by their
purpose, which leads to groups of logically similar applications.
ACTIVE SAFETY
Active safety applications are considered as the typical and most desirable group
of applications for VANETs with direct impact on road safety. The basic intention is to
make driving safer by communication, which can mean that drivers are warned about a
dangerous situation or even that the vehicle can try to avoid an accident or react
appropriately if an accident cannot be avoided anymore.
In Table 1 we categorize active safety applications according to the danger level,
which can be seen as a compilation of criteria elaborated in [3]. Dangerous road features
like curves are static and thus foreseeable; thus, danger is low. Abnormal traffic and
road conditions are still almost static, but have a dynamic notion (i.e., differ from the
expectation of drivers that regularly pass the event location). In these cases danger is
elevated. Danger is high when applications try to prevent collisions (e.g., if a vehicle
brakes heavily in dense traffic). If this does not help anymore (i.e., in case of imminent
danger when a collision cannot be avoided anymore), precrash sensing will prepare the
vehicle in order to minimize the impact of the impending crash (e.g., by closing windows
or raising dampers). Finally, when danger has turned into an incident, it is important
to warn approaching vehicles or call for help.
PUBLIC SERVICE
Vehicular networks are also intended to support the work of public service
such as police or emergency recovery units. Prominent examples of this category are
the support of emergency vehicles by virtual sirens or signal preemption capabilities.
Using these applications, emergency vehicles may be able to reach their destination
much faster than today. In addition, traffic surveillance could be simplified by
applications such as an electronic license plate. However, such an application must
not be abused by any- one, which clearly underlines security requirements and the
need for a discussion of legal aspects of vehicular communication.
NETWORK CHARACTERISTICS
Besides the application requirements another major set of constraints to the
development of applications, respective message dissemination methods and security
mechanisms is given by the network characteristics, which make VANETs a very
distinct category of ad hoc networks. In this section, we categorize these characteristics
and elaborate on consequences for development.
NODE VELOCITY
One of the most important aspects of mobility in VANETs is the potential node
velocity. Nodes either denote vehicles or roadside units (RSUs) in this case. Node
velocity may range from zero for stationary RSUs or when vehicles are stuck in a traffic
jam to over 200 km/h on highways.
In particular, these two extremes each pose a special challenge to the
communication system. In case of very high node velocities, the mutual wireless
communication window is very short due to a relatively small transmission range of
several hundred meters. For example, if two cars are driving in opposite directions at 90
km/h each, and if we assume a theoretical wireless transmission range of 300 m,
communication is only possible for 12 s. At the same time, moderate relative velocities
of vehicles driving in the same direction lead to less topology dynamics among those
vehicles. These node velocity characteristics have implications on all communication
layers. In case of high relative velocity, the transceivers have to cope with physical
phenomena like the Doppler effect. Because the link layer cannot predict when a
connection will be disrupted, link failures will occur frequently. For routing or
multihop message dissemination, the short encounters between vehicles and general
movement lead to a highly unstable topology, rendering topology-based routing
practically use- less. In their review of challenges of inter vehicle communication [6],
Blum et al. show by means of simulations that routes discovered by classic topologybased
routing protocols get invalid even before they are fully established. For
applications, high node velocities have the effect that e.g., context awareness gets
difficult because the immediate context changes very fast.
NODE DENSITY
Apart from speed and movement pattern, node density is the third key property of
vehicular mobility. As already briefly discussed, it is not hard to imagine that the number
of other vehicles in mutual radio range may vary from zero to dozens or even hundreds.
If we assume a traffic jam on a highway with four lanes, one vehicle every 20 m and a
radio range of 300 m, every node theoretically has 120 vehicles in its transmission range.
In case of very low density, immediate message forwarding becomes impossible.
In this case more sophisticated information dissemination is necessary, which can store
and forward selected information when vehicles encounter each other. In this case the
same message may be repeated by the same vehicle multiple times. In high-density
situations the opposite must be achieved. Here, a message should be repeated only by
selected nodes, because otherwise this may lead to an overloaded channel.
NODE HETEROGENEITY
Among the nodes participating in the envisioned applications, we find numerous
different kinds and types. A basic distinction can be made between vehicles and
infrastructural units (RSUs). Vehicles can be further categorized as private vehicles,
authority vehicles, road construction and maintenance vehicles, and so on. Certainly, not
all applications will be installed in all vehicles; for example, only an emergency vehicle
should be able to issue warnings about its approach. The situation is similar for RSUs.
Depending on the capabilities of the units, infrastructural nodes may simply emit data to
the network or have complete ad hoc functionality, and thus may be used for forwarding
like other vehicles. Moreover, infrastructural nodes may provide access to background
networks (e.g., to inform a traffic operation center about road conditions). In contrast to
vehicles, RSUs have widely different capabilities.