A survey on position-based routing for vehicular ad hoc networks
[8]. There have been many routing protocols proposed for
VANETs in the past ten years [3,9–26]. Routing protocols
dedicated to VANETs are roughly divided into five cate-
gories: topology-based routing, position-based (geography-
based) routing, traffic-aware routing, cluster-based routing,
and broadcast-based (multicast, geocast) routing.
Topology-based routing forwards packets using exist-
ing communication links. The classical routing protocols
are DSDV [27], WRP, AODV [28], DSR [29] and TORA.
In addition, Namboodiri V presented the PRAODV and
PRAODV-M routing protocols in 2004, which depend on
the accuracy of the predicted algorithm [30]. Naumov V
proposed the ADOV + PGB r outing protocol in 2006 [31]
and Celimuge Wu et al. presented a flexible, portable and
practical routing protocol called PFQ-AODV, which is based
on AODV in 2013. The protocol uses fuzzy logic to eval-
uate whether a wireless link is acceptable by considering
multiple metrics. Based on an evaluation of each wireless
link, the protocol uses a Q-learning algorithm to learn and
find the best route [24]. Position-based (geography-based)
routing protocols or position-based hybrid routing protocols
have been proposed by several papers [13,22,23,25,32,33],
since vehicle position, speed and road information (geog-
raphy, road condition) can be readily obtained or predicted.
Traffic-aware protocols take the traffic information (real-time
or non-real time) into account in their forwarding strategy.
Spatial and traffic aware routing (STAR) [34] is a hybrid
routing protocol, which considers traffic data based on the
position of vehicle, exploiting both street layout information
achieved from geographic information systems and informa-
tion about the spatial distribution of vehicles along the street
and vehicular traffic. Improved greedy traffic aware rout-
ing (GyTAR)[35] considers the direction of vehicle travel,
vehicular density, road conditions and traffic environment
changes. It is a road-based geographic routing protocol
which can find a robust optimal path in an urban environ-
ment. Clustering for Open IVC networks (COIN) [36] and
cluster-based location routing (CBLR) [37] are two typical
cluster-based routing protocols. The urban multi-hop broad-
cast protocol (UMB) [38], multi-hop vehicular broadcast
protocol (MHVB) [39,40], to last one protocol (TLO)[41]
and context-adaptive information diffusion protocol (CAID)
[42] are broadcast-based routing protocols that use flood-
ing to broadcast packets. Some improved flooding mecha-
nisms are addressed, such as probabilistic-based flooding,
counting-based flooding, distance-based flooding, location-
based flooding, neighbor-based flooding and clustering-
based flooding.
After analyzing relevant literature carefully, we find that,
except for some traditional topology-based routing protocols,
the best routing protocols are related to position (digital map
or road layout). We can further sub-classify these routing
protocols into position-based simple routing, position-based
hybrid routing (such as those which incorporate traffic flow
information) and non-position-based routing.
Our main contribution in this paper is to present a study
of strengths and limitations of some well-known position-
based routing protocols. Our goal is to survey the state of
the art of position-based simple and hybrid routing protocols
from the perspectives of scenario, characteristics and prereq-
uisites. We discuss pros and cons for routing protocols from
each perspective, and also make a qualitative comparison of
protocols. To the best of our knowledge, this is the first sur-
vey to assess the state of the art of routing protocols from
three perspectives. In addition, we also have discussed some
positioning algorithms, since vehicular position is a basic
element for position-based routing protocols.
The rest of the paper is organized as follows. Section 2 will
introduce the architecture of Internet of Vehicle and protocol
stacks of VANETs. Since accurate position information is an
important consideration in vehicular applications, we outline
some positioning schemes to help GNSS to acquire vehicle
location. Then we will explain the position-based routing
protocols for V2V and V2I in Sect. 3. In Sect. 4, we will
discuss and compare routing protocols for vehicular com-
munication. Section 5 reviews open issues and challenges in
vehicular communication. Finally, we conclude the paper in
Sect. 6.
2 Backgrounds and motivations
With the advent of big data technology, large volumes of data
generated from vehicular information, traffic data and road
status can now be utilized more effectively. Google’s prod-
ucts, including basic file systems (Google file system [GFS]
or Colossus), management databases (BigTable), and pro-
gramming models (MapReduce), can assist with managing
applications based on large amounts of data, such as traffic
flow forecasting. Meanwhile, the development of LTE com-
munication technologies brings faster network access than
before. The data storage mode has changed in some indus-
tries, such as wireless body area network (WBAN) [43] and
mobile cloud computing [44,45]. Similarly, the architecture
of Internet of Vehicle has changed with the introduction of
big data and the LTE technique in recent years [46,47]. Enter-
tainment and information services can now be delivered more
effectively using LTE cellular networks. In this section, we
outline the background and the latest developments in ITS,
including system architecture and protocol stacks. Position-
based routing applies in more and more applications, as vehi-
cles can obtain their location from onboard GPS receivers or
a global digital map. Accurate position information plays
an important role in location based services. However, GPS
will fail in tunnels, underground, or anywhere else where the
satellite signal is blocked. The accuracy of positioning may
123