A survey of geographic routing protocols for Vehicular Ad Hoc Networks (VANETs)
A survey of geographic routing protocols for Vehicular Ad Hoc Networks (VANETs) Jesus Gabriel Balderas Lopez New Mexico State University. Fall 2010 CS 479/579Problem definition: Routing has been a challenge in Vehicular Ad Hoc Networks(VANETs) because of the features of this kind of networks. We knowthat in VANETs exist a frequent change in the topology of the networkand this is due to the fast mobility of the nodes/vehicles. Therealso exist the problem of connectivity between the nodes for the samereason. VANETs have many applications; we can use them to improve roadtraffic safety and efficiency with real time information about thestatus of the road, if there is an accident we can know it and lookfor a new route to go to your destination; we can also use VANETs formedia sharing between two vehicles (like the example given in classwhere the driver of one of the vehicles is able to know the song thatis playing is the neighbor vehicle and sends a request to downloadthe song and play it on his vehicle). There are many applicationsof VANETs and in order to accomplish these things we need some newrouting protocols that consider the challenges mentioned above.In this survey we are going to analyze only routing strategies thatuse geographical location information obtained from street maps,traffic models or even more prevalent navigational systems on-boardthe vehicles. The reason of this is that geographic routing has beenidentified as a more promising routing paradigm for VANETs; therefore,most of the routing protocols available for VANETs use some kind of
geographic information. In the following sections, I classify the geographic routingprotocols for VANETs according to the routing type; I describe eachprotocol and describe how the protocol uses the location informationfor routing. After this classification I present my insights of theproblem, and finally I present some complexities that I ran into whiletrying to solve this problem.Classification: The easiest way to classify the geographic routing protocolsis by type of routing (Unicast, Broadcast or Geocast). Other wayto classify them is by the use that the protocol gives to theposition information (Packet forwarding, Route Selection, Clusterformation, Formation of cells, Classify Forwarding Group, or RouteRequest Forwarding). For this survey I will use the type of routingclassification and in each protocol I’ll talk about how the protocoluses the geographic information.Unicast: GPSR (Greedy Perimeter Stateless Routing)  is probablythe best known geographic routing protocol for VANETs. GPSR usesthe positions of routers and a packet’s destination to make packetforwarding decisions. The position of a packet’s destination andpositions of the candidate next hops are sufficient to make correctforwarding decisions, without any other topological information. Inthis protocol the authors assume that all wireless routers know theirown position, either from a GPS device, if outdoors, or through othermeans. They also assume bidirectional radio reachability. Finally,they assume that packet sources can determine the locations of packetdestinations, to mark packets they originate with their destination’slocation. GPSR consists of two methods for forwarding packets: greedyforwarding, which is used wherever possible, and perimeter forwarding,which is used in the regions greedy forwarding cannot be. ● Greedy forwarding A forwarding node can make a locally optimal, greedy choice in choosing a packet’s next hop. The locally optimal choice of next hop is the neighbor graphically closest to the packet’s destination. This scheme is followed successively
until the destination is reached. The figure above represents an example of greedy next hop choice. Here, the x receives a packet destined for D. x’s radio range is denoted by the dotted circle about x, and the arc with radius is equal to the distance between y and D is shown as the dashed arc about D. x forwards the packet to y, as the distance between y and D is less than that between D and any of x’s other neighbors. This greedy forwarding process repeats until the packet reaches D. Periodically, each node transmits a beacon to the broadcast MAC address, containing only its own identifier and position. This process provides all nodes with their neighbors’ positions. There are topologies in which the only route to a destination requires a packet move temporarily farther in geometric distance from the destination. An example of such topology is shown in the next figure. In this figure, x is closer to D than its neighbors w and y. Although two paths, (x -> y -> z -> D) and (x -> w -> v -> D), exists to D, x will not choose to forward to w or y using greedy forwarding because x is the local maximum in its proximity to D. Some other mechanism must be used to forward the packets in these situations.● The Right-Hand Rule: Perimeters: This rule states that when
arriving at node x from node y, the next edge traversed is the next one sequentially counterclockwise about x from edge (x, y). The figure to the right illustrates this rule. It is known that the right-hand rule traverses the interior of a closed polygonal region (a face) in clockwise edge order- In this case, the triangle bounded by the edges between nodes x, y, z, in the order (y -> x -> z -> y). They call the sequence of edges traversed by the right-hand rule a perimeter. It is important to recall that all nodes maintain a neighbortable, which stores the addresses and locations of their single-hopradio neighbors. All packet data packets are marked initially at theiroriginators as greedy mode. Packet sources also include the geographiclocation of the destination in packets. When a forwarding node receives a packet in greedy mode, itsearches its neighbor table for the neighbor geographically closest tothe packet destination. If this neighbor is closer to the destination,the node forwards the packet to that neighbor. When no neighbor iscloser, the node marks the packet into perimeter mode. GPSR forwards perimeter mode packets using a simple planar graphtraversal. GPSR works best in a free open space scenario with evenlydistributed nodes. It also suffers from several problems. First, incity scenarios, greedy forwarding is often restricted because directcommunications between nodes may not exist due to obstacles such asbuildings and trees. Second, if apply first planarized graph to buildthe routing topology and then run greedy forwarding or face routing onit, the routing performance will degrade. Third, mobility can alsoinduce routing loops for face routing, and last, sometimes packets mayget forwarded to the wrong direction leading higher delays or even
network partitions. Geographic Source Routing (GSR)  is other position-basedrouting protocol for VANETs. GSR assumes the aid of a street map incity environments. This street map is used to know the city topology.GSR uses something called Reactive Location Service (RLS) to get thedestination position. GSR combines geographic routing and topologicalknowledge from street maps; the sender determines the junctions thathave to be traversed by the packet using the Dijkstra’s shortest pathalgorithm and then forward the packet in a position-based fashionbetween the junctions. This protocol was designed for cityenvironments. GPCR (Greedy Perimeter Coordinator Routing)  is other unicastgeographic routing protocol for VANETs. The main idea of GPCR is totake advantage of the fact that streets and junctions form a naturalplanar graph, without using any global or external information such asa static street map. It consists of two parts: A restricted greedyforwarding procedure and a repair strategy which is based on thetopology of real-world streets and junctions and hence does notrequire a graph planarization algorithm. ● Restricted Greedy Routing: A special form of greedy forwarding is used to forward a data packet towards the destination. Since obstacles block radio signal, data packets should be routed along streets. Junctions are the only places where actual routing decisions are taken. Therefore packets should always be forwarded to a node on a junction rather than being forwarded across a junction. This is illustrated in the figure below where node u would forward the packet beyond the junction to node 1a if regular greedy forwarding is used. By forwarding the packet to node 2a an alternative path to the destination node can be found without getting stuck in a local optimum. They call a node located in the area of a junction a coordinator.
● Repair Strategy: The repair strategy of GPCR avoids using graph planarization by making routing decision on the basis of street and junctions instead of individual nodes and their connectivity. As a consequence the repair strategy of GPCR consist of two parts: ○ On each junction it has to be decided which street the packet should follow next. ○ In between junctions greedy routing to the next junction can be used. If the forwarding node for a packet in repair mode is a coordinator then the node needs to determine which street the packet should follow next. To this end the topology of the city is regarded as a planar graph and the well known right-hand rule is applied. The next image is taken from  This image compares Greedyforwarding (used in GPSR) vs Restricted greedy routing in the area ofjunctions (used in GPCR) in (a), and (b) illustrates the right handrule used in the repair strategy of GPCR.
Anchor-based Street and Traffic Aware Routing (A-STAR)  wasproposed for city environments. A-STAR is similar to GSR; it uses thestreet map to compute the sequence of junctions (anchors) throughwhich a packet must pass to reach its destination. Bur unlike GSR, A-STAR computes the anchor paths with traffic awareness. A-STAR is also different from other protocols because it employsa new local recovery strategy for packets routed to a local minimumthat is more suitable for a city environment than the greedy approachof GSR and the perimeter-mode of GPSR. In the local recovery state,the packet is salvaged by traversing the new anchor path. To preventother packets from traversing through the same void area, the streetat which local minimum occurred is marked as “out of service”temporarily and these streets are not used for anchor computation orre-computation during the “out of service” duration and theyresume “operational” after the time out duration.
Clustering for Open Inter-vehicular communication (IVC) Networks(COIN)  is a cluster-based protocol. In cluster-based routing, avirtual network infrastructure must be created through the clusteringnodes in order to provide scalability. Each cluster can have a clusterhead, which is responsible for intra-and inter-cluster coordination inthe network management functions. Nodes inside a cluster communicatevia direct links. Inter-cluster communication is performed via thecluster-heads. The image below illustrates clustering in VANETs. COINuses the location information for cluster formation. Cluster headelection is based on vehicular dynamics and driver intentions, insteadof ID or relative mobility as in classical clustering methods. COINproduces much more stable structures in VANETs while introducinglittle additional overhead. LORA_CBF  is other location based routing algorithm that usescluster-based flooding for VANETs. Each node can me the cluster-head,gateway or cluster member. If a node is connected to more than onecluster, it is called gateway. The cluster-head maintains informationabout its members and gateways. Packets are forwarded from a source tothe destination by protocol similar to greedy routing. If the locationof the destination is not available, the source will send out thelocation request (LREQ) packets. This phase is similar to the routediscovery phase of AODV, but only the cluster-heads and gateways willdisseminate the LREQ and LREP (Location Reply) messages.Broadcast: The simplest way to implement a broadcast service is flooding inwhich each node re-broadcast messages to all of its neighbors exceptthe one it got this message from. Flooding guarantees the messagewill eventually reach all nodes in the network. Flooding performsrelatively well for a limited small number of nodes and is easy to beimplemented. But when the number of nodes in the network increases,the performance drops quickly. Flooding may have a very significantoverhead and selective forwarding can be used to avoid networkcongestion.
BROADCOMM  is an emergency broadcast protocol based on ahierarchical structure for a highway network. In BROADCOMM, thehighway is divided into virtual cells, which moves as the vehiclemoves. The nodes in the highway are organized into two level ofhierarchy: the first level includes all the nodes in a cell; thesecond level is represented by the cell reflectors, which are a fewnodes usually located closed to the geographical center of the cell.Cell reflectors behaves for a certain time interval as a base stationor cluster head that will handle the emergency messages coming frommembers of the same cell, or close members from neighbors cells. BROADCOMM outperforms similar flooding based routing protocolsin the message broadcasting delay and routing overhead. However, it isvery simple and only works with simple highways networks. UMB (Urban Multi-Hop Broadcast)  is designed to address thebroadcast storm, hidden node, and reliability problems of multi-hopbroadcast in urban areas. This protocol assigns the duty of forwardingand acknowledging broadcast packet to only one vehicle by dividing theroad portion inside the transmission range into segments and choosingthe vehicle in the furthest non-empty segment without apriori topologyinformation. When there is an intersection in the path of the messagedissemination, new directional broadcast are initiated by repeaterslocated at the intersections. The most important goals of UMB, according to the authors, areas follows: 1. Avoiding collisions due to hidden nodes: In order to decrease the effect of hidden nodes, a mechanism similar to RTS/CTS handshake in point-to-point communication is employed by their UMB protocol. They refer to RTS and CTS as Request To Broadcast (RTB) and Crear To Broadcast (CTB), respectively. 2. Using the channel efficiently: Forwarding duty is assigned to only the furthest vehicle in the transmission range without using the network topology information. 3.Making the broadcast communication as reliable as possible: To achieve the reliability goal, an ACK packet is sent by the vehicle which was selected to forward the packet. 4. Disseminating messages in all directions at an intersection: New directional broadcast are initiated by the simple repeaters installed at the Intersection Broadcast mechanism. The next image illustrates the sequence of packets in UMB toavoid collisions due to hidden nodes.
The image below illustrates the intersection broadcast in UMBfor disseminating messages in all directions at an intersection. Vector-based TRAcking DEtection (V-TRADE) and History-enhanced V-TRADE (HV-TRADE)  are GPS based message broadcasting protocols.The basic idea is similar to the unicast routing protocol Zone RoutingProtocol (ZRP). Based on position and movement information, theirmethods classify the neighbors into different forwarding groups. Foreach group only a small subset of vehicles (called border vehicles) isselected to rebroadcast the message. They show significant improvementof bandwidth utilization with slightly loss of reachability, becausethe new protocols pick fewer vehicles to rebroadcast the messages. Butthey still have routing overhead as long as the forwarding nodes areselected in every hop.Geocast: Geocast routing is basically a location-based multicast routing.The objective of a geocast routing is to deliver the packer from a
source node to all the other nodes with a specified geographicalregion (Zone of Relevance, ZOR). Vehicles outside the ZOR are notalerted to avoid unnecessary and hasty reactions. The source node isusually inside the ZOR Most geocast routing methods are based on directed flooding,which tries to limit the message overhead and network congestion ofsimple flooding by defining a forwarding zone and restricting theflooding inside it. Non-flooding approaches (based on unicast routing)are also proposed, but inside the destination region, regionalflooding may still be used even for protocols characterized as non-flooding. Inter-Vehicles Geocast protocol (IVG)  consists in informingall the vehicles of a highway about any danger such as an accident orany other obstacle. In this case, risk areas are determined accordingto the driving direction and the position of the vehicles. The nodewhich receives an alarm message should not rebroadcast it immediatelybut has to wait some time, called defer time, to take a decision aboutrebroadcast. When this defer time expires and if it does not receivethe same alarm message from another node behind it, it deducts thatthere is no relay node behind it. Thus it has to designate itself as arelay and starts broadcasting alarm messages to inform the vehicleswhich might be behind it. The defer time of node (x) receiving amessage from another node (s) is inversely proportional to thedistance separating them that is to favorite the farthest node to waitless time and to rebroadcast faster Cached Geocast  is other geocast protocol. The main idea oftheir cached greedy geocast inside the ZOR is to add a small cache tothe routing layer that holds those packets that a node cannot forwardinstantly due to a local minimum. When a new neighbor comes into areach or known neighbors change their positions, the cached messagecan be possibly forwarded to the newly discovered node. Their distanceaware neighborhood strategy takes frequent neighborhood changes intoaccount. It chooses the closest node to destination which is inside
the range r (smaller than the transmission range) instead of the nodetransmission range in the general greedy routing mode. The improvedneighborhood selection taking frequent neighborhood changes intoaccount significantly decreases network load and decreases end-to-enddelivery delay. Beside of the classical geocast routing, there is a specialgeocast, called Abiding Geocast , where the packets need todelivered to all nodes that are sometime during the geocast lifetime(a certain period of time) inside the geocast destination region.Services like position-based advertising, position-based publish-and-subscribe, and many other location-based services profit from abidinggeocast. For VANETs, abiding geocast allows realization of informationand safety applications like virtual warning signs. Similar to realtraffic of warning signs, they are attached to a certain geographicalposition or area. When a vehicle enters such an area, the virtualwarning sign is displayed for the driver. The authors provided threesolutions: 1. A server is used to store the geocast messages. 2. An elected node inside the geocast region stores the messages. 3. Each node stores all geocast packets destined for its location and keeps the neighbor information.My insights In this work I only present a few position-based routingprotocols for Vehicular Ad hoc Networks, there are so many and more tocome because this is a field of wireless networks that is increasingso fast. We can see in this paper that have different protocolsdepending the environment, if this is for a city of highway, and therouting type. When I started to work on this project I wanted to be able tomention the best protocol for VANETs but this is a hard job. If wewant to implement one of these protocols we need to think about therequirements. If it is a city scenario maybe I would choose a unicastprotocol like A-STAR because it is also consider traffic load in thestreets. Another option is a broadcast protocol like UMB, I think thisis a good option because this protocol tries to avoid collisions dueto hidden nodes, makes broadcast a reliable way of communicationbecause implements ACK packets in the process of communicationsbetween nodes; it also adds the idea of repeater in the intersectionsto rebroadcast the packets and this is useful because you don’tnecessarily need other vehicles in the streets to forward the packet
to the destination. I think there is so much work to do in the area of routing inVANETs because in our day to day life we are not always in the citynor in the highways; it’s necessary to create a new protocol able towork in both city environments and highways but this is a bigchallenge. I never thought to find a protocol like Abiding Geocast that isdesigned for advertising and publishing depending on the region thevehicle is located and this is an amazing application for VANETs, notonly for real time traffic information or emergency information. I am impressed of all these protocols and applications and Ithink there is more to come.Complexities I think that the complexities of solving this problem of routingin VANETs are related in the implementation and experimentation.This is because in order to have better results we need no implementthe protocol in a real-life scenario and not just simulate it withsome kind of traffic simulator; also we could need to implement theprotocol not just in one city or highway, maybe 2 or more is better toobtain results not just in a particular place. I think this part isthe hardest one.References 1. F. Li and Y. Wang, “Routing in vehicular ad hoc networks: A Survey,” Vehicular Technology Magazine, IEEE, vol. 2, no. 2, pp. 12-22, jun 2007 2.B. Karp and H. T. Kung, “GPSR: Greedy perimeter stateless routing for wireless networks,” in Proceedings of the ACM/IEEE International Conference on Mobile Computing and Networking (MobiCom), 2000. 3. C. Lochert, H. Hartenstein, J. Tian, D. Herrmann, H. Füßler, and M. Mauve, “A routing strategy for vehicular ad hoc networks in city environments,” in Proceedeings of IEEE Intelligent Vehicles Symposium (IV2003), pp. 156-161, June 2003. 4. C. Lochert, M. Mauve, H. Füßler, and H. Hartenstein, “Geographic routing in city scenarios,” ACM SIGMOBILE Mobile Computing and Communications Review (MC2R), vol. 9, no. 1, pp. 69-72, Jan.
5. G. Liu, B.-S. Lee, B.-C. Seet, C. H. Foh, K. J. Wong, and L.-K. Lee, “A routing strategy for metropolis vehicular communications,” in International Conference on Information Networking (ICOIN), pp. 134-143, 2004.6. J. Blum, A. Eskandarian, and L. Hoffman, “Mobility management in IVC networks,” in IEEE Intelligent Vehicles Symposium, 2003.7. R. A. Santos, A. Edwards, R. Edwards, and L. seed, “Performance evaluation of routing protocols in vehicular adhoc networks,” The International Journal od Ad Hoc and Ubiquitous Computing, vol. 1, no. ½, pp. 80-91, 2005. 8. M. Durresi, A. Durresi, and L. Barolli, “Emergency broadcast protocol for inter-vehicle communications,” in ICPADS `05: Proceedings of the 11th International Conference on Parallel and Distributed Systems--Workshops (ICPADS`05), 2005.9. G. Korkmaz, E. Ekici, F. Özgüner, and Ü. Özgüner, “Urban multi- hop broadcast protocol for inter-vehicle communication systems,” in ACM International Workshop on Vehicular Ad Hoc Networks, pp. 76-85, 2004. 10. M. Sun, W. Feng, T.-H. Lai, K. Yamada, H. Okada, and K. Fujimura, “GPS-based message broadcasting for inter- vehicle communication,” in ICPP `00: Proceedings of the 2000 International Conference on Parallel Processinf, 2000.11. A. Bachir and A. Benslimane, “A multicast protocol in ad hoc networks inter-vehicle geocast,” in Proceedings of the 57th IEEE Semiannual Vehicular Technology Conference, vol. 4, pp. 2456– 2460, 2003. 12. C. Maihöfer and R. Eberhardt, “Geocast in vehicular environments: Caching and transmission range control for improved efficiency,” in Proceedings of IEEE Intelligent Vehicles Symposium (IV), pp. 951–956, 2004.13. C. Maihöfer, T. Leinmüller, and E. Schoch, “Abiding geocast: time-stable geocast for ad hoc networks,” in Proceedings of the 2nd ACM international workshop on Vehicular ad hoc networks (VANET ’05), pp. 20–29, 2005. 14. J. Luo and J.-P. Hubaux. A survey of inter-vehicle communication. Technical Report IC/2004/24, EPFL, Lausanne, Switzerland, 2004.