FANETs have evolved using the concept ofMANETs and VANETs. To apply routing protocolsto FANETs is one of the common challenges.
Unlike FANETs, generic ADHOC networksare not capable of sending informationquickly and correctly is some situations like atthe time of natural disaster like flooding, earthquakesand even in military battlefield.Optimized Link-State Routing (OLSR), andPredictive-OLSR (P-OLSR) are the techniquesfor dynamic topology adaptation in accordancewith the mobility pattern. OLSR is a routingprotocol where every node maintains one ormore tables representing the entire topologyof the network. P-OLSR which is designed forFANETs is an extension of Optimized Link-StateRouting. OLSR have a benefit of having routesquickly available whenever required.
Simulationresults verify that the proposed schemecould significantly increase the performance offlying ad hoc networks.1 INTRODUCTIONA Wireless ad hoc network is a set of differentnetwork nodes forming the temporary networkswithout the aid of any infrastructure orany centralized administrator. They do nothave gateway, each node can act as the gateway.The wireless arena has been experiencing exponentialgrowth in 21st Century. The IEEEStandard 802.11 is used for ad hoc networkswhich specifies Medium Access control andPhysical layer. In ad hoc network every nodehas its own transmission range, which in turncombine to form a bigger transmission area.Nodes use hopping technique to transmit data.
To make this process more effective , an appropriaterouting algorithm should be implemented.Routing protocols are designed based on therequirement such as when the data is to be exchanged,what data should be exchanged, howroutes are calculated etc. In proactive algorithmsroutes are calculated in the before hand, so nodes use these routes for information ex1change. OLSR routing algorithm is an exampleof proactive.
Whereas in reactive algorithmsroute is calculated only when there are nodesthat need to be communicated. AODV is anexample of reactive algorithm. Multiple pathsbetween any two given nodes can have betterthroughput and recovering connection failureis also easy , but overhead of route discovery inmulti path routing more when compared witha single path routing algorithm.In table driven routing protocol informationabout each node to other node is updated.Where as in on demand routing ,nodes calculateif they have to exchange data from eachother. They don’t have to update information.For a routing algorithm to be best, it shouldovercome the routing problems in ad hoc network.
Some of the routing problems are givenbelow:• Asymmetric Link: If node 1 can hear node2, that does not imply vice versa, hencerouting algorithms should be asymmetric.• Redundant connection: It consumes lot oftime to update routing table , if there aremany redundant connections.• Interference: Interference is very commonas the data transfer is done trough waves.Natural effects like weather, scattering cancause interference.• Dynamic topology : Since nodes can enterand leave network freely, it causes nodeto make frequent changes in their routingtable.FANETs are the special form of MANETs andVANETs. They have much mobility degree andbecause of this topology changes more frequentlythan that of MANETs and VANETs. Distancebetween the nodes is also longer thanthat of the VANETs and MANETs.
Hence communicationrange in FANETs also should belonger than that of others. Since the topologychanges in an unpredictable manner, robustalgorithm with dynamic routing are better tomaintain communication among the nodes inthe FANETs. Dynamic routing in ad hoc networkis an reactive algorithm where the nodesrecover the routes each time a node transfersdata to other node.
Since FANETs have high mobilitynodes, route breakage is very common.Hence route is maintained in order to minimizethe breakage.Network Simulator widely known as NS2, isan event driven simulation tool that has proveduseful in studying dynamic nature of FANETs.NS2 enables users by providing with ways tospecify network protocols and simulate theirbehavior.
The paper has three sections. In thefirst section , OLSR protocol is analyzed , in thesecond part P-OLSR is analyzed. Third sectionis about the simulation results.
2 DESCRIPTION2.1 OLSROLSR is a proactive routing protocol that obtainthe strength of link state algorithm. By proactive,it refers that the entire topology of the networkis maintained by every node. Some of theOLSR performance include: 1. It declares onlylinks with its multipoint relay selectors. 2.
Everynode select from its neighbor, who can read2or who can broadcast messages, those nodesare termed as multipoint relays (MPRs). 3. Thelink state information is partially distributed inthe network, where the link is obtained only betweenan MPR and its MPR selectors. Only MPRcan retransmit in this protocol. For instance,consider a node A having nodes B, C, and Das its neighbors. So, if A considers C as MPR,then whatever packet A sends is received byall the other three nodes.
However, only C canrebroadcast that packet. Hence, this protocolsignificantly reduces the number of retransmissionsin a broadcast procedure and can play asignificant role for dense ad hoc networks.OLSR is not a central entity based but workson a completely distributed manner.
Since thenodes send periodic control messages, it is notcompulsory to have a reliable transmission ofmessages. Types of packets:1. Hello packets: Hello packets are mainlyused to know the state of the link. Everynode transmits their neighbor’s informationand the information about their linkto its neighbors.2. Topology Control Packets: Topology controlpackets are used to give informationabout the MPRs of that particular node.3.
MID packet: If a node has multiple interface,then these MID packets are used tosend the information of those interfacesA HELLO message consists information of itsneighbors and their link status. The addressesof all the neighbors to which there exists a validbi-directional links. And addresses of the neighborthat received the HELLO messages. All onehopnodes receive the message. Every nodein the network periodically sends the HELLOmessages to its neighbors. As a node receivesa HELLO message, it updates its MPR selectortable.
The question is how one can select the multipoint relay. It is selected independently by eachnode on the network. The set of MPRs is selectedin such a way that all the two hop neighborsare selected, and this information can beretrieved from the HELLO messages receivedby this node. Only the information of bi directionallink nodes is taken.
Whenever the nodeof the network is updated or a node fails, themultipoint relay is recalculated.The information of MPR should also be mentionedto other nodes. This can be sent withtopology control packets. The TC messages3sent by the MPR includes its own sequencenumber to the list. The topology control messagesthat are sent throughout the networkhelps in building the topology of the network.The interval of TC messages may depend on theMPR selector set, that if a change occurs thensent early.
A node records information aboutall the multipoint relays, using the informationobtained from TC messages. The informationof all the multipoint relays in a topology table.The table consists of 1. MPR selector set sequencenumber, 2. Destination address, 3.
Lasthop node address. This topology table is executedin the following way:The entry in the topology table is discardedif the MPR selected sequence number is greaterthan the sequence number than sequencenumber in the received message. Second, ifthere is an entry of last-hop address is smallerthan the sequence number in the received, theentry is removed. By using theThird, for each MPR selector received in theTC message, a new topology entry is recordedin the topology table.Each node consists of a routing table, to sendthe packets for other destinations in the network.Routing has information of all the nodes,hence if any of the nodes are updated or declinedthen this routing table needs to be recalculated.
The following procedure1. All the entries from the routing table aredeleted.2. Starting with one hop neighbors as destinationnodes, new entries are added intothe table.
3. The topology table records that are notused while calculating the routing tableare removed.