research article protocol independent adaptive route...

Research ArticleProtocol Independent Adaptive Route Update for VANET

Asim Rasheed1 Sana Ajmal2 and Amir Qayyum1

1 Muhammad Ali Jinnah University Islamabad Pakistan2 Center for Advanced Studies in Engineering Islamabad Pakistan

Correspondence should be addressed to Asim Rasheed asimcorenetorgpk

Received 9 August 2013 Accepted 19 January 2014 Published 27 February 2014

Academic Editors P Garcia-Teodoro and Y Qi

Copyright copy 2014 Asim Rasheed et al This is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

High relative node velocity and high active node density have presented challenges to existing routing approaches within highlyscaled ad hoc wireless networks such as Vehicular Ad hoc Networks (VANET) Efficient routing requires finding optimum routewith minimum delay updating it on availability of a better one and repairing it on link breakages Current routing protocols aregenerally focused on finding and maintaining an efficient route with very less emphasis on route update Adaptive route updateusually becomes impractical for dense networks due to large routing overheads This paper presents an adaptive route updateapproach which can provide solution for any baseline routing protocol The proposed adaptation eliminates the classification ofreactive and proactive by categorizing them as logical conditions to find and update the route

1 Introduction

Dynamics of wireless networks have changed due to roundthe clock data connectivity requirements all over the placeNew and specialized data networks have emerged requiringhigh mobility and scalability These networks may havedisconnected topologies sudden change in active nodedensities and broadcast storms Specialized and complexdeployment and movement patterns in wide areas challengeQoS support Vehicular Ad hoc Networks (VANETs) are acase of the highly fluent wireless networks [1 5 14 22]

Efficient routing aims finding optimum route updatingit on availability of better one and then maintaining itby keeping low overheads Accordingly researchers haveproposed a number of routing protocols using a variety ofmetrics for example hop count node location and so forth[2 6 20 23 24 28 32] Accordingly route finding andmaintenance are done through the following approaches

(1) periodic metrics sharing (proactive routing)(2) event based metrics sharing (reactive routing)(3) derivatives of (1) and (2) for example hybrid or

history oriented approach

Route update on availability of a better one requiresupdated information of network conditions Considering the

definition route update is technically not possible for eventtriggered routing In such protocols route is only updatedonce the old route fails or a new connection or packetexchange is initiated At the same time the periodic updateapproach may add significant overheads to the networktraffic The approach (proactive or reactive) is generally fixedand predefined in the protocol instead of being based onruntime network conditions Adaptation of routes has beenproposed by researchers through multiple approaches forexample [20 28 32]

(i) the inclusion or exclusion of a specific node from theroute according to the changes in runtime conditionssuch as traffic load mobility or node density Thisapproach is generally used for energy conservation orload balancing [32]

(ii) switching between precomputed multiple routesaccording to runtime conditions known as AdaptiveMultipath Routing [7]

(iii) proactive updating of routes using geographical loca-tions and so forth

For Adaptive Multipath Approach one observation sug-gests that as all routes are precomputed though the bestamongst the pool the selected route may not be holistically

Hindawi Publishing Corporatione Scientific World JournalVolume 2014 Article ID 403918 10 pageshttpdxdoiorg1011552014403918

2 The Scientific World Journal

Table 1 Routing update strategy comparison

Primary factor Options Routingpreference

Application needs Control messages ReactiveBroadcast Reactive

Node deploymentAd hoc Proactive

Hybrid (Clustered biased) HybridHybrid (Infrastructure biased) Reactive

Mobility modelRandom Proactive

Bidirectional probabilistic ReactiveMultidirectional probabilistic Hybrid

Node density High ReactiveLow Proactive

Scalability High ReactiveLow Proactive

Relative mobilityHigh Hybrid

Medium ProactiveLow Reactive

QoS requirementThroughput Reactive

Jitter ProactiveDelay Proactive

the optimum one under given conditions For proactiveupdating using geographical locations the requirement oftimely and accurate sharing of updated node locations mightincur additional overheads

Route update strategies must support realistic but diversedeployment mobility patterns and QoS requirements Wepropose an adaptive route update scheme independent of thebaseline routing algorithm which uses logical conditions tofind and update the route

The rest of this paper is organized as follows Section 2analyses the current routing strategies Section 3 explains themodel for adaptive route update Section 4 deals with thesimulation results and their analysis Section 5 concludes thepaper

2 Analysis of Current Routing Strategies

Mobile wireless networks suffer from sudden link break-ages due to topological changes changes in node densitiesand reduction in end-to-end link capacities Resultantlyeach VANET node requires flexible generic and adaptiveroute update and maintenance strategy [29] Different fac-tors directly affect the routing strategies even before routedetermination mechanism such as deployment and mobilitypatterns Table 1 explains the effect of different factors onroute update strategy It can be observed that no singlerouting approach satisfies all conditions A single node mayface a variety of conditions during the same session such aschange in node density and variation in relative node velocity

In subsequent sections we have evaluated current routingmetric sharing strategies with emphasis on state of the art inVANETs

0

01

02

03

04

05

06

07

08

09

Thro

ughp

ut (M

bps)

Time (s)

ThroughputOLSR OHOLSR OH

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95

Total OH 2n

Total OH 86n2n

86n

86n

Figure 1 Comparison of per node routing overheads total over-heads and net throughput

21 Simple Evaluation of Current Metric Sharing ApproachesTo observe the effect of increased routing overheads forproactive routing protocols state-of-the-art OLSRv2 [34]protocol was tested for high node densities We computedthroughput using simple but realistic node topologies withhigh throughput MAC protocol IEEE 80211n To start withtwo mobile nodes from two hop distance were movedtowards each other While staying in direct communicationnode density was gradually increased for more mediumcontention

On increasing the node density significant increase inrouting as well as total overheads was observed at eachnode With the increase in the number of nodes number ofrouting messages as well as size of HELLO packets was alsoincreased Due to increased control as well as data packetseach node faced medium congestion The increase in trans-mission retries also added overheads After increasing thenode density to 86 nodes we observed that total overheadsapproached the actual application throughput for each nodeFor such high rate of overheads randomly selected nodesfaced lack of communication resources for even routing andcontrol packets

Figure 1 shows the increase of total overheads withincrease in routing overheadsThe 119909-axis shows time scale inseconds whereas 119910-axis shows the per node throughput andoverheads in Mbps Five different curves show comparisonbetween routing overheads total overheads and end-to-endthroughput among end nodes for 2 and 86 nodes

The curves in the Figure 1 show that increase in nodedensity also significantly increased routing as well as totaloverheads Both overheads increased approximately eighttimes from initial values Due to high rate of overheadsmany nodes even fail to share the control informationOwingto design limitations the phenomenon of increased totaloverhead will replicate for all protocols with proactive metricsharing scheme regardless of selected metric that is nodelocation or link history

22 VANET Routing Protocols With the emergence of newand more complex requirements on routing new routing

The Scientific World Journal 3

schemes are being proposed by researchers at a fast rateUsing a variety of metrics and three basic metric sharingschemes we have classified the existing routing protocols intothe following major categories

(i) link state and distance vector MANET routing proto-cols such as OLSRv2 and AODVv2 (DYMO) gener-ally perform topology based routing In many casesthese protocols face performance degradation issueswith increased scalability and rapid link breakages[30]

(ii) Broadcast based routing protocols typically floodthe data in entire network Although this approachensures delivery it can only work for small scalenetworks Modifications of this approach such as V-TRADE and HV-TRADE routing protocols [3] limitthe flooding and show improvement over traditionalapproach by reorganizing the network in subgroupsHowever there are significant routing overheads forrebroadcasting

(iii) In overlay routing the routing protocol operateson a set of representative nodes laid over networktopology for example GPCR [31] and CAR [4] Inthe dense environments (eg urban scenarios) streetjunctions can be used as decision points for subse-quent selection of route Similarly use of geographicalfeatures can also help in decision making for routingin highway scenarios Appropriate selection of overlaymap for example junction points can assist in timelydelivery of data using shortest path

(iv) Cluster based routing for example CBRP [15] andCOIN [8] is the combination of the above twotechniques In such schemes each node designatesa cluster-head within a subset of nodes The cluster-head node broadcasts the required packet to clus-ter members Although these protocols answer thescalability issue additional delays and overhead areincurred while forming and maintaining clusters

(v) Infrastructure or road side unit (RSU) based routingprotocols for example RAR [9] and MOVE [25]forms the concept of hybrid networks Being static innature each RSU maintains information about otherRSUs and directly connectedmobile nodes Hence insuch networks maximum reliance is given to RSU forselecting a route to destination

(vi) Location based routing protocols for exampleCMGR [23] and GPSR [12] are generally claimedto be suitable for highly scaled networks such asVANETs The information of nodes all along thepath reduces delay in route determination Use oflocation information instead of hierarchical routingtables significantly reduces routing overheads Theseprotocols answer scalability and delay in routedetermination issues However lack of updatedand exact location of all the nodes can degradethe routing performance [10 21] They can alsosuffer from routing loops and disconnected networktopologies [12]

(vii) Geocast routing for example BBR [17] is a com-bination of broadcast routing and position basedrouting In this scheme data is broadcast within aspecific geographical region around the source Thisscheme is useful for control and safety informationdissemination Some other scheme can be used fordata transmission outside geographical region Thenetwork partitioning and mapping of geographicalregions on road layout are major limitations of thisapproach

(viii) Delay tolerant routing for example VADD [16] andGeOpps [18] is generally a new concept for thenodes spread in the sparse areas As establishingan end-to-end route may not be possible in theabsence of next hop neighbour under disconnectedtopologies packets are buffered till availability of nexthop neighbour This approach is generally known ascarry-and-forward

(ix) A quality of service (QoS) based routing for examplePBR [33] generally performs resource reservationprior to the start of data transfer Such guarantees aredifficult for highly dynamic networks in a determin-istic manner butmay be given in a probabilistic sense

Above stated simple evaluation confirms that currentrouting approaches with predefined and fixed route updatescheme cannot perform even for the simple high nodedensity scenarios Such network scenarios are very commonfor complex networks such as sports stadium and trafficjams These circumstances can cause absolute routing failurefor many successful routing protocols For these routinesituations node density may even increase to thousands ofnodes highlighting need for optimised and adaptive routeupdate

3 Adaptive Route Update Strategy Model

Researchers have observed that localised route maintenanceperforms better than end-to-end route repair due to involve-ment of less number of nodes in the process [2] For efficiencyall nodes are required to perform local route update on twoindependent conditions that is

(i) when the link with next hop neighbour is about tobreak

(ii) when 2nd hop neighbour comes in direct range

For the overall analysis study of the combined impact ofboth conditions is necessaryThe possibility of any neighbournode to affect the host node depends upon two factors

(i) in which direction Δ120579 the neighbour node is movingrelative to host node

(ii) how much distance Δ119889 a neighbour node can coverrelative to the host node during time 119905

Accordingly Δ119889 can be computed as

Δ119889 = (ΔV) 119905 = (V119899cos (120579

119899) minus Vℎ) 119905 (1)


A

H

N(a)

N(e)

D(1)

Y

X

N(c)

D(2)

N(g)

N(h)

N(b)

N(f)

N(d)

B

R1R2

R4

R3

Figure 2 Link ranges for the test topology

where V119899is velocity of next hop neighbour 120579

119899is the relative

angle between the node and its neighbour Vℎis velocity of the

host node and 119905 is the lapsed time since last measurementThe change in topology around any host node 119867 can

be determined by the movement of two hops neighbouringnodes during time 119905 Whereas only nodes capable of cover-ingΔ119889 distance from themaximumhop distance during time119905 will be able to either leave the next hop region or enter in it

Behaviour of the neighbours can be explained by studyingthe area around 119867 Using (1) Figure 2 defines differentcommunication ranges with respect to the host node 119867 Fora relatively simpler model we assumed that 119867 is movingon strip segment 119883119884 for example nodes moving on a roadin VANETs Segment 119883119884 is divided into lanes Direction ofarrow shows that themovement direction of119867119873(119886 119887 119888 and119889) depicts the four next hop neighbours whereas 119873(119890 119891 119892and ℎ) depicts the four 2nd hop neighbours moving in thesame or opposite direction of node119867 respectively119863(1) and119863(2) depict the destination nodes

The probability of selection of next or 2nd hop neighbourin any region will vary according to the selected routingmetric The probability of selecting a next hop node closerto the edge will be more if the routing metric is minimumhop count However the probability will be less for theconsideration ofmaximum stable link For ourmodel we canconsider different generic probabilities for different regionsas

(i) 119875(1) is the probability of next hop neighbour in 1198771for example nodes N(a) and N(c)

(ii) 119875(2) is the probability of next hop neighbour in 1198772for example nodes N(b) and N(d)

(iii) 119875(3) is the probability of 2nd hop neighbour in1198773 forexample nodes N(e) and N(g)

(iv) 119875(4) is the probability of 2nd hop neighbour in1198774 forexample nodes N(f ) and N(h)

For a simple genericmodel it can be assumed thatmobilenodes within the same lane are moving with almost equalspeed Due to bidirectional behaviour of lanes nodes canmove in line or in opposite direction to 119867 or independentlyadapt static behaviour Accordingly variations due to changeof speed and direction of movement will create different linkstability behaviours Resultantly effect of the relative distanceΔ119889 and relative angle of movement Δ120579 between the next and2nd hop neighbour nodes require detailed analysis

For better understanding of node movement Δ119889 and Δ120579we can divide the next hop region into two equal halves onthe axis perpendicular to direction of movement Accordingto simple geometrical layout of node deployment we cancompare both halves as follows

(i) the 2nd hop neighbours which exist towards thedirection of motion of 119867 have a higher probabilityof coming into direct range of119867

(ii) On the other hand next hop neighbour nodes exist-ing in the half opposite to the direction of motion of119867 have a high probability of going out of range of119867

We assume that the maximum transmission range of thehost node is up to the 2nd region (1198772) Hence next hopneighbours can exist within first two regions only Similarly2nd hop neighbours can exist outside 2nd region Moreoverduring 119905 two neighbours can covermaximumdistance of lessthan half of communication range of host nodeHence nodespresent in 1st region (1198771) will remain in communicationduring 119905

Due to limited covered distance in time 119905 Nodes within1198772 will have the chance of link breakage because of its closeproximity to the maximum transmission range boundary Inthe area opposite to direction of movement of119867 (lower half)neighbour nodes moving in line with 119867 but with negativerelative velocity will also be able to go out of range from119867

Neighbouring nodes within same region but movingopposite to 119867 will have double dispersion from 119867 duringtime 119905 This dispersion will introduce expansion in 1198772 ascompared to upper half

Similar to 1198772 nodes in the 3rd region (1198773) will have highprobability of coming within the range of 119867 Like 1198772 nodesin1198773 (upper half)moving opposite to119867will also have doubleconvergence in 119905 Similar to 1198771 nodes in 4th region (1198774) willnot directly affect119867 in time 119905

31 Effect of Δ119889 on Metric Sharing Approaches Routing pro-tocols with time based metric sharing approach use HELLOpackets for assessment of network topology As definedearlier only the nodes present inΔ119889 distance frommaximumhop boundary can go out of range between two consecutiveHELLO intervals (time 119905)Hence sharing of network topologyon all nodes during 119905 will cause resource wastage for nodesoutside Δ119889 The rate of possible nodes involved in change oftopology increases with increase in node speed and durationof HELLO interval


Table 2 Possibility of link breakage in 1198772

Host node status Next hop neighbourstatus

Neighbour node movementdirection

Link breakagepossibility

Probability ofgoing out

Moving Moving Same No

16

Moving Moving Opposite NoMoving Static NoStatic Moving Same YesStatic Moving Opposite NoStatic Static No

Reactive metric sharing approach does not performroute update whereas proactive metric sharing performsroute update on all nodes regardless of their involvementin topology change Hence these protocols do not performoptimum route update and fail for the complicated scenariosAlthoughnumber of adaptive routing protocols are proposedperiodic or event based metric sharing approaches do notconsider this important conclusion

32 Neighbouring Link Behaviour For both of the halvesdepicted in Figure 2mobility can generatemultiple scenariosaccording to change in relative velocity According tomobilityof next or 2nd hop neighbouring nodes with respect to 119867four different possible relations are possible as

(i) both nodes are static(ii) one node is static and other is moving(iii) both nodes are moving in same direction(iv) both nodes are moving in opposite direction

To start with simple relation of two nodes in upper halfof Figure 2 can be evaluated Table 2 defines the relationbetween two neighbouring nodes in right half of 1198772 duringtime 119905 Presence of neighbour node can have six possiblescenarios according to our assumptions and mobility rela-tions defined above Considering the same speed in case ofmobility only one scenario can face link breakage possibilityas defined in Table 2 Hence according to different possibleoptions of mobility and direction of move link breakage willbe possible only for the cases where neighbour node haspositive relative mobility

The same relation can be expanded to multiple lanes forthe next hop neighbour inside right half of 1198772 To make ageneric relation we can consider nodes with uniform density

Considering the different possible scenarios effect ofmobility on link breakage probability can be computedaccording to its relation with the number of lanes Using theresults of Table 2 an expression can be formulated for theprobability of the next hop neighbour going out of range of119867 in time 119905 that is 119875(119873

1(out)) as

119875 (1198731(out)) = sum (119897 minus 119886)

119897 (119897 + 1) + (119897 + 1)2 (2)

where 119886 is the dummy variable for the number of lanes 119897 and119886 = 0 1 2 (119897 minus 1)

From the assumption of uniform node density we knowthat number of nodes 119899 is directly proportional to the numberof lanes 119897 and can have the worst case of single node per laneHence using (2) the probability of next hop neighbour nodeexisting in upper half of 1198772 and going out of range from hostnode is

119875 (1198731(out)) = 1

2

119875 (2) (

sum (119899 minus 119886)

119899 (119899 + 1) + (119899 + 1)2) (3)

similarly in other regions we can formulate themathematicalexpression for probability for next hopneighbour going out ofrange that is 119875(119873

1(out)) or for 2nd hop neighbour coming

in range that is 119875(1198732(in)) Subsequently the overall prob-

ability for optimal route update conditioned on breakage ofthe next hop link or establishment of a direct link with the2nd hop neighbour can be computed using the combinedprobability as

119875 [1198731(out) or 119873

2(in)]

= (

1

2119899 (119899 + 1) + (119899 + 1)2)

times ( ((119875 (2) + 119875 (3)) (sum (119899 minus 119911) + sum(119899 + 119910)))

+ ((119875 (2) + 119875 (3)) 1198992

) minus (

1

2119899 (119899 + 1) + (119899 + 1)2)

times ((119875 (2) 119875 (3)) (sum (119899 minus 119911) + sum(119899 + 119910)) + 1198992

)

2

)

(4)

where 119910 and 119911 are dummy variables for number of nodes119899 119911 = 0 1 2 (119899 minus 1) and 119910 = 0 1 2 119899 119875(2)

and 119875(3) are the probabilities of any node in 1198772 and 1198773and 119875[119873

1(out) or 119873

2(in)] is the probability of next hop

neighbour going out from or 2nd hop neighbour coming indirect communication range of119867

33 Optimum Probability for Route Update The route updateis generally not required by the static networks Howeverfor the networks with high mobility lack of route updatemechanism leads to significant decrease in efficiency Hencewe can state that need of route update or probability ofchange in link status is proportional to the value of changein topology (layout number of nodes in direct range etc)


1 16 32 48 64 80 96 112 128 144 160 176 1920

005

01

015

02

025

03

035

Prob

abili

ty o

f opt

imum

rout

e upd

ate

Node density

Stable linkMinimum hop

Figure 3 Optimum link probability

Thus the change in topology is directly proportional to119875[1198731(out) or 119873

2(in)]

As a test case we can consider the example of routingstrategy ofminimumhop count towards the destinationThisapproach will have higher value of 119875(2) as each node will tryto find the next hop neighbour closest to its communicationboundary The same approach being adapted by the nexthop neighbour will tend to select 2nd hop neighbour also atfarthest distanceThis intention will cause lower value of119875(3)for 2nd hop neighbour On the other hand the behaviour ofa routing strategy based on maximum stable link will havelower 119875(2) and higher 119875(3) value However this behaviourmay not stand true for strategies which do not depend onnode distance such as minimum cost or load balance

Equation (4) provides a very interesting result as shownin Figure 3 The 119909-axis shows the node density whereas 119910-axis shows the probability of optimum route update due tolink breakage with next hop neighbour or link establishmentwith 2nd hop neighbour For the computation of com-bined probability behaviour curve of minimum hop countapproach pdf curve for linear distribution is used Similarlyfor maximum stable link approach pdf curve for CentralizedPareto Distribution is used

Both the curves drawn for routing approach of stable linkand minimum hop count follow the same pattern and aremonotonically increasing with increase in number of nodes119899 It suggests that the requirement for route update is directlyproportional to change in network topology

The proposedmodel as defined in (4) is based on differentcommunicating regions defined by (1) The time 119905 is directlyproportional to region size as shown in (1) which is againdirectly proportional to number of nodes 119899 as shown in (4)Hence we can draw two important conclusions as

(i) by keeping the 119899 constant the probability of changein network conditions for optimum route updateincreases with increase in the update interval 119905

(ii) by keeping the update interval 119905 constant the prob-ability of change in network conditions for optimumroute update increases with increase in 119899

The above stated conclusions are the baseline factorsfor our adaptive route update strategy The modificationof current routing protocols by introducing adaptive routeupdate can significantly improve the efficiency of the net-work Hence irrespective of route finding approach therequirement for route update increases with increase in nodedensity

The readermay argue that with a decreasing node densitythe existing route may not be the best route any moreHowever it must not be ignored that if the node densityis decreasing then in an average sense the probability ofavailability of a better route also decreases (as there are lesserchoices) Subsequently the same conclusion can also be statedas follows

A higher node density requires a higher rate of changein the network conditions for a state where route updatebecomes an optimum choice that is where either next hopnode is about to face link breakage or 2nd hop node hasalready come in next hop range

4 Evaluation of Adaptive Routing Strategy

As described earlier many other studies along with alreadystated results confirm that current predefined and fixedroute update schemes may not perform efficiently underall circumstances Resultantly emphasis should be given totwo basic but important issues related to routing strategiesA generic and flexible but realistic solution to these issuescan enhance the routing efficiency These issues include thefollowing

(i) Which type(s) of metrics can be shared for routefinding

(ii) How these metrics can be shared that is adaptivelyor through fixed scheme

Literature review of routing algorithms reveals that majoremphasis is given to first issue and improved solutions arestill being suggested whereas the other issue need reemphasisaccording to requirements of upcoming complex networks

To utilize the results achieved through (4) we propose theterm adaptation for route update strategy Although adap-tation has already been proposed for routing optimizationfor adaptation has been generally neglected in our currentstatic routing approaches The principal design dissimilaritybetween adaptive and optimized adaptive approach is the useof different metrics according to runtime network conditionsto adaptively find and update the best possible route Theoptimized adaptation will use the logical conditions to findand update the route by abolishing the taxonomy of reactiveand proactive Optimized adaptation leads to use of metricsat runtime for route update strategy Accordingly metricsneed to be redefined as per their use for route finding andupdate and maintenance as a single metric may not performoptimally for all routing roles In the light of (4) differentthreshold values for each metric or combination of metricscomputed on runtime can develop adaptive route updateapproach


As already described a localized metric can show morepromising results as compared to end-to-end metric forlocal route repair Analysis of networks with large topol-ogy changes emphasizes use of different types of met-rics for efficient routing strategy for example QoS relatedmetrics position related metrics and PHY layer metricsThese routing metrics which are also cross layer in designindependently define changes in network topology Thedifferent threshold values for each metric or combina-tion of metrics computed on runtime can be used foradaptive route update However routing strategies withmultimetrics schemes can also be researched for the pur-pose

QoS related metrics such as throughput delay andpacket pair delay are often used in routing schemes [11]Considering the use of local route repair QoS metrics fornext hop only performs more efficiently than end-to-endmetrics

The use of position relatedmetrics [8] such as neighbour-ing nodes average node distance and number of neighbourshas significant importance for networks such as VANETsThese metrics are also considered for geographical address-ing For highly fluent networks one hop neighbours listprovidesmore realistic results for change in network topologythan average neighbour count or neighbour distance

The use of physical layer metrics such as SINR andreceived power require more complicated algorithm designsbut provide more stable routes [19] Being a combinationof received signal strength noise and interference SINRprovides more promising results than considering receivedsignal strength only

To verify our research for different possible scenarios ofVANETs we performed various simulations in NS-2 Aftercomprehensive verification of the proposed schemewe testedit against some state-of-the-art routing protocols

As a test platform for the adaptive route update approachwe modified standard AODV routing protocol and namedit as Adaptive AODV (AAODV) AODV being a reactiverouting protocol update its route on link breakage onlyHow-ever it continuously shares HELLO messages to learn aboutits neighbours AODV was modified in a manner whereeach node was defined to continuously measure the metricfor route update for example SINR and adaptively updateits route if change in metric value achieves a predefinedthreshold

As a test metric one hop neighbours list next hopthroughput and SINR were selected All three metricsshowed significant improvement in overall performanceSelection of SINR can be argued considering real life imple-mentation as one hop neighbours list can easily be com-puted using standard HELLO messages However with theadvancement in cross layer design the use of physical layermetrics is emerging Hence considering the slight edge inperformance SINR was selected for subsequent evaluation

SINR also indirectly shows the change in node topologyaround 119867 The change in received signal power can showthe change in either signal (119878) strength or noise (119873) plusinterference (119868) strength Increase in SINR means eitherreduction in interfering node density or next hop node

distance and vice versa Each node can compute SINR valuearound itself for each next hop neighbour using formula [13]

SINR =

119878

119873 + 119868

=

119875119905ℎ119905119897 (119903)

119873 + sum120580120598120601119904

(119875119894ℎ119894119897119870

1003816100381610038161003816119909minus119886

119894

1003816100381610038161003816)

(5)

where 119875119905= Transmitted power (constant) ℎ

119905= Channel gain

of transmitter (constant) 119875119894= Transmitted power from 119894th

node (constant) 119903=Transmitter to receiver distance and119909119894=

Distance between 119894th interfering node and receiverFor the simulation purpose highly mobile nodes

(20ndash250) moving at variable velocities (1ndash130 kmph) weredeployed in sparse area Nodes were converged fromhighways to a single road crossing Node density wasgradually increased by converging all nodes within one hopregion After reaching the road crossing region all nodeswere forced to adapt a temporary static behaviour Afterstaying for a while in one hop region all nodes moved indifferent directions

For the comparative study of different routing protocolsagainst optimized AAODV 28 different simulations were runfor 10 repetitions each 7 different topologies were simulatedunder different scenarios and conditions by varying the num-ber of nodes Nodes were randomly selected with randomtimings to generate TCP and UDP traffic

Using the percentage change in various metrics valuesthe optimum local route update conditions can be workedout To find the optimum threshold value for route updatequality of service (QoS) parameters were measured by keep-ing different threshold values of the metric under differentnode densities After acquiring the coarse grained opti-mum threshold values of SINR we compared our adaptiveapproach with other state-of-the-art routing protocols Theparameters considered for the simulations were as under

(i) number of active nodes 4 8 16 32 56 64 and 128(ii) number of passive nodes 9 and 18(iii) MAC protocol IEEE 80211p(iv) number of lanes 2 4 and 8(v) node speeds 16 32 50 65 80 96 112 and 130 kmph(vi) comparative protocols AAODV AODVv2 (DYMO)

OLSRv2 FROMR and XORi

SINR value can vary from zero to approaching infinity(very high) that is either signal strength approaches zero(near link breakage) or approaches infinity (no noise andinterference) Hence different threshold values of SINRchange were considered from 1 to 1 Billion QoS parameterswere measured by keeping different threshold values of SINRto find the optimum values under different node densitiesOptimum threshold values for different node densities werecomputed at which adaptive route update provided bestresults

Figure 4 shows the graph of optimum change in SINRthreshold for different node densities The 119909-axis of thegraph shows the combination of different node densitieswhereas119910-axis shows the percentage value of change at whichlocal route update provided maximum data transfer for the


10

100

1000

10000

100000

1000000

10000000

100000000

1000000000

24 32 48 72 82 125 150 175 200 225 250

Opt

imum

thre

shol

d ch

ange

(Log

)

Nodes

Figure 4 Optimum threshold change

particular topology As the difference between lowest andhighest threshold value is too large the logarithmic scale wasselected for better analyses of the curve It can be observedthat the value of optimum change threshold increases withincrease in the node density This result shows that theoptimum threshold is dependent on total nodes involved indata transfer Minimum value of change in SINR (50) wasobserved for 24 nodes whereas maximum value of 1 Billionwas recorded for 250 nodes topology

Interestingly the optimum threshold curve for the SINRchange follows the samepattern as themathematicalmodel in(4)The little variations can be considered due to randomnessin topologies and variations in routing approach Similarto Figure 2 the difference in threshold values for sparsetopologies is more as compared to the dense ones Thisresult confirms the analysis that optimum threshold value isdependent on active nodes within given area

Optimum threshold value starts from lower values andthen increases to achieve somemaximum value After attain-ing the higher threshold values for higher node densitiesthe change in optimum values decreases significantly andthe curve becomes flat The threshold curve flattens for largenode densities Although being very low increasing trendcontinues for all values

After attaining the coarse grained optimum thresholdvalues of selected test metrics for route update the com-parison of optimized AAODV with other standard protocolswas done Figures 5 and 6 show the comparison of AAODVagainst AODVv2 (DYMO) OLSRv2 FROMR [26] and XORi[27] AODVv2 and OLSRv2 are designed as routing protocolsfor Mobile Ad hoc Network (MANET) whereas the lattertwo are specifically designed for VANETs In Figure 5 the119909-axis of the graph shows the combination of different nodedensitiesThe 119910-axis shows the average of ratio of normalizedthroughput independently computed for each node densityand topology for different protocols against AAODV

The comparison of curves for the low node densitiesshows better performance by XORi OLSRv2 and AAODVWith change in topology XORi and OLSRv2 timely updatedtheir routes Resultantly due to low overheads at lower nodedensity both protocols showed comparative performancewith AAODV However due to drastic increase in overheads

0

01

02

03

04

05

06

07

08

09

1

20 24 32 48 72 82 150

Thro

ughp

ut (n

orm

aliz

ed)

Node density

FROMRXORiAODVv2

OLSRv2AAODV

Figure 5Throughput comparison of proposed approach (indepen-dently normalised for each topology)

at higher node densities both protocols showed significantreduction in performance against AAODV

On increasing the node density higher contention andcontrol overheads caused lack of sufficient bandwidth forcontrol messages for OLSRv2 The curve confirmed theprevious analysis of Figure 1

From other two protocols FROMR and AODVv2 per-formed equivalently to each other However FROMR had aslight edge due to use of alternate available routes For lownode densities both routing protocols remained inefficientdue to nonadaptation to topology changes By increasingthe node density more options were made available forroute to destination However performance of both protocolsremained low even for higher node densities For all nodedensities both protocols did not optimize their routing tableson availability of more suitable path Resultantly AAODVoutperformed in all cases

On increasing the node density further the gap betweencurves of both protocols and AAODV started decreasingDue to design limitation of three retries higher contentionrate caused repeated false route failures for these protocolsForced route error messages caused route update even inpresence of old route Repeated unintentional route updateimproved the performance of reactive protocols The samebehaviour can be seen at node density of 150 active nodesHowever the unintentional update did not cause all nodesto update their route Similarly the behaviour cannot beguaranteed under all topologies

Figure 6 shows the comparison of delay for selectedprotocols The 119909-axis of the graph shows the combination ofdifferent node densitiesThe 119910-axis shows the average of end-to-end delay for each topology We can observe that for thelow node densities OLSRv2 offered minimum delay owingto its simple proactive design However on increasing thenode density increased overheads caused significant increasein end-to-end delay For higher node densities AODVv2showed minimum delay due to its proactive design and falseroute update as already defined


0

01

02

03

04

05

06

07

08

09

1

20 24 32 48 72 82 150

Aver

age d

elay

(s)

Node density

FROMRXORiAODVv2

OLSRv2AAODV

Figure 6 Delay comparison of proposed approach

Use of adaptive route update can make any routing pro-tocol more efficient Optimization of adaptive route updatecan be performed on different node densities and typeof networks However the threshold level for optimizedadaptive route update may vary for different networks

5 Conclusion

Routing protocols can be divided into many categoriesaccording to the algorithm and modifications proposedagainst other protocols Using different metrics routingprotocols can be grouped into three categories according torouting metric sharing methodThese types include periodictopology sharing event based topology sharing and theirderivatives (hybrid and history oriented)

Behaviour of any routing algorithms differs for staticand mobile network scenarios Contrary to static networksmobile wireless networks suffer from sudden link breakagesdue to topological changes change in node densities andreduction in average link capacities Resultantly any singleprotocol may not perform well under all scenarios andconditions The analysis of routing strategies proves theirinefficiency for complex networks

Without incorporating adaptive route update existingprotocols can provide satisfactory results for networks withlimited topology changes and limited number of nodesHowever for large scale networks or networks involvingrapid topology changes current routing strategies will faceperformance issues

We developed a mathematical model to describe thebehaviour of changes in network conditions The curves ofthe model provide a very interesting result as the prob-ability for optimum route update is directly proportionalto time interval and node density Rapid topology changesdemand adaptive use of runtime intelligence for route updateThe proposed adaptive route update scheme which can beimplemented with any baseline routing algorithm will allownodes to locally optimize their routes Hence adaptive useof runtime conditions will replace the terms reactive or

proactive protocols with logical conditions to find the mostoptimum route at any given time

Regardless of baseline routing approach adaptive routeupdate based on different metrics can make a routingprotocol more efficient than a routing protocol with staticroute update approach As a future work we intend to extendtheoretical analysis of the proposed approach includingcomplexity analysis We also intend to modify some otherrouting protocols according to the proposed approach toverify the improvements against their standard versions

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

References

[1] A Rasheed A Qayyum and S Ajmal ldquoSecurity architectureparameters in VANETsrdquo in Proceedings of the 2nd InternationalConference on Information and Emerging Technologies (ICIETrsquo10) pp 1ndash6 Karachi Pakistan June 2010

[2] A Xeros M Lestas M Andreou and A Pitsillides ldquoAdaptiveprobabilistic flooding for InformationHovering in VANETsrdquo inProceedings of the IEEE Vehicular Networking Conference (VNCrsquo10) pp 239ndash246 Jersey City NJ USA December 2010

[3] M T Sun W C Feng T H Lai K Yamada H Okadaand K Fujimura ldquoGPS-based message broadcast for adaptiveinter-vehicle communicationsrdquo in Proceedings of the IEEE 52ndVehicular Technology Conference pp 279ndash286 September 2000

[4] Y C Tseng S Y Ni Y S Chen and J P Sheu ldquoThe broadcaststorm problem in a mobile ad hoc networkrdquoWireless Networksvol 8 no 2-3 pp 153ndash167 2002

[5] A Dahiya and R Chauhan ldquoA comparative study of manet andVANET environmentrdquo Journal of Computing vol 2 no 7 pp87ndash92 2010

[6] S Mohseni R Hassan A Patel and R Razali ldquoComparativereview study of reactive and proactive routing protocols inMANETsrdquo in Proceedings of the 4th IEEE International Confer-ence onDigital Ecosystems andTechnologies (DEST rsquo10) pp 304ndash309 Dubai United Arab Emirates April 2010

[7] Y Shinohara Y Chiba and H Shimonishi ldquoAn adaptivemultipath routing algorithm formaximizing flow throughputsrdquoin Proceedings of theWorld Telecommunications Congress (WTCrsquo12) pp 1ndash6 March 2012

[8] F Li and Y Wang ldquoRouting in vehicular ad hoc networks asurveyrdquo IEEE Vehicular Technology Magazine vol 2 no 2 pp12ndash22 2007

[9] M Slavik I Mahgoub and M Rathod ldquoStatistical broadcastprotocol design with WiBDAT wireless Broadcast design andanalysis toolrdquo in Proceedings of the IEEE Wireless Communi-cations and Networking Conference (WCNC rsquo11) pp 1236ndash1241Cancun Mexico March 2011

[10] R C Shah A Wolisz and J M Rabaey ldquoOn the performanceof geographical routing in the presence of localization errorsrdquoin Proceedings of the IEEE International Conference on Commu-nications (ICC rsquo05) pp 2979ndash2985 May 2005

[11] M S Rainer Baumann S Heimlicher and AWeibel ldquoA surveyon routing metricsrdquo 2007

[12] B Karp and H T Kung ldquoGPSR Greedy Perimeter StatelessRouting for wireless networksrdquo in Proceedings of the 6th Annual


International Conference on Mobile Computing and Networking(MOBICOM rsquo00) pp 243ndash254 August 2000

[13] A GoldsmithWireless Communications Cambridge universitypress New York NY USA 2005

[14] S A Mohammad A Rasheed and A Qayyum ldquoVANETarchitectures and protocol stacks a surveyrdquo in CommunicationTechnologies for Vehicles pp 95ndash105 Springer Berlin Germany2011

[15] B Zarei M Zeynali and VMajid Nezhad ldquoNovel cluster basedrouting protocol in wireless sensor networksrdquo InternationalJournal of Computer Science Issues vol 7 no 4 pp 32ndash36

[16] J Zhao and G Cao ldquoVADD Vehicle-assisted data deliveryin vehicular ad hoc networksrdquo in Proceedings of the 25thIEEE International Conference on Computer Communications(INFOCOM rsquo06) pp 1ndash12 Barcelona Spain April 2006

[17] M Zhang and R S Wolff ldquoBorder node based routing protocolfor VANETs in sparse and rural areasrdquo in Proceedings of theIEEE Globecom Workshops pp 1ndash7 Washington DC USANovember 2007

[18] I Leontiadis and C Mascolo ldquoGeOpps Geographical oppor-tunistic routing for vehicular networksrdquo in Proceedings of theIEEE International Symposium on a World of Wireless Mobileand Multimedia Networks (WOWMOM rsquo07) pp 1ndash6 EspooFinland June 2007

[19] A Rasheed and SAjmal ldquo3D-a doppler directivity anddistancebased architecture for selecting stable routing links inVANETsrdquoin Proceedings of the 2nd International Conference on ComputerControl and Communication (IC4 rsquo09) pp 1ndash5 Karachi Pak-istan February 2009

[20] M Slavik and I Mahgoub ldquoSpatial distribution and channelquality adaptive protocol for multi-hop wireless broadcastrouting in VANETrdquo IEEE Transactions on Mobile Computingvol 12 no 4 pp 722ndash734 2013

[21] Y Kim J J Lee and A Helmy ldquoModeling and analyzing theimpact of location inconsistencies on geographic routing inwireless networksrdquo ACM SIGMOBILE Mobile Computing andCommunications Review vol 8 no 1 pp 48ndash60 2004

[22] A Rasheed H Zia F Hashmi U Hadi W Naim and SAjmal ldquoFleet amp convoy management using VANETrdquo Journal ofComputer Networks vol 1 no 1 pp 1ndash9 2013

[23] K Shafiee and V C M Leung ldquoConnectivity-aware minimum-delay geographic routing with vehicle tracking in VANETsrdquo AdHoc Networks vol 9 no 2 pp 131ndash141 2011

[24] M Al-Rabayah and R Malaney ldquoA new scalable hybrid routingprotocol for VANETsrdquo IEEE Transactions on Vehicular Technol-ogy vol 61 no 6 pp 2625ndash2635 2012

[25] R Barr Z J Haas and R Van Renesse ldquoJiST an efficientapproach to simulation using virtual machinesrdquo Software vol35 no 6 pp 539ndash576 2005

[26] C S Wu S C Hu and C S Hsu ldquoDesign of fast restorationmultipath routing in VANETsrdquo in Proceedings of the Inter-national Computer Symposium (ICS rsquo10) pp 73ndash78 TainanTaiwan December 2010

[27] R Oliveira A Garrido R Pasquini et al ldquoTowards the use ofXOR-based routing protocols in vehicular ad hoc networksrdquo inProceedings of the IEEE 73rd Vehicular Technology Conference(VTC Spring rsquo11) pp 1ndash6 Budapest Hungary May 2011

[28] H Saleet R Langar O Basir and R Boutaba ldquoAdaptivemessage routing with QoS support in vehicular Ad Hoc net-worksrdquo in Proceedings of the IEEE Global TelecommunicationsConference (GLOBECOM rsquo09) pp 1ndash6 Honolulu Hawaii USADecember 2009

[29] S Zeadally R Hunt Y S Chen A Irwin and A HassanldquoVehicular ad hoc networks (VANETs) status results andchallengesrdquo Telecommunication Systems vol 50 no 4 pp 217ndash241 2012

[30] M B S Jaap and L Wolf ldquoEvaluation of routing protocolsfor vehicular ad hoc networks in city traffic scenariosrdquo inProceedings of the 5th International Conference on ITS Telecom-munications no 5 Brest France 2005

[31] T Song W Xia T Song and L Shen ldquoA cluster-based direc-tional routing protocol in VANETrdquo in Proceedings of the 12thIEEE International Conference on Communication Technology(ICCT rsquo10) pp 1172ndash1175 Nanjing China November 2010

[32] D Murray M Dixon and T Koziniec ldquoAn experimentalcomparison of routing protocols inmulti hop ad hoc networksrdquoin Proceedings of the Australasian Telecommunication Networksand Applications Conference (ATNAC rsquo10) pp 159ndash164 Auck-land New Zealand November 2010

[33] V Namboodiri and L Gao ldquoPrediction-based routing forvehicular Ad Hoc networksrdquo IEEE Transactions on VehicularTechnology vol 56 no 4 pp 2332ndash2345 2007

[34] T Clausen C Dearlove P Jacquet et al ldquoThe optimized linkstate routing protocol version 2rdquo draft-ietf-manet-olsrv2-19Work in progress 2013

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

RoboticsJournal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Active and Passive Electronic Components

Control Scienceand Engineering

Journal of



RotatingMachinery


Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Submit your manuscripts athttpwwwhindawicom

VLSI Design



Shock and Vibration


Civil EngineeringAdvances in

Acoustics and VibrationAdvances in



Electrical and Computer Engineering

Journal of

Advances inOptoElectronics


Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of


Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014


Chemical EngineeringInternational Journal of Antennas and

Propagation




Navigation and Observation



DistributedSensor Networks



Table 1 Routing update strategy comparison

Primary factor Options Routingpreference

Application needs Control messages ReactiveBroadcast Reactive

Node deploymentAd hoc Proactive

Hybrid (Clustered biased) HybridHybrid (Infrastructure biased) Reactive

Mobility modelRandom Proactive

Bidirectional probabilistic ReactiveMultidirectional probabilistic Hybrid

Node density High ReactiveLow Proactive

Scalability High ReactiveLow Proactive

Relative mobilityHigh Hybrid

Medium ProactiveLow Reactive

QoS requirementThroughput Reactive

Jitter ProactiveDelay Proactive

the optimum one under given conditions For proactiveupdating using geographical locations the requirement oftimely and accurate sharing of updated node locations mightincur additional overheads

Route update strategies must support realistic but diversedeployment mobility patterns and QoS requirements Wepropose an adaptive route update scheme independent of thebaseline routing algorithm which uses logical conditions tofind and update the route

The rest of this paper is organized as follows Section 2analyses the current routing strategies Section 3 explains themodel for adaptive route update Section 4 deals with thesimulation results and their analysis Section 5 concludes thepaper

2 Analysis of Current Routing Strategies

Mobile wireless networks suffer from sudden link break-ages due to topological changes changes in node densitiesand reduction in end-to-end link capacities Resultantlyeach VANET node requires flexible generic and adaptiveroute update and maintenance strategy [29] Different fac-tors directly affect the routing strategies even before routedetermination mechanism such as deployment and mobilitypatterns Table 1 explains the effect of different factors onroute update strategy It can be observed that no singlerouting approach satisfies all conditions A single node mayface a variety of conditions during the same session such aschange in node density and variation in relative node velocity

In subsequent sections we have evaluated current routingmetric sharing strategies with emphasis on state of the art inVANETs

0

01

02

03

04

05

06

07

08

09

Thro

ughp

ut (M

bps)

Time (s)

ThroughputOLSR OHOLSR OH

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95

Total OH 2n

Total OH 86n2n

86n

86n

Figure 1 Comparison of per node routing overheads total over-heads and net throughput

21 Simple Evaluation of Current Metric Sharing ApproachesTo observe the effect of increased routing overheads forproactive routing protocols state-of-the-art OLSRv2 [34]protocol was tested for high node densities We computedthroughput using simple but realistic node topologies withhigh throughput MAC protocol IEEE 80211n To start withtwo mobile nodes from two hop distance were movedtowards each other While staying in direct communicationnode density was gradually increased for more mediumcontention

On increasing the node density significant increase inrouting as well as total overheads was observed at eachnode With the increase in the number of nodes number ofrouting messages as well as size of HELLO packets was alsoincreased Due to increased control as well as data packetseach node faced medium congestion The increase in trans-mission retries also added overheads After increasing thenode density to 86 nodes we observed that total overheadsapproached the actual application throughput for each nodeFor such high rate of overheads randomly selected nodesfaced lack of communication resources for even routing andcontrol packets

Figure 1 shows the increase of total overheads withincrease in routing overheadsThe 119909-axis shows time scale inseconds whereas 119910-axis shows the per node throughput andoverheads in Mbps Five different curves show comparisonbetween routing overheads total overheads and end-to-endthroughput among end nodes for 2 and 86 nodes

The curves in the Figure 1 show that increase in nodedensity also significantly increased routing as well as totaloverheads Both overheads increased approximately eighttimes from initial values Due to high rate of overheadsmany nodes even fail to share the control informationOwingto design limitations the phenomenon of increased totaloverhead will replicate for all protocols with proactive metricsharing scheme regardless of selected metric that is nodelocation or link history

22 VANET Routing Protocols With the emergence of newand more complex requirements on routing new routing





















Δ119889 = (ΔV) 119905 = (V119899cos (120579

119899) minus Vℎ) 119905 (1)


A

H

N(a)

N(e)

D(1)

Y

X

N(c)

D(2)

N(g)

N(h)

N(b)

N(f)

N(d)

B

R1R2

R4

R3





























16








1(out)) as

119875 (1198731(out)) = sum (119897 minus 119886)

119897 (119897 + 1) + (119897 + 1)2 (2)



119875 (1198731(out)) = 1

2

119875 (2) (

sum (119899 minus 119886)

119899 (119899 + 1) + (119899 + 1)2) (3)





119875 [1198731(out) or 119873

2(in)]

= (

1

2119899 (119899 + 1) + (119899 + 1)2)


+ ((119875 (2) + 119875 (3)) 1198992

) minus (

1

2119899 (119899 + 1) + (119899 + 1)2)


)

2

)

(4)



1(out) or 119873





1 16 32 48 64 80 96 112 128 144 160 176 1920

005

01

015

02

025

03

035

Prob

abili

ty o

f opt

imum

rout

e upd

ate

Node density




2(in)]


























SINR =

119878

119873 + 119868

=

119875119905ℎ119905119897 (119903)

119873 + sum120580120598120601119904

(119875119894ℎ119894119897119870

1003816100381610038161003816119909minus119886

119894

1003816100381610038161003816)

(5)














10

100

1000

10000

100000

1000000

10000000

100000000

1000000000

24 32 48 72 82 125 150 175 200 225 250

Opt

imum

thre

shol

d ch

ange

(Log

)

Nodes







0

01

02

03

04

05

06

07

08

09

1

20 24 32 48 72 82 150

Thro

ughp

ut (n

orm

aliz

ed)

Node density

FROMRXORiAODVv2

OLSRv2AAODV








0

01

02

03

04

05

06

07

08

09

1

20 24 32 48 72 82 150

Aver

age d

elay

(s)

Node density

FROMRXORiAODVv2

OLSRv2AAODV



5 Conclusion









References







































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation





























Δ119889 = (ΔV) 119905 = (V119899cos (120579

119899) minus Vℎ) 119905 (1)


A

H

N(a)

N(e)

D(1)

Y

X

N(c)

D(2)

N(g)

N(h)

N(b)

N(f)

N(d)

B

R1R2

R4

R3





























16








1(out)) as

119875 (1198731(out)) = sum (119897 minus 119886)

119897 (119897 + 1) + (119897 + 1)2 (2)



119875 (1198731(out)) = 1

2

119875 (2) (

sum (119899 minus 119886)

119899 (119899 + 1) + (119899 + 1)2) (3)





119875 [1198731(out) or 119873

2(in)]

= (

1

2119899 (119899 + 1) + (119899 + 1)2)


+ ((119875 (2) + 119875 (3)) 1198992

) minus (

1

2119899 (119899 + 1) + (119899 + 1)2)


)

2

)

(4)



1(out) or 119873





1 16 32 48 64 80 96 112 128 144 160 176 1920

005

01

015

02

025

03

035

Prob

abili

ty o

f opt

imum

rout

e upd

ate

Node density




2(in)]


























SINR =

119878

119873 + 119868

=

119875119905ℎ119905119897 (119903)

119873 + sum120580120598120601119904

(119875119894ℎ119894119897119870

1003816100381610038161003816119909minus119886

119894

1003816100381610038161003816)

(5)














10

100

1000

10000

100000

1000000

10000000

100000000

1000000000

24 32 48 72 82 125 150 175 200 225 250

Opt

imum

thre

shol

d ch

ange

(Log

)

Nodes







0

01

02

03

04

05

06

07

08

09

1

20 24 32 48 72 82 150

Thro

ughp

ut (n

orm

aliz

ed)

Node density

FROMRXORiAODVv2

OLSRv2AAODV








0

01

02

03

04

05

06

07

08

09

1

20 24 32 48 72 82 150

Aver

age d

elay

(s)

Node density

FROMRXORiAODVv2

OLSRv2AAODV



5 Conclusion









References







































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










A

H

N(a)

N(e)

D(1)

Y

X

N(c)

D(2)

N(g)

N(h)

N(b)

N(f)

N(d)

B

R1R2

R4

R3





























16








1(out)) as

119875 (1198731(out)) = sum (119897 minus 119886)

119897 (119897 + 1) + (119897 + 1)2 (2)



119875 (1198731(out)) = 1

2

119875 (2) (

sum (119899 minus 119886)

119899 (119899 + 1) + (119899 + 1)2) (3)





119875 [1198731(out) or 119873

2(in)]

= (

1

2119899 (119899 + 1) + (119899 + 1)2)


+ ((119875 (2) + 119875 (3)) 1198992

) minus (

1

2119899 (119899 + 1) + (119899 + 1)2)


)

2

)

(4)



1(out) or 119873





1 16 32 48 64 80 96 112 128 144 160 176 1920

005

01

015

02

025

03

035

Prob

abili

ty o

f opt

imum

rout

e upd

ate

Node density




2(in)]


























SINR =

119878

119873 + 119868

=

119875119905ℎ119905119897 (119903)

119873 + sum120580120598120601119904

(119875119894ℎ119894119897119870

1003816100381610038161003816119909minus119886

119894

1003816100381610038161003816)

(5)














10

100

1000

10000

100000

1000000

10000000

100000000

1000000000

24 32 48 72 82 125 150 175 200 225 250

Opt

imum

thre

shol

d ch

ange

(Log

)

Nodes







0

01

02

03

04

05

06

07

08

09

1

20 24 32 48 72 82 150

Thro

ughp

ut (n

orm

aliz

ed)

Node density

FROMRXORiAODVv2

OLSRv2AAODV








0

01

02

03

04

05

06

07

08

09

1

20 24 32 48 72 82 150

Aver

age d

elay

(s)

Node density

FROMRXORiAODVv2

OLSRv2AAODV



5 Conclusion









References







































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation
















16








1(out)) as

119875 (1198731(out)) = sum (119897 minus 119886)

119897 (119897 + 1) + (119897 + 1)2 (2)



119875 (1198731(out)) = 1

2

119875 (2) (

sum (119899 minus 119886)

119899 (119899 + 1) + (119899 + 1)2) (3)





119875 [1198731(out) or 119873

2(in)]

= (

1

2119899 (119899 + 1) + (119899 + 1)2)


+ ((119875 (2) + 119875 (3)) 1198992

) minus (

1

2119899 (119899 + 1) + (119899 + 1)2)


)

2

)

(4)



1(out) or 119873





1 16 32 48 64 80 96 112 128 144 160 176 1920

005

01

015

02

025

03

035

Prob

abili

ty o

f opt

imum

rout

e upd

ate

Node density




2(in)]


























SINR =

119878

119873 + 119868

=

119875119905ℎ119905119897 (119903)

119873 + sum120580120598120601119904

(119875119894ℎ119894119897119870

1003816100381610038161003816119909minus119886

119894

1003816100381610038161003816)

(5)














10

100

1000

10000

100000

1000000

10000000

100000000

1000000000

24 32 48 72 82 125 150 175 200 225 250

Opt

imum

thre

shol

d ch

ange

(Log

)

Nodes







0

01

02

03

04

05

06

07

08

09

1

20 24 32 48 72 82 150

Thro

ughp

ut (n

orm

aliz

ed)

Node density

FROMRXORiAODVv2

OLSRv2AAODV








0

01

02

03

04

05

06

07

08

09

1

20 24 32 48 72 82 150

Aver

age d

elay

(s)

Node density

FROMRXORiAODVv2

OLSRv2AAODV



5 Conclusion









References







































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










1 16 32 48 64 80 96 112 128 144 160 176 1920

005

01

015

02

025

03

035

Prob

abili

ty o

f opt

imum

rout

e upd

ate

Node density




2(in)]


























SINR =

119878

119873 + 119868

=

119875119905ℎ119905119897 (119903)

119873 + sum120580120598120601119904

(119875119894ℎ119894119897119870

1003816100381610038161003816119909minus119886

119894

1003816100381610038161003816)

(5)














10

100

1000

10000

100000

1000000

10000000

100000000

1000000000

24 32 48 72 82 125 150 175 200 225 250

Opt

imum

thre

shol

d ch

ange

(Log

)

Nodes







0

01

02

03

04

05

06

07

08

09

1

20 24 32 48 72 82 150

Thro

ughp

ut (n

orm

aliz

ed)

Node density

FROMRXORiAODVv2

OLSRv2AAODV








0

01

02

03

04

05

06

07

08

09

1

20 24 32 48 72 82 150

Aver

age d

elay

(s)

Node density

FROMRXORiAODVv2

OLSRv2AAODV



5 Conclusion









References







































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation



















SINR =

119878

119873 + 119868

=

119875119905ℎ119905119897 (119903)

119873 + sum120580120598120601119904

(119875119894ℎ119894119897119870

1003816100381610038161003816119909minus119886

119894

1003816100381610038161003816)

(5)














10

100

1000

10000

100000

1000000

10000000

100000000

1000000000

24 32 48 72 82 125 150 175 200 225 250

Opt

imum

thre

shol

d ch

ange

(Log

)

Nodes







0

01

02

03

04

05

06

07

08

09

1

20 24 32 48 72 82 150

Thro

ughp

ut (n

orm

aliz

ed)

Node density

FROMRXORiAODVv2

OLSRv2AAODV








0

01

02

03

04

05

06

07

08

09

1

20 24 32 48 72 82 150

Aver

age d

elay

(s)

Node density

FROMRXORiAODVv2

OLSRv2AAODV



5 Conclusion









References







































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










10

100

1000

10000

100000

1000000

10000000

100000000

1000000000

24 32 48 72 82 125 150 175 200 225 250

Opt

imum

thre

shol

d ch

ange

(Log

)

Nodes







0

01

02

03

04

05

06

07

08

09

1

20 24 32 48 72 82 150

Thro

ughp

ut (n

orm

aliz

ed)

Node density

FROMRXORiAODVv2

OLSRv2AAODV








0

01

02

03

04

05

06

07

08

09

1

20 24 32 48 72 82 150

Aver

age d

elay

(s)

Node density

FROMRXORiAODVv2

OLSRv2AAODV



5 Conclusion









References







































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation










0

01

02

03

04

05

06

07

08

09

1

20 24 32 48 72 82 150

Aver

age d

elay

(s)

Node density

FROMRXORiAODVv2

OLSRv2AAODV



5 Conclusion









References







































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation



































RoboticsJournal of





Journal of



RotatingMachinery





VLSI Design



Shock and Vibration







Journal of



Volume 2014


SensorsJournal of





Propagation









research article protocol independent adaptive route...

Documents