investigation of a new handover approach in lte and wimax

Research ArticleInvestigation of a New Handover Approach in LTE and WiMAX

Mohammad Nour Hindia Ahmed Wasif Reza and Kamarul Ariffin Noordin

Department of Electrical Engineering Faculty of Engineering University of Malaya 50603 Kuala Lumpur Malaysia

Correspondence should be addressed to Ahmed Wasif Reza awreza98yahoocom

Received 4 July 2014 Revised 3 September 2014 Accepted 6 September 2014 Published 14 October 2014

Academic Editor Chun-Wei Tsai

Copyright copy 2014 Mohammad Nour Hindia et al This is an open access article distributed under the Creative CommonsAttribution License which permits unrestricted use distribution and reproduction in any medium provided the original work isproperly cited

Nowadays one of the most important challenges in heterogeneous networks is the connection consistency between the mobilestation and the base stations Furthermore along the roaming process between the mobile station and the base station the systemperformance degrades significantly due to the interferences from neighboring base stations handovers to inaccurate base stationand inappropriate technology selection In this paper several algorithms are proposed to improve mobile station performanceand seamless mobility across the long-term evolution (LTE) and Worldwide Interoperability for Microwave Access (WiMAX)technologies along with a minimum number of redundant handovers Firstly the enhanced global positioning system (GPS) andthe novel received signal strength (RSS) prediction approaches are suggested to predict the target base station accurately Thenthe multiple criteria with two thresholds algorithm is proposed to prioritize the selection between LTE and WiMAX as the targettechnology In addition this study also covers the intercell and cochannel interference reduction by adjusting the frequency reuseratio 3 (FRR3) to work with LTE and WiMAX The obtained results demonstrate high next base station prediction efficiency andhigh accuracy for both horizontal and vertical handovers Moreover the received signal strength is kept at levels higher than thethreshold while maintaining low connection cost and delay within acceptable levels In order to highlight the combination of theproposed algorithmsrsquo performance it is compared with the existing RSS and multiple criteria handover decision algorithms

1 Introduction

With rapid development and deployment of wireless tech-nologies (WiMAX LTE) mobile networks should providefull mobility for all mobile stations simultaneously and at thesame time guarantee the required quality of services Oneof the main challenges of seamless mobility is the availabilityof efficient horizontal and vertical handovers (HHO VHO)The handover which occurs between two networks using thesame technology is called horizontal handover (HHO) forinstance WiMAX-to-WiMAX or LTE-to-LTE handovers [1]whereas the handover occuring between different technolo-gies is called vertical handover (VHO) for instance LTE-to-WiMAX handover or vice versa [2]

Handover process experiences many obstacles such as topredict an accurate target base station (BS) and to select anappropriate technology to connect with Furthermore com-plex calculations are required during the selection processof the target BS and technology from the BSs suggestion list[3] The selection process between technologies should be

accurate and able to satisfy the userrsquos preferences Otherwisethe mobile station (MS) keeps Ping-Ponging betweentechnologies to search for better connection The Ping-Pongeffect causes unnecessary handoff processes and brings someweaknesses including low network throughput long handoffdelay and high dropping probability [4 5] Another issueis the two interferences the intercell interference (ICI) andthe cochannel interference (CCI) from the surrounding BSswhich sharply degrades the received signal strength [6ndash8]In some cases the interferences surpass the acceptable levelsleading to connection loss between the MS and BS Theinterference avoidance scheme is proposed to reduce theinterferences from neighboring BSs by avoiding collisionsbetween similar frequencies used by neighboring BSs Thisgoal can be achieved either in a staticmanner which allocatesdifferent frequencies to neighboring BSs (such as frequencyreuse factor) or in an intelligent way which adjusts thecellsrsquo radius based on the interference level Consideringthe signaling overhead and complexity in implementingthe intelligent technique only the static method is widely

Hindawi Publishing Corporatione Scientific World JournalVolume 2014 Article ID 246206 10 pageshttpdxdoiorg1011552014246206

2 The Scientific World Journal

adopted These issues inherit the connection between theMS and BS such as increasing the data loss rate end-to-enddelay and connection cost

Recently several handover prediction methodologieshave been proposed to predict and select the target BS withproper technology network namely

(1) a history prediction approach that depends on theuserrsquos history and current location firstly the usermobilityrsquos history andhandover ratio from the servingto the target BS are recorded and secondly the predic-tion is performed based on the frequency of previoushandovers between the serving and the target BS [9]Some of the drawbacks of such methods are the factthat they are more suitable for the static nature of MSrsquomobility and the target BS has to be stored in theuserrsquos mobility history

(2) another prediction approach that is based on deter-mining the exactMS locations to predict the target BSso that MS could handover itself that is GPS [10 11]this approach provides an accurate prediction of thetarget BS but is still insufficient for certain cases interms of high cost long time process and high powerconsumption

(3) an important improvement in the handover predic-tion that utilizes the received signal strength forboth target and serving BSs is proposed in [12] thisapproach does not require any parameters exceptthose for the RSSrsquo measurements from the surround-ing BSs Once the RSS of the serving base stationdecreases below a predefined threshold level the MSshifts to the neighboring BS This approach suffersfrom signal attenuation such as fading and shadow-ing that is where the handover prediction probabilitydecreases as the signal attenuation increases

Many algorithms have been proposed for vertical han-dover procedure based on a variety of criteria such as avail-able bandwidth received signal strength signal to inferenceratio (SIR) connection cost handover delay MSrsquo velocitybattery consumption and quality of service (QoS) Forinstance in [13] theMS checks the surrounding BSs and thenhands over to the one which can offer the lowest delay and thehighest bandwidth Therefore it is clear that it is not possibleto make an appropriate handover decision only by evaluatingthose criteria (delay and bandwidth) leading to an increasein the wrong handover prediction ratio whereas in [14] thehandover technique offers a less complex algorithm whilemaintaining a robust VHO decision among heterogeneousnetworks Furthermore this method is based on evaluatingthe multiple criteria received from the neighboring BSs anddetermines the potential target BS These criteria are theavailable bandwidth cost of service received signal strengthexpected time to stay in practical network and powerconsumption Despite the fact that this approach shows highefficiency in low interference environments it cannot workproperly in high interference areas

From the attenuation side several studies have success-fully decreased the interference from the surrounding BSs

and efficiently utilized the available frequency spectrumby avoiding the collisions between the similar operatingfrequenciesThis avoidance is achieved by several approachesand techniques For example in [15] the authors propose anew method to determine a proper frequency operation foreach one of the inner cell outer cell and femtocell whereasin [16] the ICI cancellation is realized using the biorthogonalfrequency division multiple access cellular system along withmultiple angle division reuse scheme Moreover anothertechnique is reported in [17] it suggests a novel dynamicinterference cancellation method based on two levels In thefirst level it determines the coordination of each intercell inthe network then at the second level the central controllerallocates the most appropriate chunk for the user terminalwhich causes no conflict between the terminals Further-more in [18] a method based on the power control andproper reuse of the frequency offers an attractive solution tothe margining problem between the high power node (mainbase station) and low power node (relay node) in the samenetwork

Up till now the most efficient method to eliminate theinterferences in LTE technology is proposed in [6] Theproposed mechanism divides the cell into two regions theinner and outer region and selects the optimal size as wellas the optimal frequency allocation between these regionsHowever this technique seems to be insufficient to cooperatewith other handover approaches due to the setup timeand operation procedures that add much delay comparedto static ones (fractional frequency reuse ratio technique)The FRR3 (frequency reuse ratio 3) exhibits an acceptablelevel of efficiency in terms of interference reduction fromsurrounding BSs while maintaining system simplicity betterthan the one reported in [6] The authors have demonstrateda target BS prediction mechanism and technology selectionmethod based on the userrsquos preference without taking intoconsideration the prediction scenarios and the investigationof the FRR3 technique [19]

In this paper we propose and demonstrate an enhance-ment on the existing target BS prediction algorithms (GPSand RSS) by introducing a virtual trigger threshold Theimplementations of the two enhanced algorithms allow theuser to tradeoff between predictionsrsquo criteria namely costpower consumption and accuracy Furthermore a modifiedmultiple criteria with two thresholds algorithm is suggestedto permit the user to select the target technology (WiMAX orLTE) based on its priorities such as connection cost delayavailable bandwidth and received signal strength Moreoverthe combination of target BS prediction approach technologyselection approach and FRR3 technique is reported in thispaper for the first time

The rest of this paper is organized as follows Section 2describes the system model which consists of the enhancedGPS and novel RSS prediction approaches Moreover theselection procedure of appropriate technology and theoverview of FRR3 with related equation are also discussedand elaborated in detail Section 3 contains a detailed studyof the performance evaluation of the proposed approachesFinally the conclusion and the related future work arepresented in Section 4

The Scientific World Journal 3

Proposed algorithms

Enhanced GPSnovel RSS

MMTT

FRR3 technique

Predict target BS

Select technology satisfyinguserrsquos preferences

Keep RSS at acceptancelevel

Figure 1 The system model

2 System Model

As shown in Figure 1 the proposed algorithms are dividedinto three stages In the first stage based on the userrsquos prefer-ences either enhancedGPS or novel RSS is selected to predictthe target BS In the next stagemodifiedmultiple criteriawithtwo handover thresholds (MMTT) algorithm selects themostappropriate technology (LTE or WiMAX) which satisfies theuserrsquos preferences Finally FRR3 technique decreases the ICIand CCI interferences from the surrounding BSs

21 Prediction Approaches of the Target BS Figure 2 illus-trates the necessity behind applying the prediction approach-es Due to the random movement of the MS its suggestionlist combines 6 possible BSs as a target BS Each BS offerstwo technologies (LTE and WiMAX) thus the MS shouldgo through 12 options as a searching process for the optimalconnection as follows (1 LTE)(1 WiMAX) (6 LTE)(6WiMAX) The introduction of prediction approachessignificantly reduces the suggestion list to a maximum oftwo BSs with four possibilities Therefore at least 60 ofthe search process is reduced out of the calculations Con-sequently a sharp decrease in probability of connection lossprediction time and system complexity is observed In thefollowing subsections two efficient prediction approaches areproposed namely enhanced GPS and novel RSS predictionapproaches

211 GPS Prediction Approach The enhanced GPS predic-tion approach is subject to MSrsquo behavior such as angle ofmovement cross-distance and velocity Once the MSrsquo RSSreaches the trigger threshold level (trigger threshold is theproposed virtual RSS level located at level +10 dBm higherthan handover threshold) the GPS device is activated todetermine the current coordination in the layout of threedimensions (latitude (119909) longitude (119910) and ellipsoid height(119911)) Then the coordination is kept updating up to 119899-timeintervals (119899 is the range between the trigger and handoverthreshold) For each coordination measurement the crosseddistance is calculated as the difference between the currentand previous MSrsquo location during one time interval (1) Also

0

1000

2000

3000

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

34

5

2

71

6

minus3000

minus2000

minus1000

0 1000 2000minus2000 minus1000

y-c

oord

inat

e

x-coordinate

Figure 2 Potential movements of the MS

the angle of movement is determined based on the 119909-axis as0-degree

119863 (119899) = ((119909 (119899) minus 119909 (119899 minus 1))2+ (119910 (119899) minus 119910 (119899 minus 1))

2

+(119911 (119899) minus 119911 (119899 minus 1))2)12

(1)

where 119863(119899) is the MS crossed distance (119909(119899 minus 1) 119910(119899 minus

1) 119911(119899minus1)) is theMSrsquo coordination at 119899minus1 time interval and(119909(119899) 119910(119899) 119911(119899)) is the MSrsquo coordination at 119899-time interval

TheMSrsquo velocity (V) is theMS crossed distance divided bythe required time to cross it (119905) (2)

V =119863 (119899)

119905 (2)

By calculating the movement angle crossed distanceand velocity the next BS is predicted accurately The mainenhancements added to GPS approach are power consump-tion and prediction cost along with high prediction accuracysince the GPS device is activated during 119899-time intervals(between the trigger and handover thresholds) instead ofkeeping the GPS device on all the time

212 RSS Prediction Approach The main objective of thenovel RSS approach is to foretell the target BS in lower costin a simpler and faster way than the existing algorithms Itdoes not require any additional information neither fromthe BS nor from the MS MSrsquo RSS is measured frequently for119899-time intervals (from trigger to handover thresholds) fromall surrounding BSs (3) Then the highest RSS accumulativevalue is set as the targetrsquos BS (4) The robustness of theapproach is that it takes 119899 RSS measurements instead ofone RSS measurement so even if the interferences blurthe prediction decision for a while the target BS keeps


maintaining the highest RSS accumulative value which ishigher than others The RSS is calculated as the following

RSS119894(MS) = PT

119894+ GR + GT

119894minus PL119894minus LT119894minus LR (3)

tarBS119894= MAX

119899

sum119895=1

119873

sum119894=1

RSS119894119895(MS) (4)

where RSS119894(MS) represents the RSS received at the MS from

BS119894 the transmission powers of BS

119894(PT119894) GR andGT

119894are the

antenna gain of bothMS and BS119894 respectively PL

119894is the path

loss betweenBS119894andMS LR andLT

119894are the thermal receivers

noise in both MS and BS119894 respectively tarBS

119894represents the

target base station 119894 119899 is the time interval between the triggerand handover threshold 119895 is the interval time index and119873 isthe total number of BSs

22MMTTApproach to Select the Accurate Target TechnologyThe MMTT approach is proposed to determine the mostappropriate target technology (WiMAX or LTE) which cansatisfy the userrsquos preferences This approach is based on eval-uating many criteria such as RSS connection cost handoverprocess delay and offered bandwidthTheMMTT guaranteesstability of the connection by constantly maintaining theMSrsquo RSS at an acceptable level and minimizing the numberof redundant handovers between technologies due to thenetwork selection which depends on the userrsquos preferencesThe MMTT approach is illustrated as follows

221 Handover RSS Threshold and Triggered RSS ThresholdCalculations While the MS is moving across cells it keepstracking the RSSrsquo serving BS Once it equals the RSS triggerthreshold level the selection process of the most appropriatetarget technology starts If the MSrsquo RSS of technology 119896 atserving BS

119909is less than the handover threshold and the MSrsquo

RSS of technology 119896 at target BS119910is bigger than or equal to

the handover threshold then the RSS condition is satisfied1003816100381610038161003816RSS119898119896119909 lt RSSth119909

1003816100381610038161003816

10038161003816100381610038161003816RSS119898119896119910

ge RSSth11991010038161003816100381610038161003816

(5)

where RSS119898119896119909

is RSS received at the MS119898from the tech-

nology 119896 at service BS119909 and RSS

119898119896119910is RSS received at the

MS119898from the technology 119896 at target BS

119910 while RSSth119909 and

RSSth119910 are the handover thresholds for the service BS119909 andtarget BS

119910 respectively To increase the accuracy of handover

and trigger thresholds for both serving and target BSs a self-learning algorithm is developed as shown below

RSSth119909 =1

119903times

119897

sum119894=1

RSS119894119898119909119896

RSSth119910 =1

119903times

119897

sum119894=1

RSS119894119898119910119896

(6)

where 119903 is the number of previous handover processes thatoccur between the serving BS

119909and target BS

119910 119894 is the index

of the handover event and RSS119894119898119909119896

and RSS119894119898119910119896

are the RSS

Select accurate target technology(1) inputWC WB WD WRSS RSSth 119862th 119889th 119887th(2) Let119873 be the number of base stations(3) Let 119895 be the number of available technologies(4) Let 119894 be the index of base stations(5) Let 119896 be the index of technologies(6) for each 119894 isin 119873 119896 isin 119895 do

(7) if RSS119896gt RSSth

(8) and 119862119896lt 119862th

(9) and 119889119896lt 119889th

(10) and 119887119896gt 119887th

(11) then V(119894 119896) larr (119894 119896)

(12) else remove it from suggestion list(13) end if(14) Calculate (7) for all V(119894 119896)(15) (119894 119896) larr argmax V(119894 119896)(16) end for

Algorithm 1 Evaluation of the BSrsquo technology

measurements at the MS119898from the serving BS

119909and target

BS119910 respectively at handover eventThe self-learning algorithm is triggered after two han-

dover events and then it keeps calculating and updatingthe threshold values of serving and target BSs up to 119899-time intervalsThemain enhancement added by self-learningalgorithm is the determination of themost accurate handoverand trigger threshold values experimentally which helpsto prevent the sudden disconnections especially in a highattenuation area since the threshold values are set to bedynamically adopted with the surrounding area

222 The Technology Selection Process The MS movementdirection of the BSrsquo technologies (WiMAX and LTE) willbe evaluated by Algorithm 1 one of the technologies isselected as the accurate target technology for handover Eachtechnology has to satisfy four conditions to be added to thetarget technology suggestion list as follows

(1) The RSS of evaluating technology (RSS119896) is bigger

than the RSS threshold (RSSth) (Algorithm 1 (line 7))(2) The technology connection cost (119862

119896) is less than the

connection cost threshold (119862th) (line 8)(3) The technology handover process delay (119889

119896) is less

than the delay threshold (119889th) (line 9)(4) The technology offered bandwidth (119887

119896) is bigger than

the bandwidth threshold (119887th) (line 10)

Finally the user will commence handover to the besttechnology 119896 in BS

119894 which can provide the maximum value

of (7) (line 15)

TQE = WC times (1119862

max (1119862)) +WB times ( 119861

max119861)

+WD times (1119863

max (1119863)) +WRSS times ( RSS

maxRSS)

(7)


where TQE is the technology quality evaluation WC WBWD and WRSS are the weights of connection cost availablebandwidth delay and received signal strength respectivelymax (1119862) max 119861 max (1119863) and max RSS are themaximum values of connection cost available bandwidthdelay and received signal strength respectively among thetechnologies in the suggestion list

The technologies added to the suggestion list should bevalidated by these steps Then the highest TQE value will beset as the target technology It is worth mentioning that thesum of the weights is 100

23 FRR3 Technique Themain objective behind utilizing theFRR3 is decreasing the CCI and ICI interferences in orderto significantly improve the received signal strength and thehandoversrsquo stability (VHO and HHO) The FRR3 maintainsmutual interference between MS and surrounding BSs belowa harmful level especially at a high attenuation area Thisguarantees a good performance of theMS-BS connection andthe handover process Moreover FRR3 is simple and easyto apply The FRR3 technique divides WiMAX and LTE BSsinto 3 sectors and 3 hexagons respectively Each BSrsquo sectorand hexagon has its unique operating frequency taking intoconsideration the minimum frequency reuse distance (dm)As a result the ICI and CCI interferences sharply decreasecompared to those without applying the FRR3 techniquesince the interferences come only from the adjacent sectorsand hexagons in the same propagation angle (using the samefrequency operation) instead of from all surrounding BSs asillustrated below [6]

SINR119898119891

=PL119894119898

times 119879119894119891times ℎ119894119898119891

1205902119891+ sum119910

119878=1119868119878119909

times 119879119878119891

times ℎ119878119909119891

(8)

where SINR119898119891

is the signal-to-interference-plus-noise ratioPL119894119898

refers to the path loss associated with the channelbetweenMS

119898and BS

119894119879119894119891

is the transmit power of the BS119894on

subcarrier 119891 ℎ119894119898119891

is the exponentially distributed channelfast-fading power1205902

119891is the noise power of theAdditiveWhite

Gaussian Noise channel 119878 is the BS index and 119910 is thenumber of cochannel BSs

3 Results and Discussion

To clarify the enhancement of the previous proposed algo-rithms prediction and the overall scenario are tested bythe NS-2 simulator and MATLAB program The predictionscenario is validated at the University of Malaya area Thecoordination data have been collected from the GoogleMap database and called to simulate inputs The predictingtarget BS using the enhanced GPS and novel RSS predictionapproaches is presented in Sections 311 and 312 respec-tively and comparedwith the existing RSS approach whereasthe overall scenario is presented in Section 32 and evaluatedin terms of target BS prediction HHO and VHO RSS delayand cost Then it is compared with the existing RSS andmultiple criteria approach

0 1000 2000

0

1000

2000

3000

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

34

2

17

56

minus3000

minus2000

minus1000

minus2000 minus1000

y-c

oord

inat

e

x-coordinate

Figure 3 MS movement at the University of Malaya area

31 Prediction Scenario The two prediction approaches areapplied and tested at the University of Malaya area as shownin Figure 3 The MS is assumed to be roaming from pointA located in BS 1 to point B located in BS 7 These BSs aresupported with LTE and WiMAX technologies

311 Enhanced GPS Approach Results Once the MS reachesthe trigger threshold level of the BS 1 the GPS deviceis activated The GPS prediction approach determines MSrsquocoordination (119909 119910 and 119911) and keeps tracking the coordinatesup to the handover threshold (during 119899-time intervals) asillustrated in Section 211 In Figure 4 the enhancedGPS pre-diction approach proves a high prediction quality comparedto the existing RSS approach [12] Between 38 and 52 secondsthe existing RSS approach shows inaccurate and unstabledetermination of the target BS since high attenuation ofBSrsquo edges lead to fluctuation in RSS level which makes theBS prediction decision unclear and the MS is continuouslyshifted between BS 1 and BS 7 which leads to a high ratioof Ping-Pong effects In terms of the handover accuracy ratio(the proportion of the number of accurate handovers over thetotal number of handovers) the enhanced GPS shows a highlevel of accuracy compared to the existing RSS with only twomissed handovers for the enhanced GPS

312 Novel RSS Prediction Approach Results In order todemonstrate the improvement of the novel RSS predictionapproach it is compared to the existing RSS predictionapproach in [12] (Figure 5) From 45 to 55 seconds the novelRSS remarkably reduces the handover failures compared tothe existing RSS The robustness of the novel RSS is in theway of decisionmaking process which is based on the BSrsquo RSSaccumulative values during 119899-time intervals whereas theexisting RSS is dependent on one RSS measurement whichmakes it more susceptible to signal fluctuating and at thesame time it provides a high accuracy ratio when comparedto the existing RSS The novel RSSrsquos missed handovers areconsidered few despite being 7 when compared to the exist-ing RSS approaches with more than 20 missed handovers


0 10 20 30 40 50 60 70 80 90 10001234567

BS ID

Existing RSSEnhanced GPS

Time (s)

Figure 4 BS prediction using enhanced GPS prediction approach

0 10 20 30 40 50 60 70 80 90 10001234567

BS ID

Existing RSSNovel RSS

Time (s)

Figure 5 BS prediction using RSS prediction approach

As a conclusion from Figures 4 and 5 the enhanced GPSapproach has higher accuracy compared to the novel RSS pre-diction approach On the contrary the RSS approach showsless power consumption no additional data requirementsand lower cost The reason behind proposing these twoprediction approaches is to add flexibility to the predictionprocess based on the userrsquos preferencesThe user tradeoffs areamong accuracy power consumption and cost of prediction

32 Overall Scenario Figure 6 expresses the simulation stepsseries The simulation sequence runs as the following steps

(i) Input parameters are set according to Table 1(ii) The existing RSS multiple criteria (MC) approach

and proposed algorithm (modified multiple criteriaapproach with two handover thresholds) are pre-sented

(iii) FRR3 technique is implemented to all the approaches

All the approaches proposed above have been simulatedtested and evaluated for BS prediction HHO and VHO RSSdelay and cost as simulation outputs The details result andthe decision of each approach are described below

321 Prediction of Target BS and Target Technology In thissection the mechanismrsquos behavior for the prediction andselection of the most appropriate target BS and technology is

Simulation inputs

Sim

ulat

ion

outp

uts

RSS without FRR3MC without FRR3Proposed without

FRR3 BS predictionHHO and VHO

RSS delaycost

Algorithms without FRR3

2nd step

RSS with FRR3MC with FRR3

Proposed with FRR3

Algorithms with FRR3

1st step

3rd step

Figure 6 The simulation steps

0 500 1000 1500

0

500

1000

1500

2000

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

14

6

7

5

Deactivate GPS

Activate GPS

minus500

minus500

minus1000

minus2000 minus1000minus1500

y-c

oord

inat

e

x-coordinate

Figure 7 MS movement in the overall scenario simulation

investigated The topology of Figure 7 consists of 7 BSs eachBS supports two technologies (LTE andWiMAX)The resultsbeing studied are comprised of three approaches with andwithout the FRR3 technique

The studied approaches are named as follows existingRSS approach MC approach and the proposed approachThe weighting factors for each criterion (userrsquos preferences)are applied as inputs to the proposed approach FromFigure 7 we assume that the MS is roaming from BS 4 to BS1 Based on the userrsquos preference (Table 1) the enhanced GPSprediction approach is chosen as the prediction approachsince the cost weight is only 5 as opposed to the novel RSSapproach

A comparison of the results of the enhanced GPSapproach with other competitive approaches regarding targetBS prediction without FRR3 is presented in Figure 8 Duringthe simulation period the enhanced GPS approach provesan accurate BS prediction and high stability compared tothe other two approaches It shows only two handoverfailures (connection to BS 3 instead of BS 1) Since the MS


Table 1 Overall scenario of simulation parameters

Input parameters Units ValuesNumber of cell 7Number MS 1RSS weight 50

Bandwidth weightCost weightDelay weight

20525

Angle of movement Degree 0

Mobility type PRWMM probabilistic randomwaypoint mobility model

Path loss type dB Macro urban path lossPL = 3381 times log 10(fc) minus 794 + 3 + 3504 times log 10(119889)

Simulation time Second 100MS highest Meter 1BS highest Meter 40

0 10 20 30 40 50 60 70 80 90 1000

1

2

3

4

5

6

7

BS ID

MC Enhanced GPS

Existing RSS

Time (s)

Figure 8 BS prediction without FRR3

experiences high ICI and CCI at BS 1 (central BS) and BS3 seems to be having more acceptable factors than BS 4the MC approach connects to BS 3 instead of 4 for quite along time (keep MS without connection between 3 and 43seconds) Therefore this kind of prediction is not convenientor practical for utilization in high attenuation areas Theexisting RSS approach shows average redundant handoversratio between the MC and the enhanced GPS approach

The three approaches show a significant enhancementin the prediction process by applying the FRR3 as can beseen clearly from Figure 9 where the enhancedGPS approachprovides no redundant handovers The MCrsquos efficiency hasincreased remarkably since it is converted to a valid con-nection with few redundant handovers that is from 0ndash60seconds High handovers stability leads to more efficiencyfor exploiting the available resources from both technologies(LTE and WiMAX) That is shown in Figures 10 and 11where the MMTT algorithm shows the highest value ofthe handover quality indicator compared to the other twoapproaches The other approaches show many unwanted

0 10 20 30 40 50 60 70 80 90 1000

1

2

3

4

5

6

7BS

ID

MC Enhanced GPS

Existing RSS

Time (s)

Figure 9 BS prediction with FRR3

handovers and waste available resources since they reserveand use a lot of resources from both technologies for atoo short period inefficiently As it can be observed thedegradation of MSrsquo QoS is a conclusion for both of theexisting RSS and MC

The MMTT approach keeps the signal strength withinan acceptable range during the whole simulation periodexcept at the 57th second (Figure 12) As illustrated inFigures 12 and 13 it is obvious that there is remarkableenhancement between the received signal strength beforeand after applying the FRR3 technique At the 57th secondthe MMTT algorithm is recovered from degradation in thereceived signal strength to an acceptable level At the sametime from 4 to 41 seconds the MC approach demonstrates ahuge enhancement since it is converted from insufficient tosufficient handover approach

Even theMMTTapproach has the highest handover delaycompared to the other two approaches but it still satisfiesthe userrsquos preferences (Figure 14) The FRR3 decreases the


0 10 20 30 40 50 60 70 80 90 1000

1

2

LTE

WiMAX

Tech

nolo

gy ID

MC MMTT

Existing RSS

Time (s)

Figure 10 HHO and VHO without FRR3

0 10 20 30 40 50 60 70 80 90 1000

1

2

LTE

WiMAX

Tech

nolo

gy ID

MC MMTT

Existing RSS

Time (s)

Figure 11 HHO and VHO with FRR3

0 10 20 30 40 50 60 70 80 90 100

0

Existing RSSMC MMTT

RSS

(dBm

)

minus100

minus150

minus50

Time (s)

Figure 12 Received signal strength without using FRR3

0 10 20 30 40 50 60 70 80 90 100

0

Existing RSSMC MMTT

RSS

(dBm

)

minus100

minus150

minus50

Time (s)

Figure 13 Received signal strength with FRR3

0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

Time (s)

Del

ay (m

s)

Existing RSSMC MMTT

Figure 14 Delay without FRR3

handover process delay of theMMTT approach up to roughly15 in comparison to the scenario with no FRR3 (Figure 15)This means that the proposed algorithm determines thetarget BS and technology faster than without using the FRR3technique

From Figure 16 it is surmised that the proposed algo-rithm achieves the lowest average cost value equal to 0883while MC results in 08915 and the existing RSS has thehighest value of 09

4 Conclusions

In this paper the enhancement of GPS and RSS algorithms byadding the virtual threshold (trigger threshold) is presentedIn the RSS algorithm a novel ldquo119899rdquo number of measurementsmethod is used instead of the standard one measurement


0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

Time (s)

Del

ay (m

s)

Existing RSSMC MMTT

Figure 15 Delay with FRR3

0 10 20 30 40 50 60 70 80 90 1000

02

04

06

08

1

12

14

16

18

2

Time (s)

Cos

t (un

it)

Existing RSSMC MMTT

Figure 16 Connection cost

techniqueTherefore the RSS approach becomes less affectedby the interferences from the surrounding BSs whereas inthe GPS algorithm the enhancement decreases the cost andpower consumption due to the fact that theGPS device is onlyactive between the trigger and handover thresholds instead ofall the time

Moreover we demonstrate a modified multiple criteriawith two thresholds algorithm resulting in an increase inthe efficiency of target network selection The selection isbased on the user preferences since it uses the self-learningalgorithm to determine the trigger and handover thresholdsdynamically Finally by adding the FRR3 technique to thesystem the efficiency of the prediction of target BS and theselection of target technology is increased and the delayis decreased by approximately 15 As far as we know

this is the first investigation of a setup combining targetBS prediction approach technology selection approach andFRR3 technique

A further enhancement may be added to this work suchas a generic and extensible media access control layer (MAC)for the networks This enhancement will allow the MS tohave a smoother transmission and an ability to receive dataamong different BSs with different network types Since theMSwill be supportedwith smooth connectivity with differentBSs technologies and not requiring any extra equipment thesystem is expected to have less delay lower cost and betterperformance

Conflict of Interests

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

Acknowledgments

This research work is supported by the Science fund Grant(Grant no SF002-2014) and University of Malaya ResearchGrant (UMRG) scheme (RG286-14AFR)

References

[1] D Cheelu M R Babu and P V Krishna ldquoA study of verticalhandoffdecision strategies in heterogeneouswireless networksrdquoInternational Journal of Engineering and Technology vol 5 no3 pp 2541ndash2554 2013

[2] M Khanand and K Han ldquoAn optimized network selection andhandover triggering scheme for heterogeneous self-organizedwireless networksrdquo Mathematical Problems in Engineering vol2014 Article ID 173068 11 pages 2014

[3] X Yan Y A Sekercioglu and S Narayanan ldquoA survey ofvertical handover decision algorithms in Fourth Generationheterogeneous wireless networksrdquo Computer Networks vol 54no 11 pp 1848ndash1863 2010

[4] S B Johnson P Saranya Nath and T Velmurugan ldquoAn opti-mized algorithm for vertical handoff in heterogeneous wirelessnetworksrdquo in Proceedings of the IEEE Conference on Informationand Communication Technologies (ICT rsquo13) pp 1206ndash1210 Jeju-do Republic of Korea April 2013

[5] E Kim S Kim and L Choonhwa ldquoSupporting seamlessmobility for P2P live streamingrdquo The Scientific World Journalvol 2014 Article ID 134391 8 pages 2014

[6] D Bilios C Bouras V Kokkinos A Papazois and G TseliouldquoSelecting the optimal fractional frequency reuse scheme inlong term evolution networksrdquo Wireless Personal Communica-tions vol 71 no 4 pp 2693ndash2712 2013

[7] B Choi S Lim and T-J Lee ldquoSequential frequency reuse withpower control for OFDMA systemsrdquoWireless Communicationsand Mobile Computing vol 13 no 1 pp 37ndash46 2013

[8] C Kosta B Hunt A U Quddus and R Tafazolli ldquoOn inter-ference avoidance through inter-cell interference coordination(ICIC) based on OFDMA mobile systemsrdquo IEEE Communica-tions Surveys and Tutorials vol 15 no 3 pp 973ndash995 2013

[9] Z Becvar ldquoEfficiency of handover prediction based on han-dover historyrdquo Journal of Convergence Information Technologyvol 4 no 4 pp 41ndash47 2009


[10] S Michaelis N Piatkowski and K Morik ldquoPredicting nextnetwork cell IDs for moving users with discriminative and gen-erative modelsrdquo in Proceedings of the Mobile Data Challenge byNokia Workshop in Conjunction with International Conferenceon Pervasive Computing 2012

[11] S V Saboji and C B Akki ldquoA client-based vertical handoffsystem in 4G wireless networksrdquo Journal of Advances in Infor-mation Technology vol 1 no 4 pp 197ndash203 2010

[12] Z Becvar P MacH and B Simak ldquoImprovement of handoverprediction in mobile WiMAX by using two thresholdsrdquo Com-puter Networks vol 55 no 16 pp 3759ndash3773 2011

[13] R RathiyaAAnitha and J Jayakumari ldquoEfficientQoSorientedvertical handoff scheme in the integration of WiMAXWLANnetworksrdquo in Proceedings of the IEEE Conference on Informationand Communication Technologies (ICT rsquo13) pp 378ndash381 April2013

[14] DHe C Chi S Chan C Chen J Bu andM Yin ldquoA simple androbust vertical handoff algorithm for heterogeneous wirelessmobile networksrdquo Wireless Personal Communications vol 59no 2 pp 361ndash373 2011

[15] C-Y Oh M Y Chung H Choo and T-J Lee ldquoResourceallocation with partitioning criterion for macro-femto overlaycellular networks with fractional frequency reuserdquo WirelessPersonal Communications vol 68 no 2 pp 417ndash432 2013

[16] H Wang ldquoBiorthogonal frequency division multiple accesscellular system with angle division reuse schemerdquo WirelessPersonal Communications vol 70 no 4 pp 1553ndash1573 2013

[17] M Rahman and H Yanikomeroglu ldquoEnhancing cell-edge per-formance a downlink dynamic interference avoidance schemewith inter-cell coordinationrdquo IEEE Transactions on WirelessCommunications vol 9 no 4 pp 1414ndash1425 2010

[18] Q C Li R Q Hu Y Xu and Y Qian ldquoOptimal fractional fre-quency reuse and power control in the heterogeneous wirelessnetworksrdquo IEEE Transactions on Wireless Communications vol12 no 6 pp 2658ndash2668 2013

[19] MNHindia AWReza KNNoordin andA SM Z KausarldquoEnhancement the handovers accuracy and performance ofWiMAX and LTE networksrdquo in Proceedings of the InternationalConference on Computer Science and ComputationalMathemat-ics (ICCSCM rsquo14) 2014


adopted These issues inherit the connection between theMS and BS such as increasing the data loss rate end-to-enddelay and connection cost

Recently several handover prediction methodologieshave been proposed to predict and select the target BS withproper technology network namely

(1) a history prediction approach that depends on theuserrsquos history and current location firstly the usermobilityrsquos history andhandover ratio from the servingto the target BS are recorded and secondly the predic-tion is performed based on the frequency of previoushandovers between the serving and the target BS [9]Some of the drawbacks of such methods are the factthat they are more suitable for the static nature of MSrsquomobility and the target BS has to be stored in theuserrsquos mobility history

(2) another prediction approach that is based on deter-mining the exactMS locations to predict the target BSso that MS could handover itself that is GPS [10 11]this approach provides an accurate prediction of thetarget BS but is still insufficient for certain cases interms of high cost long time process and high powerconsumption

(3) an important improvement in the handover predic-tion that utilizes the received signal strength forboth target and serving BSs is proposed in [12] thisapproach does not require any parameters exceptthose for the RSSrsquo measurements from the surround-ing BSs Once the RSS of the serving base stationdecreases below a predefined threshold level the MSshifts to the neighboring BS This approach suffersfrom signal attenuation such as fading and shadow-ing that is where the handover prediction probabilitydecreases as the signal attenuation increases

Many algorithms have been proposed for vertical han-dover procedure based on a variety of criteria such as avail-able bandwidth received signal strength signal to inferenceratio (SIR) connection cost handover delay MSrsquo velocitybattery consumption and quality of service (QoS) Forinstance in [13] theMS checks the surrounding BSs and thenhands over to the one which can offer the lowest delay and thehighest bandwidth Therefore it is clear that it is not possibleto make an appropriate handover decision only by evaluatingthose criteria (delay and bandwidth) leading to an increasein the wrong handover prediction ratio whereas in [14] thehandover technique offers a less complex algorithm whilemaintaining a robust VHO decision among heterogeneousnetworks Furthermore this method is based on evaluatingthe multiple criteria received from the neighboring BSs anddetermines the potential target BS These criteria are theavailable bandwidth cost of service received signal strengthexpected time to stay in practical network and powerconsumption Despite the fact that this approach shows highefficiency in low interference environments it cannot workproperly in high interference areas

From the attenuation side several studies have success-fully decreased the interference from the surrounding BSs

and efficiently utilized the available frequency spectrumby avoiding the collisions between the similar operatingfrequenciesThis avoidance is achieved by several approachesand techniques For example in [15] the authors propose anew method to determine a proper frequency operation foreach one of the inner cell outer cell and femtocell whereasin [16] the ICI cancellation is realized using the biorthogonalfrequency division multiple access cellular system along withmultiple angle division reuse scheme Moreover anothertechnique is reported in [17] it suggests a novel dynamicinterference cancellation method based on two levels In thefirst level it determines the coordination of each intercell inthe network then at the second level the central controllerallocates the most appropriate chunk for the user terminalwhich causes no conflict between the terminals Further-more in [18] a method based on the power control andproper reuse of the frequency offers an attractive solution tothe margining problem between the high power node (mainbase station) and low power node (relay node) in the samenetwork

Up till now the most efficient method to eliminate theinterferences in LTE technology is proposed in [6] Theproposed mechanism divides the cell into two regions theinner and outer region and selects the optimal size as wellas the optimal frequency allocation between these regionsHowever this technique seems to be insufficient to cooperatewith other handover approaches due to the setup timeand operation procedures that add much delay comparedto static ones (fractional frequency reuse ratio technique)The FRR3 (frequency reuse ratio 3) exhibits an acceptablelevel of efficiency in terms of interference reduction fromsurrounding BSs while maintaining system simplicity betterthan the one reported in [6] The authors have demonstrateda target BS prediction mechanism and technology selectionmethod based on the userrsquos preference without taking intoconsideration the prediction scenarios and the investigationof the FRR3 technique [19]

In this paper we propose and demonstrate an enhance-ment on the existing target BS prediction algorithms (GPSand RSS) by introducing a virtual trigger threshold Theimplementations of the two enhanced algorithms allow theuser to tradeoff between predictionsrsquo criteria namely costpower consumption and accuracy Furthermore a modifiedmultiple criteria with two thresholds algorithm is suggestedto permit the user to select the target technology (WiMAX orLTE) based on its priorities such as connection cost delayavailable bandwidth and received signal strength Moreoverthe combination of target BS prediction approach technologyselection approach and FRR3 technique is reported in thispaper for the first time

The rest of this paper is organized as follows Section 2describes the system model which consists of the enhancedGPS and novel RSS prediction approaches Moreover theselection procedure of appropriate technology and theoverview of FRR3 with related equation are also discussedand elaborated in detail Section 3 contains a detailed studyof the performance evaluation of the proposed approachesFinally the conclusion and the related future work arepresented in Section 4


Proposed algorithms


MMTT

FRR3 technique

Predict target BS




2 System Model




0

1000

2000

3000

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

34

5

2

71

6

minus3000

minus2000

minus1000

0 1000 2000minus2000 minus1000

y-c

oord

inat

e

x-coordinate




2

+(119911 (119899) minus 119911 (119899 minus 1))2)12

(1)




V =119863 (119899)

119905 (2)





RSS119894(MS) = PT

119894+ GR + GT


tarBS119894= MAX

119899

sum119895=1

119873

sum119894=1

RSS119894119895(MS) (4)



119894(PT119894) GR andGT

119894are the


119894is the path











1003816100381610038161003816

10038161003816100381610038161003816RSS119898119896119910

ge RSSth11991010038161003816100381610038161003816

(5)

where RSS119898119896119909









RSSth119909 =1

119903times

119897

sum119894=1

RSS119894119898119909119896

RSSth119910 =1

119903times

119897

sum119894=1

RSS119894119898119910119896

(6)


119909and target BS

119910 119894 is the index


and RSS119894119898119910119896

are the RSS



(8) and 119862119896lt 119862th

(9) and 119889119896lt 119889th

(10) and 119887119896gt 119887th

(11) then V(119894 119896) larr (119894 119896)




119909and target








119896) is less






of (7) (line 15)


max (1119862)) +WB times ( 119861

max119861)

+WD times (1119863


maxRSS)

(7)





SINR119898119891

=PL119894119898

times 119879119894119891times ℎ119894119898119891

1205902119891+ sum119910

119878=1119868119878119909

times 119879119878119891

times ℎ119878119909119891

(8)




119898and BS

119894119879119894119891


subcarrier 119891 ℎ119894119898119891






0 1000 2000

0

1000

2000

3000

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

34

2

17

56

minus3000

minus2000

minus1000

minus2000 minus1000

y-c

oord

inat

e

x-coordinate






0 10 20 30 40 50 60 70 80 90 10001234567

BS ID


Time (s)


0 10 20 30 40 50 60 70 80 90 10001234567

BS ID


Time (s)









Simulation inputs

Sim

ulat

ion

outp

uts



RSS delaycost


2nd step


Proposed with FRR3


1st step

3rd step


0 500 1000 1500

0

500

1000

1500

2000

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

14

6

7

5

Deactivate GPS

Activate GPS

minus500

minus500

minus1000


y-c

oord

inat

e

x-coordinate









20525





0 10 20 30 40 50 60 70 80 90 1000

1

2

3

4

5

6

7

BS ID

MC Enhanced GPS

Existing RSS

Time (s)




0 10 20 30 40 50 60 70 80 90 1000

1

2

3

4

5

6

7BS

ID

MC Enhanced GPS

Existing RSS

Time (s)






0 10 20 30 40 50 60 70 80 90 1000

1

2

LTE

WiMAX

Tech

nolo

gy ID

MC MMTT

Existing RSS

Time (s)


0 10 20 30 40 50 60 70 80 90 1000

1

2

LTE

WiMAX

Tech

nolo

gy ID

MC MMTT

Existing RSS

Time (s)


0 10 20 30 40 50 60 70 80 90 100

0

Existing RSSMC MMTT

RSS

(dBm

)

minus100

minus150

minus50

Time (s)


0 10 20 30 40 50 60 70 80 90 100

0

Existing RSSMC MMTT

RSS

(dBm

)

minus100

minus150

minus50

Time (s)


0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

Time (s)

Del

ay (m

s)

Existing RSSMC MMTT




4 Conclusions



0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

Time (s)

Del

ay (m

s)

Existing RSSMC MMTT


0 10 20 30 40 50 60 70 80 90 1000

02

04

06

08

1

12

14

16

18

2

Time (s)

Cos

t (un

it)

Existing RSSMC MMTT








Acknowledgments


References






















Proposed algorithms


MMTT

FRR3 technique

Predict target BS




2 System Model




0

1000

2000

3000

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

34

5

2

71

6

minus3000

minus2000

minus1000

0 1000 2000minus2000 minus1000

y-c

oord

inat

e

x-coordinate




2

+(119911 (119899) minus 119911 (119899 minus 1))2)12

(1)




V =119863 (119899)

119905 (2)





RSS119894(MS) = PT

119894+ GR + GT


tarBS119894= MAX

119899

sum119895=1

119873

sum119894=1

RSS119894119895(MS) (4)



119894(PT119894) GR andGT

119894are the


119894is the path











1003816100381610038161003816

10038161003816100381610038161003816RSS119898119896119910

ge RSSth11991010038161003816100381610038161003816

(5)

where RSS119898119896119909









RSSth119909 =1

119903times

119897

sum119894=1

RSS119894119898119909119896

RSSth119910 =1

119903times

119897

sum119894=1

RSS119894119898119910119896

(6)


119909and target BS

119910 119894 is the index


and RSS119894119898119910119896

are the RSS



(8) and 119862119896lt 119862th

(9) and 119889119896lt 119889th

(10) and 119887119896gt 119887th

(11) then V(119894 119896) larr (119894 119896)




119909and target








119896) is less






of (7) (line 15)


max (1119862)) +WB times ( 119861

max119861)

+WD times (1119863


maxRSS)

(7)





SINR119898119891

=PL119894119898

times 119879119894119891times ℎ119894119898119891

1205902119891+ sum119910

119878=1119868119878119909

times 119879119878119891

times ℎ119878119909119891

(8)




119898and BS

119894119879119894119891


subcarrier 119891 ℎ119894119898119891






0 1000 2000

0

1000

2000

3000

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

34

2

17

56

minus3000

minus2000

minus1000

minus2000 minus1000

y-c

oord

inat

e

x-coordinate






0 10 20 30 40 50 60 70 80 90 10001234567

BS ID


Time (s)


0 10 20 30 40 50 60 70 80 90 10001234567

BS ID


Time (s)









Simulation inputs

Sim

ulat

ion

outp

uts



RSS delaycost


2nd step


Proposed with FRR3


1st step

3rd step


0 500 1000 1500

0

500

1000

1500

2000

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

14

6

7

5

Deactivate GPS

Activate GPS

minus500

minus500

minus1000


y-c

oord

inat

e

x-coordinate









20525





0 10 20 30 40 50 60 70 80 90 1000

1

2

3

4

5

6

7

BS ID

MC Enhanced GPS

Existing RSS

Time (s)




0 10 20 30 40 50 60 70 80 90 1000

1

2

3

4

5

6

7BS

ID

MC Enhanced GPS

Existing RSS

Time (s)






0 10 20 30 40 50 60 70 80 90 1000

1

2

LTE

WiMAX

Tech

nolo

gy ID

MC MMTT

Existing RSS

Time (s)


0 10 20 30 40 50 60 70 80 90 1000

1

2

LTE

WiMAX

Tech

nolo

gy ID

MC MMTT

Existing RSS

Time (s)


0 10 20 30 40 50 60 70 80 90 100

0

Existing RSSMC MMTT

RSS

(dBm

)

minus100

minus150

minus50

Time (s)


0 10 20 30 40 50 60 70 80 90 100

0

Existing RSSMC MMTT

RSS

(dBm

)

minus100

minus150

minus50

Time (s)


0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

Time (s)

Del

ay (m

s)

Existing RSSMC MMTT




4 Conclusions



0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

Time (s)

Del

ay (m

s)

Existing RSSMC MMTT


0 10 20 30 40 50 60 70 80 90 1000

02

04

06

08

1

12

14

16

18

2

Time (s)

Cos

t (un

it)

Existing RSSMC MMTT








Acknowledgments


References























RSS119894(MS) = PT

119894+ GR + GT


tarBS119894= MAX

119899

sum119895=1

119873

sum119894=1

RSS119894119895(MS) (4)



119894(PT119894) GR andGT

119894are the


119894is the path











1003816100381610038161003816

10038161003816100381610038161003816RSS119898119896119910

ge RSSth11991010038161003816100381610038161003816

(5)

where RSS119898119896119909









RSSth119909 =1

119903times

119897

sum119894=1

RSS119894119898119909119896

RSSth119910 =1

119903times

119897

sum119894=1

RSS119894119898119910119896

(6)


119909and target BS

119910 119894 is the index


and RSS119894119898119910119896

are the RSS



(8) and 119862119896lt 119862th

(9) and 119889119896lt 119889th

(10) and 119887119896gt 119887th

(11) then V(119894 119896) larr (119894 119896)




119909and target








119896) is less






of (7) (line 15)


max (1119862)) +WB times ( 119861

max119861)

+WD times (1119863


maxRSS)

(7)





SINR119898119891

=PL119894119898

times 119879119894119891times ℎ119894119898119891

1205902119891+ sum119910

119878=1119868119878119909

times 119879119878119891

times ℎ119878119909119891

(8)




119898and BS

119894119879119894119891


subcarrier 119891 ℎ119894119898119891






0 1000 2000

0

1000

2000

3000

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

34

2

17

56

minus3000

minus2000

minus1000

minus2000 minus1000

y-c

oord

inat

e

x-coordinate






0 10 20 30 40 50 60 70 80 90 10001234567

BS ID


Time (s)


0 10 20 30 40 50 60 70 80 90 10001234567

BS ID


Time (s)









Simulation inputs

Sim

ulat

ion

outp

uts



RSS delaycost


2nd step


Proposed with FRR3


1st step

3rd step


0 500 1000 1500

0

500

1000

1500

2000

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

14

6

7

5

Deactivate GPS

Activate GPS

minus500

minus500

minus1000


y-c

oord

inat

e

x-coordinate









20525





0 10 20 30 40 50 60 70 80 90 1000

1

2

3

4

5

6

7

BS ID

MC Enhanced GPS

Existing RSS

Time (s)




0 10 20 30 40 50 60 70 80 90 1000

1

2

3

4

5

6

7BS

ID

MC Enhanced GPS

Existing RSS

Time (s)






0 10 20 30 40 50 60 70 80 90 1000

1

2

LTE

WiMAX

Tech

nolo

gy ID

MC MMTT

Existing RSS

Time (s)


0 10 20 30 40 50 60 70 80 90 1000

1

2

LTE

WiMAX

Tech

nolo

gy ID

MC MMTT

Existing RSS

Time (s)


0 10 20 30 40 50 60 70 80 90 100

0

Existing RSSMC MMTT

RSS

(dBm

)

minus100

minus150

minus50

Time (s)


0 10 20 30 40 50 60 70 80 90 100

0

Existing RSSMC MMTT

RSS

(dBm

)

minus100

minus150

minus50

Time (s)


0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

Time (s)

Del

ay (m

s)

Existing RSSMC MMTT




4 Conclusions



0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

Time (s)

Del

ay (m

s)

Existing RSSMC MMTT


0 10 20 30 40 50 60 70 80 90 1000

02

04

06

08

1

12

14

16

18

2

Time (s)

Cos

t (un

it)

Existing RSSMC MMTT








Acknowledgments


References

























SINR119898119891

=PL119894119898

times 119879119894119891times ℎ119894119898119891

1205902119891+ sum119910

119878=1119868119878119909

times 119879119878119891

times ℎ119878119909119891

(8)




119898and BS

119894119879119894119891


subcarrier 119891 ℎ119894119898119891






0 1000 2000

0

1000

2000

3000

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

34

2

17

56

minus3000

minus2000

minus1000

minus2000 minus1000

y-c

oord

inat

e

x-coordinate






0 10 20 30 40 50 60 70 80 90 10001234567

BS ID


Time (s)


0 10 20 30 40 50 60 70 80 90 10001234567

BS ID


Time (s)









Simulation inputs

Sim

ulat

ion

outp

uts



RSS delaycost


2nd step


Proposed with FRR3


1st step

3rd step


0 500 1000 1500

0

500

1000

1500

2000

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

14

6

7

5

Deactivate GPS

Activate GPS

minus500

minus500

minus1000


y-c

oord

inat

e

x-coordinate









20525





0 10 20 30 40 50 60 70 80 90 1000

1

2

3

4

5

6

7

BS ID

MC Enhanced GPS

Existing RSS

Time (s)




0 10 20 30 40 50 60 70 80 90 1000

1

2

3

4

5

6

7BS

ID

MC Enhanced GPS

Existing RSS

Time (s)






0 10 20 30 40 50 60 70 80 90 1000

1

2

LTE

WiMAX

Tech

nolo

gy ID

MC MMTT

Existing RSS

Time (s)


0 10 20 30 40 50 60 70 80 90 1000

1

2

LTE

WiMAX

Tech

nolo

gy ID

MC MMTT

Existing RSS

Time (s)


0 10 20 30 40 50 60 70 80 90 100

0

Existing RSSMC MMTT

RSS

(dBm

)

minus100

minus150

minus50

Time (s)


0 10 20 30 40 50 60 70 80 90 100

0

Existing RSSMC MMTT

RSS

(dBm

)

minus100

minus150

minus50

Time (s)


0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

Time (s)

Del

ay (m

s)

Existing RSSMC MMTT




4 Conclusions



0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

Time (s)

Del

ay (m

s)

Existing RSSMC MMTT


0 10 20 30 40 50 60 70 80 90 1000

02

04

06

08

1

12

14

16

18

2

Time (s)

Cos

t (un

it)

Existing RSSMC MMTT








Acknowledgments


References






















0 10 20 30 40 50 60 70 80 90 10001234567

BS ID


Time (s)


0 10 20 30 40 50 60 70 80 90 10001234567

BS ID


Time (s)









Simulation inputs

Sim

ulat

ion

outp

uts



RSS delaycost


2nd step


Proposed with FRR3


1st step

3rd step


0 500 1000 1500

0

500

1000

1500

2000

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

2

3

14

6

7

5

Deactivate GPS

Activate GPS

minus500

minus500

minus1000


y-c

oord

inat

e

x-coordinate









20525





0 10 20 30 40 50 60 70 80 90 1000

1

2

3

4

5

6

7

BS ID

MC Enhanced GPS

Existing RSS

Time (s)




0 10 20 30 40 50 60 70 80 90 1000

1

2

3

4

5

6

7BS

ID

MC Enhanced GPS

Existing RSS

Time (s)






0 10 20 30 40 50 60 70 80 90 1000

1

2

LTE

WiMAX

Tech

nolo

gy ID

MC MMTT

Existing RSS

Time (s)


0 10 20 30 40 50 60 70 80 90 1000

1

2

LTE

WiMAX

Tech

nolo

gy ID

MC MMTT

Existing RSS

Time (s)


0 10 20 30 40 50 60 70 80 90 100

0

Existing RSSMC MMTT

RSS

(dBm

)

minus100

minus150

minus50

Time (s)


0 10 20 30 40 50 60 70 80 90 100

0

Existing RSSMC MMTT

RSS

(dBm

)

minus100

minus150

minus50

Time (s)


0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

Time (s)

Del

ay (m

s)

Existing RSSMC MMTT




4 Conclusions



0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

Time (s)

Del

ay (m

s)

Existing RSSMC MMTT


0 10 20 30 40 50 60 70 80 90 1000

02

04

06

08

1

12

14

16

18

2

Time (s)

Cos

t (un

it)

Existing RSSMC MMTT








Acknowledgments


References

























20525





0 10 20 30 40 50 60 70 80 90 1000

1

2

3

4

5

6

7

BS ID

MC Enhanced GPS

Existing RSS

Time (s)




0 10 20 30 40 50 60 70 80 90 1000

1

2

3

4

5

6

7BS

ID

MC Enhanced GPS

Existing RSS

Time (s)






0 10 20 30 40 50 60 70 80 90 1000

1

2

LTE

WiMAX

Tech

nolo

gy ID

MC MMTT

Existing RSS

Time (s)


0 10 20 30 40 50 60 70 80 90 1000

1

2

LTE

WiMAX

Tech

nolo

gy ID

MC MMTT

Existing RSS

Time (s)


0 10 20 30 40 50 60 70 80 90 100

0

Existing RSSMC MMTT

RSS

(dBm

)

minus100

minus150

minus50

Time (s)


0 10 20 30 40 50 60 70 80 90 100

0

Existing RSSMC MMTT

RSS

(dBm

)

minus100

minus150

minus50

Time (s)


0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

Time (s)

Del

ay (m

s)

Existing RSSMC MMTT




4 Conclusions



0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

Time (s)

Del

ay (m

s)

Existing RSSMC MMTT


0 10 20 30 40 50 60 70 80 90 1000

02

04

06

08

1

12

14

16

18

2

Time (s)

Cos

t (un

it)

Existing RSSMC MMTT








Acknowledgments


References






















0 10 20 30 40 50 60 70 80 90 1000

1

2

LTE

WiMAX

Tech

nolo

gy ID

MC MMTT

Existing RSS

Time (s)


0 10 20 30 40 50 60 70 80 90 1000

1

2

LTE

WiMAX

Tech

nolo

gy ID

MC MMTT

Existing RSS

Time (s)


0 10 20 30 40 50 60 70 80 90 100

0

Existing RSSMC MMTT

RSS

(dBm

)

minus100

minus150

minus50

Time (s)


0 10 20 30 40 50 60 70 80 90 100

0

Existing RSSMC MMTT

RSS

(dBm

)

minus100

minus150

minus50

Time (s)


0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

Time (s)

Del

ay (m

s)

Existing RSSMC MMTT




4 Conclusions



0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

Time (s)

Del

ay (m

s)

Existing RSSMC MMTT


0 10 20 30 40 50 60 70 80 90 1000

02

04

06

08

1

12

14

16

18

2

Time (s)

Cos

t (un

it)

Existing RSSMC MMTT








Acknowledgments


References






















0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

Time (s)

Del

ay (m

s)

Existing RSSMC MMTT


0 10 20 30 40 50 60 70 80 90 1000

02

04

06

08

1

12

14

16

18

2

Time (s)

Cos

t (un

it)

Existing RSSMC MMTT








Acknowledgments


References





















investigation of a new handover approach in lte and wimax

Documents