THROUGHPUT ENHANCEMENT USING RESOURCE ALLOCATION AND MASSIVE MIMO TECHNOLOGY IN CELLULAR SYSTEM

Abstract

In thecomingyears,itisanticipatedthattheproliferationofsmartphones,tablets,smart terminals, andemergingapplications,includingmachine-typecommunications,willcause an explosionincellularnetworkdatatraffic.Consequently,futureradioaccessnetworks shall beabletoaccommodatevastconnectivity,diverseusersets,andtheuniquerequire- ments ofapplications.However,inadequateresources,suchasbandwidthandbasestation powerwhicharegenerallylimitedinnature,poseasignificantobstacle.Theunavailability of asufficientnumberofchannelshindersuseraccommodationwithinthecellandincreases powerconsumption.Moreover,massiveconnectivityamonguserscausesnetworkconges- tion withinthecellandleadstocoverageissueswhichaffectthequalityofcommunication and userexperience. In viewofthis,inthisthesis,wefirstdiscusstheresourceallocationissueandproposea channel assignmenttechniquewithauserfairness-basedpowerallocationstrategyforNon- Orthogonal MultipleAccess(NOMA)cellularnetworks.Ourgoalistomaximizethesystem throughput whileensuringthateveryusermaintainsaminimumdataratetosatisfyqualityof service requirementsandsuccessiveinterferencecancellationconstraints.Tosolvethisopti- mization issue,weemploytheexhaustivesearchmethodandCVX2.0MOSEKsolver-based approach forchannelallocationfollowedbypowerallocationacrossassignedchannels.We use thedifferencebetweentwoconvexfunctionsprogrammingtechniqueformodifyingthe nonconvexoptimizationproblemintoaconvexsub-problem;efficientandfairpoweralloca- tion acrosschannelsisachievedbysolvingthisconvexsub-problemiteratively. Wenextdiscusstheusers’coverageissueandconsidermulti-pairtwo-wayhalf-duplex massivemultiple-inputmultiple-output(mMIMO)relayingsystemwithmultipleuserpairs. Most oftheexistingmMIMOrelayinginvestigatespatiallycorrelatedRayleighfadingand only fewworkshaveconsidereduncorrelatedRicianfadingchannels.Theyalsoignorethe impact ofphaseshiftsintheline-of-sightcomponentduetochangesinuserlocationand hardwareeffects.We,therefore,consideramulti-pairmMIMOamplify-and-forwardre- lay systemwithspatiallycorrelatedRicianchannelalongwithphaseshifts.Wederivethe asymptotic closed-formexpressionsforSpectralEfficiency(SE)atthreedifferentpower scaling schemesbyusingacombinationofmaximal-ratio-combiningandmaximal-ratio- transmission (MRC/MRT)techniquesattherelay.Weassumethatthechannelisnotknown at therelay.Accordingly,phase-awareminimummeansquareerrorandphase-unawarelin- ear minimummeansquareerror-basedestimatorsareusedforchannelestimation. Wenextextendourabovehalf-duplexworktothefull-duplexrelayingsystemoverthe same channelscenarioandderivetheasymptoticclosed-formSEexpressionsatdifferent powerscalingschemesusingzero-forcingreceivers.ThenumericalvaluesoftheSEob- tained fromthederivedexpressionshowtheimpactofvaryingpilotpowerandloopinterfer- ence duetofull-duplexprotocol. After analyzingtheasymptoticSEperformances,wefinallyderivetheclosed-formex- pression forSignal-to-Interference-Noise-Ratio(SINR)usingMRC/MRTtechnique.The expression,soderived,isvalidforanyarbitrarynumberofantennas.Toperformavalidity check, wesimulatetheclosed-formandergodicSEexpressionsandshowthevalidityofthe lowerboundSINR(i.e.,closed-formSINR).Moreover,wealsoshedlightontheimpactof loop interferenceinfull-duplexsystems. Thus, themaincontributionofthethesisistheenhancementofspectralefficiencyand energyefficiencythroughappropriateandoptimalresourceallocationwhileprovidingim- provedusercoverage.Wefocusonoptimizingchannelandpowerdistributionindownlink NOMA systems.Inordertoextendcoveragetousers,weproposeamulti-pairamplify-and- forwardrelayschemeusingasharedhalf-duplexmMIMOrelay.Thisisfurtherextendedto a full-duplexsystemunderthesamechannelscenario.