认知无线电网络中一种有效的分布式功率控制(IJITCS-V8-N1-6)

I.J.InformationTechnologyandComputerScience,2016,01,48-53PublishedOnlineJanuary2016inMECS()DOI:10.5815/ijitcs.2016.01.06Copyright©2016MECSI.J.InformationTechnologyandComputerScience,2016,01,48-53AnEfficientDistributedPowerControlinCognitiveRadioNetworksMohammadHosseinFaridi,AliJafari,EnsiehDehghaniDepartmentofComputerEngineeringandInformationTechnologyAmirkabirUniversityofTechnology,Tehran,IranE-mail:[e-mail],[e-mail],[e-mail]—Incognitiveradionetworks,whenprimaryusers(PUs)donotuseasharedfrequencyband,secondaryusers(SUs)areallowedtousethatfrequencyband.However,whenSUusingfrequencybandandreclaimingPUalsowantstousethisband,therearetwosolutions:SUjumpstoanotherexistingspectrumbandorstaysinthesamefrequencybandandchangesitspowersothatinterferenceinthereclaimingPUsdoesnotexceedathreshold.SincethefirstsolutioninterruptstheSUs'work,thesecondsolutionismoreappropriate.Therefore,powercontrolincognitivenetworksisabsolutelysignificantwhichshouldreceivemoreattention.Inthepreviouswork(TPC-CBS)[2],TPCalgorithmhasbeenusedforpowercontrol.ThedrawbackofthisalgorithmisfixedtargetSIRs.Inotherwords,whileusingthisalgorithm,alltheusershavingenoughsourcesorextrasourcesreachthesametargetSIR,whichisnotdesirable.Therefore,inthispaper,wearegoingtosolveitbymeansoftheproposedalgorithmcalledAnEfficientDistributedPowerControlAlgorithmforCognitiveNetworks(EDPC).Bytheproposedalgorithm,whenauserhasabadchannel,itsSIRsetstotheminimumtargetSIRand,whenithasagoodchannel,itsSIRwillbemorethantheminimumvalue.Moreover,bymeansoffuzzylogicsystems(FLS),thevalueofinterferenceonreclaimingPUcausedbySUsischeckedineachiterationandpowerlevelofSUsisdecreasedifthisvalueexceedsathreshold.SimulationresultsshowthatusingourproposedalgorithmnotonlyallowsSUsandPUsusefrequencybandsimultaneously,butalsoenhancesthroughputsignificantlyincomparisonwiththepreviousapproaches(TPC-CBS).IndexTerms—CognitiveRadioNetwork,PowerControl,Interference,Throughput.I.INTRODUCTIONNowadays,cognitiveradiotechnologyisinfluentialintheefficientuseofscareradioresourcesandfrequencyspectrum.Incognitiveradiotechnology,spectrumissharedbetweenlicensedusersandcognitiveradiousersinanopportunisticmanner.Inthistechnology,whenlicenseduserorprimaryusers(PU)donotusethefrequencyband,cognitiveradioorsecondaryusers(SU)areallowedtouseit;but,whileSUusesfrequencybandandPUalsowantstousethisband,therearetwowaysforSU:SUleavesthatbandandjumpstoanotherexistingspectrumband;SUstaysinthesamefrequencyband,butchangesitspowersothatinterferenceonthePUdoesnotexceedathreshold.Obviously,jumpingtoanotherfrequencybandcausesaninterruptionforSU,becauseithastofindanemptyone;therefore,itreducesperformanceforsomeoftheusers.Consideringtheseissues,ifwecouldcontrolSUpowersothatinterferenceonthePUdoesnotcrossathreshold,thesecondsolutionwillbemoreappropriate.Therearetwowaystocontroltransmitpowerincognitivenetworks:distributedpowercontrolandcentralizedpowercontrol.Sinceindistributedpowercontrol,thebasestationneedslesslocalinformationandfeedback,distributedpowercontrolismorepreferablethancentralizedpowercontrol.Oneofthedistributedalgorithmswhichwasproposedin[1]andusedinthepreviouswork(TPC-CBS)[2]isfixed-target-SIRtrackingpowercontrol(TPC).Bymeansofthisalgorithm,whenthesystemisfeasible,alltheusersreachtheirtargetSIRswiththeminimumpower.Accordingto[3],[4],[5],and[6],thisalgorithmconvergestoauniquefixedpoint.TheTPCalgorithmalsohasadrawback;itsfixedtargetSIRs.Inotherwords,usingthisalgorithmwhilethesystemisfeasible,alltheusers,withenoughorextrasources,reachthesametargetSIR,whichisnotdesirable.Forinstance,thisfixedtargetSIRdoesnotlimitsomeoftheserviceslikevoice(duetothecharacteristicsoftheserviceandhumanears);but,forsomeserviceslikedata,reachingSIRbeyondtheleasttargetSIRenhancesthesystemperformance.Hence,weshouldfindanotherdistributedpowercontrolalgorithmtoimprovesystemperformance.Moreover,anotherchallengeinthiscognitivenetworkisrelatedtodetermininguser'spowerinadistributemanner,whichshouldbecontrolledfrequentlysothatSUinterferenceinthePUdoesnotexceedathreshold.Inthispaper,wearegoingtousethealgorithmin[7]whilekeepingthesystemfeasible;notonlyalltheusersreachtheminimumtargetSIR,butalsothoseuserswhohavemoreresourcesreachmoreSIRand,therefore,theAnEfficientDistributedPowerControlinCognitiveRadioNetworks49Copyright©2016MECSI.J.InformationTechnologyandComputerScience,2016,01,48-53systemthroughputimprovesincomparisontothepreviouswork(TPC-CBS).Moreover,bymeansoffuzzylogicsystems(FLS),thisalgorithmcontrolsthepowerofSUstousethefrequencybandwithouttoomuchinterferenceintheprimaryusers.Therestofthispaperisorganizedasfollows:InSection2,thesystemmodelisintroduced.InSection3,first,drawbacksofthepreviousworkarestatedand,then,theproposedmethodforsolvingthesedrawbacksisdiscussed.Simulationresults,relatedworksandconclusionsarepresentedinSections4,5and6respectively.II.SYSTEMMODELInthissection,weconsideranetworkwith4primarycellsand1cognitiveradiocell.Inthisnetwork,1,2,...,KKand1,2,...,MMarethebase