3848IEEETRANSACTIONSONWIRELESSCOMMUNICATIONS,VOL.9,NO.12,DECEMBER2010
TemporalSpectrumSharing
BasedonPrimaryUserActivityPrediction
KiWonSung,Member,IEEE,Seong-LyunKim,Member,IEEE,andJensZander,Member,IEEE
Abstract—Inthispaperweinvestigatetheopportunisticspec-trumaccessintemporaldomainwhereasecondaryusersharesaradiochannelwithaprimaryuserduringtheOFFperiodoftheprimaryuser.WeconsiderpracticalON/OFFtraf cmodelswhoseburstynaturesarenotproperlydescribedbyaMarkovianassumption.Anoptimalstrategytodeterminethetransmissionpowerofthesecondaryuserisproposed,whichcanbeadaptedtoanysourcetraf cmodeloftheprimaryuser.Thisstrategywillmaximizethespectrumutilizationofthesecondaryuserwhilekeepinginterferenceviolationstotheprimaryuserbelowathreshold.Numericalresultsshowthatthetransmissionpowerofthesecondaryuserdependsontheprobabilitydistributionoftheprimarytraf caswellastheelapsedtimeoftheOFFperiod.IndexTerms—Opportunisticspectrumaccess,temporalspec-trumsharing,transmissionpowercontrol,traf cmodel.
I.INTRODUCTION
ADIOspectrumhasbecomeascarceresourcewiththerapidgrowthofthedemandforwirelesscommunica-tions.Theneedforthespectrumresourceisexpectedtoincreasemoreaspervasivewirelessservicesandapplicationsarelikelytoprevailinthenearfuture.Incontrasttotheapparentshortageofspectrum,measurementresultsforthespectrumusagesuggestthatonlyafractionofthespectrumisactuallyusedatanygiventimeandlocation[1],[2].Thisdiscrepancybetweenthespectrumdemandandtheactualutilizationnecessitatesmoreef cientmethodsofexploitingthespectrum.Onepromisingsolutionistheopportunisticspectrumaccess(OSA)thatisenvisionedbythebroaderconceptofcognitiveradio[3],[4].TheOSAallowssecondary(cognitive)userstoshareradiochannelswithprimary(legacy)usersprovidedthattheinterferencecausedbythesecondaryusersisnotintrusivetotheprimaryusers.
Thespectrumopportunitycanbediscoveredinspatialandtemporaldomains.ThespatialaspectofOSAreliesontheattenuationofradiosignalwhichenablesanenoughseparationbetweenasecondaryuserandaprimaryusertopreventharmfulinterference.ThecapacityofthespatialOSAhasbeenextensivelystudied.Thecapacitylimitofthespectrumsharing
ManuscriptreceivedMarch19,2010;revisedAugust10,2010;acceptedSeptember12,2010.TheassociateeditorcoordinatingthereviewofthispaperandapprovingitforpublicationwasD.Huang.
ThisworkwassupportedbytheVINNOVAprojectMulti-OperatorDy-namicSpectrumAccess(MODyS).ThesecondauthoracknowledgesthatthisresearchwassupportedbytheInternationalResearch&DevelopmentProgramoftheNationalResearchFoundationofKorea(NRF)fundedbytheMinistryofEducation,ScienceandTechnology(MEST)ofKorea(K20901001406-09B1300-03110,FY2010).
K.W.SungandJ.ZanderarewithWireless@KTH,RoyalInstituteofTech-nology(KTH),SE16440,Kista,Sweden(e-mail:{sungkw,jenz}@kth.se).S-.L.KimiswithSchoolofElectricalandElectronicEngineering,YonseiUniversity,Seoul120-749,Korea(e-mail:slkim@yonsei.ac.kr).DigitalObjectIdenti er10.1109/TWC.2010.101810.100433
R
wasinvestigatedin[5].Theimpactofaggregateinterferenceinducedbymultiplesecondaryuserswasexaminedin[6],[7].In[8],spectrumsharinginafrequency-plannedcellularenvironmentwasevaluated.Reliabledetectionoftheprimaryuseriscrucialinrecognizingandexploitingthespatialop-portunity.Performanceofvarioussensingschemestodetectspatialopportunitieswerecomparedinarealisticpropagationenvironmentin[9].AnextensivesurveyofOSAcanbefoundin[10].
Thispaperfocusesontemporalaspectofthespectrumsharing.ThetemporalOSAexploitstheburstynatureofthedatatraf c.AprimaryuseralternatesON(busy)andOFF(idle)statesduringitsservicetime.Thechannelbecomesavailabletoasecondaryuserwhentheprimarytraf cisidleeveniftheprimaryuserisincloseproximitytothesecondaryuser.In[11]–[13],channelselection/switchingproblemswereinvestigatedundermultipleON/OFFchannelenvironmentsaimingatminimizingthedisruptiontoprimaryusers.Theauthorsof[14]examinedtheimpactofprimaryON/OFFtraf conthecapacityofspectrumsharing.Schedulingofspectrumsensingintervalwasstudiedin[15]–[17],wherethetradeoffbetweensensingef ciencyandinterferencetoprimaryuserswasaddressedbydynamicallyadjustingtheinter-sensingduration.However,the exiblesensingintervalmayintroduceasynchronizationproblemtothesecondarytransmitterandreceiverpair.
Inthispaper,weconsidera xed-lengthframestructureforthesecondaryusertohelpasimplesynchronization.Theframeisassumedtoconsistofmultipletimeslots.Weproposeanoptimalpowercontrolschemeforthesecondaryuser.Atthebeginningofeachframe,thesecondaryuserdecideswhethertotransmitornotfortheremainingtimeslotsintheframe.Thedecisionisbasedonthepredictionoftheprimaryuseractivity,i.e.theexpectedremainingtimeoftheOFFperiod.Theobjectiveoftheproposedschemeistwofold:tomaximizethespectrumutilizationofthesecondaryuserandtokeeptheinterferenceperceivedbytheprimaryuserbelowathreshold.Itisknownthattheburstinessofthewirelesssourcetraf cisnotproperlydescribedbytheexponentialdistribution[18],[19].Thus,weconsiderpracticalnon-Markoviansourcetraf cmodels.Peer-to-peer(P2P)andinteractivegamingservicesareassumedfortheprimaryusertraf cinthisstudybecausethesearepromisingapplicationsforthewirelessInternet[20]–[22].Interestingly,theseapplicationsshowoppositestatisticalbehaviorsintermsoftheremainingtimeoftheOFFperiod.Theframeworkdiscussedinthispapercanbereadilyadaptedtoanysourcetraf cmodel.Numericalexperimentsareper-formedwiththeconsiderationofthedetectionerrorsandtheinaccuracyinthetraf cparameterestimation.Ourresults
c2010IEEE1536-1276/10$25.00
Study on fatigue life prediction under spectrum loading_专业资料。第1第 2期 0卷21年 6月01 材 料 与冶 金 学 报 W7.0No211 . Junlo trl n tlryo...
Motion Prediction Based on Temporal Layering for Layered Video Coding ...user This work was supported by the Korea Science and Engineering Foundation...
Based on Linear Prediction QINGBO YIN RUBO ZHANG ...E. Detection of Anomalous Computer Session Activity...It is believed that the temporal signature method...
memory for the on-line recognition and prediction of temporal sequences J. ...Associative memories have an autonomous learning capability based on a ...
memory for the on-line recognition and prediction of temporal sequences J. ...Associative memories have an autonomous learning capability based on a ...
热门文档