conventionalCoSaMPalgorithmcannotworkduetoN<p
algorithmwithR=4worksbetterthanthatwithR=1,andbothofthemperformwellandapproachtotheexactLS
method(performancebound).Fortheproposedsuperimposedpilotdesign,thetotalN=800pilotsoccupy19.53%
p_total
ofthetotalN=4096subcarriers,andtheequivalentaveragepilotoverheadpertransmitantennaisjustN=12.5
p_avg
(only0.30%ofthetotalN=4096subcarriers)comparedwiththeconventionalorthogonalpilots.Suchlowpilot
overheadisalmostimpossiblefortheconventionalalgorithmstorealizeaccuratechannelestimation.
-1
-2
10
10
MSE
19.92020.1
-2
10
ProposedAlgorithm,R=1
ProposedAlgorithm,R=4
ClassicalCoSaMPAlgorithm
ExactLS(Bound)
-3
10
1012141618202224262830
SNR(dB)
Figure2:MSEperformancecomparisonovertheITUVehicularBchannel
M
ItisworthnotingthatsensingmatrixΦdependsonthepilotposition?andpilotsequences{}s.Although
mm?1
M
simulationresultsindicatethatthespecific?andasmentionedinthispaperhavereliableperformancedueto
{}s
mm?1
thenear-orthogonalcolumnsofΦ,theoptimaldesignofthepilotpositionandpilotsequenceremainsaninteresting
problemtobestudiedinthefuture.
5.Conclusion
Thispaperfocusesonthedownlinktrainingandchannelestimationforlarge-scaleMIMOsystems.Incontrast
tostandardizedorthogonalpilotswiththeprohibitiveoverheadincreasingwiththenumberoftransmitantennas,the
proposedsuperimposedpilotdesignbasedonstructuredCScanefficientlysolvethepilotoverheadproblem.Atthe
receiver,theproposedstructuredCoSaMPalgorithmcanexploitthespatialandtemporalcorrelationsoflarge-scale
MIMOchannelsforsimultaneousrecoveryofmultiplechannels.Moreover,theproposedsuperimposedpilotdesign
andthecorrespondingchannelestimatorcanbeappliedintheuplinktoo,andconventionalsmall-scaleMIMOcanalso
adopttheproposedschemetoreducethepilotoverheadandimprovethechannelestimationperformance.Theremained
problemtobesolvednextistheoptimaldesignofthepilotpositionandpilotsequenceforlarge-scaleMIMO.
6.References
1.T.L.Marzetta,“Noncooperativecellularwirelesswithunlimitednumbersofbasestationantennas,”IEEETrans.
WirelessCommun.,vol.9,no.11,pp.3590-3600,Nov.2010.
2.F.Rusek,D.Persson,B.K.Lau,E.G.Larsson,T.L.Marzetta,O.Edfors,andF.Tufvesson,“ScalingupMIMO:
Opportunitiesandchallengeswithverylargearrays,”IEEESignalProcess.Mag.,vol.30,no.1,pp.40-60,Jan.2013.
3.F.Fernandes,A.E.Ashikhmin,andT.L.Marzetta,“Inter-cellinterferenceinnoncooperativeTDDlargescale
antennasystems.”IEEEJ.Sel.AreasCommun.,vol.31,no.2,pp.192-201,Feb.2013.
4.M.DuarteandY.Eldar,“Structuredcompressedsensing:fromtheorytoapplications,”IEEETrans.SignalProcess.,
vol.59,no.9,pp.4053-4085,Sep.2011.
5.L.NeedellandJ.A.Tropp,“Cosamp:Iterativesignalrecoveryfromincompleteandinaccuratesamples,”Commun.
ACM,vol.53,no.12,pp.93–100,Dec.2010.
6.Y.BarbotinandM.Vetterli,“EstimationofsparseMIMOchannelswithcommonsupport,”IEEETrans.Commun.,
vol.60,no.12,pp.3705-3716,Dec.2012.
7.L.Dai,J.Wang,Z.Wang,P.Tsiaflakis,andM.Moonen,“Spectrumandenergy-efficientOFDMbasedon
simultaneousmulti-channelreconstruction,”IEEETrans.SignalProcess.,vol.61,no.23,pp.6047-6059,Dec.2013. |
|
