Beamforming Techniques in Wireless …

1 Beamforming Techniques in WirelessCommunicationsJaved AkhtarSupervisor:Dr. Ketan RajawatDepartment of Electrical EngineeringIndian Institute of Technology, KanpurKanpur, Uttar PradeshBeamforming Techniques in Wireless Communications1 / 51 OutlineBeamforming: Introduction & OverviewTransmit & Receive Beamforming :- ExamplesCommon Beamforming TechniquesBeamformingIn MIMO SystemsIn Massive MIMO SystemsIn MIMO-OFDM SystemsFor Interference Mitigation in Multi-antenna, Multi-carrier SystemsIn Wireless Sensor NetworksFuture WorkBeamforming Techniques in Wireless Communications2 / 51 BeamformingDirectional transmission/reception to optimize a design criterion [1]Objective:Design transmit & receive beamformers to nullify theinterference and enhance system performanceCan be used at both transmitting and receiving endsApplication in different areas of Wireless communications.

2 Multiple-Input and Multiple-output(MIMO)Massive MIMO(m-MIMO)Interference mitigationWireless sensor Networks(WSN)[1] Mietzner, J.; et al., Multiple-antenna Techniques for Wireless communications - acomprehensive literature survey, in Communications Surveys & Tutorials, 2009 Beamforming Techniques in Wireless Communications3 / 51 System Model: An OverviewPWChannel(H)X=Psyr=WysTransmit BeamformerReceive BeamformerTxRxThe general model for a MIMO Wireless communication is given asy=Hx+n;H CNR NT(1)=HPs+n(2)x=Psis the precoded andr=Wyis the filtered outputBeamforming Techniques in Wireless Communications4 / 51 Transmit & Receive Beamforming :- ExamplesBeamforming Techniques in Wireless Communications5 / 51 Transmit and Receive Beamforming - ExampleConsider singular value decomposition(SVD) ofHH=U VH(3)where is a diagonal matrix with entries i(singular values)ChoosingP=VandW=UHin eq.

3 (2) givesr= s+UHn(4)= s+ n(5)Then, one data stream per singular value can be transmitted asri= isi+ ni; 1 i min{NR,NT}(6)[2] E. Telatar, Capacity of multiantenna Gaussian channels, European Transactions onTelecommunications , 1999 Beamforming Techniques in Wireless Communications6 / 51 Precoder: Transmit Beamforming - ExampleMultiple Input Single Output(MISO) system[3]y=hHfs+n(7)Here,P=fandW=IAntenn a Selection: Send data on the antenna that maximizes thereceive SNR=|hHf|2mopt= arg max1 m NT|h(m)|2(8)Transmit Beamforming vector is restricted to rank one covariancematrixOptimal selected antenna can be fedback to the transmitter usingdlog2(NT)ebits[3] D. Love, R. Heath, et al., An overview of limited feedback in Wireless communicationsystems , IEEE Journal on Selected Areas in Communications, Techniques in Wireless Communications7 / 51 Common Beamforming TechniquesBeamforming Techniques in Wireless Communications8 / 51 MRC: Maximal Ratio CombiningAll received signals are coherently combined at the receiver [5]Here,W=w(vector withNRelements) andP=Ir=wHy=wHhx+wHn(9)Maximizes the output SNR for the intended user =|wHh|2E|wHn|2=|wHh|2 2E|wHw|(10)By Cauchy-Schwartz inequality it is found that the SNR ismaximized when,w h[4] Tse, David, and Pramod Viswanath,Fundamentals of Wireless communication, Cambridgeuniversity press, 2005 Beamforming Techniques in Wireless Communications9 / 51ZF: Zero ForcingPrecoding.

4 MISO system with (K)-user is considered in [6]Received signal atkthuser user is given asyk=hHkx+nk;k= 1,2, ,K(11)where,x= Ki=1siwiis theNT 1 transmitted symbolswiis the linear precoder, orthogonal to all other user channel vectorsHence,yk=hHk Ki=1wisi+nk=hHkwksk+nk;k= 1,2, ,K(12)[5] Peel, et. al., A vector-perturbation technique for near-capacity multiantenna multiusercommunication-part I: channel inversion and regularization, IEEE Tran. on Comm., 2005 Beamforming Techniques in Wireless Communications10 / 51ZF: Zero ForcingEqualization:Considering the MIMO signal model given in eq. (1)To decouple the detection of each symbol at the receiverW=H so,r=H y=x+H n(13)where,H is the pseudo inverse given as (HHH) 1HH[6] Tse, David, and Pramod Viswanath,Fundamentals of Wireless communication, Cambridgeuniversity press, 2005 Beamforming Techniques in Wireless Communications11 / 51 State of the ArtMulti-cell MIMO Networks [a],[b]Cognitive Radio Networks [c],Massive-MIMO Networks [a]Dirty Paper Coding (DPC) based algorithms[d][a] Lakshminarayana, S.

5 ; Assaad, M.; Debbah, M., Coordinated Multicell Beamforming forMassive MIMO: A Random Matrix Approach , IEEE Tran. on Inform. Th., 2015[b]Venkategowda, ; et al., MVDR-Based Multicell Cooperative Beamforming Techniques forUnicast/Multicast MIMO Networks With Perfect/Imperfect CSI , Tran. on Veh. Tech., 2015[c]Afana, A.; et al., Distributed Beamforming for Two-Way DF Relay Cognitive Networks UnderPrimarySecondary Mutual Interference , Tran. on Veh. Tech., 2015[d] N. Jindal and A. Goldsmith,Dirty-paper coding versus TDMA for MIMO broadcast channels,IEEE Trans. Inf. Theory, 2010 Beamforming Techniques in Wireless Communications12 / 51 Beamforming in MIMOB eamforming Techniques in Wireless Communications13 / 51 Beamforming in MIMOIn [7], transmission of data symbols subject to (possibly different)QoSconstraints is consideredSystem model is same as in eq.

6 (1)Optimum Receiver: for a given transmit matrixPminWTr[( x x)( x x)H](14)where x=Wxand the optimalWis the linear minimum MSE(LMMSE) filter[8][7] D. P. Palomar, et al.,Optimum linear joint transmit-receive processing for MIMO channels withQoS constraints,IEEE Trans. Signal Process., 2004[8] Palomar, and Jiang, Y., MIMO transceiver design via majorization theory. Foundationsand trends in communications and information theory , 2006 Beamforming Techniques in Wireless Communications14 / 51 Beamforming in MIMO [7] Precoder:minPTr(PPH)(15) [(I+PHRHP) 1]ii i,1 i L(16)where,Lis the number of established links andRH=HHR 1nHAbove problem is nonconvex inPMajorization theory is used to reformulate it as a simple convexoptimization problemContribution:A practical and efficient multilevel water-filling algorithm is proposedA simple robust design under channel estimation errors is alsoproposedBeamforming Techniques in Wireless Communications15 / 51 Beamforming in MIMO.

7 Robust FrameworkIn [9], the design of linear transceivers with robustness for imperfectCSI is consideredThe MIMO channel matrix distribution is modeled asH=H+ (RRxH)1/2G(RTxH)(1/2)H(17)whereGis the unknown part in the fading estimateHis the channel mean & covariance matrixRH= (RRxH) (RTxH)(18)[9] Xi Zhang; et al., Statistically Robust Design of Linear MIMO Transceivers, in TSP, 2008 Beamforming Techniques in Wireless Communications16 / 51 Beamforming in MIMO [9] Problem:minW,PFo{Tr[E{( x x)( x x)H}]}(19) Tr[PP ] pmax(20)where,Fois an arbitrary increasing function of the average MSECase1: Imperfect CSIR & CSITA closed-form expression for the average MSE matrix isE{E(W,P)}= [(WHHP Il)(PHHHW Il) +WHR nW](21)where,R n=Rn+ Tr[PPH(RTxH)T]RRxHThe optimal receiverWis similar to the Wiener filterThe optimal design of the transmitterPis derived forRTxH=INTB eamforming Techniques in Wireless Communications17 / 51 Beamforming in MIMO [9].

8 Perfect CSIR & imperfect CSITThe optimal receiverWis exactly the same as for perfect CSI caseA robust transmitter design is proposed based on a tight lower boundofE{E}Cost function of the design problem (19) is also lower-boundedBeamforming Techniques in Wireless Communications18 / 51 Beamforming in Massive MIMOB eamforming Techniques in Wireless Communications19 / 51 Beamforming in Massive MIMOM assive MIMO employs hundreds of antennas to enable a largebeamforming gainIn [10], quantized versions of the channel estimates are used forconjugate beamformingDesirable to use A/D and D/A converters with resolutions as coarseas possibleM-antenna base station,Kautonomous single antenna terminalswith Time Divison Duplexing(TDD) operation is consideredThe minimum mean-square error (MMSE) estimate forHis H= u u1+ u uywhere uis coherence slot for up-link pilots uis a measure of the expected SNR of the up-link[10] Hong Yang.

9 Marzetta, , Quantized Beamforming in Massive MIMO , Annual Conferenceon Information Sciences and Systems (CISS), 2015 Beamforming Techniques in Wireless Communications20 / 51 Beamforming in Massive MIMO[10] Beamforming :x= HHq MK u u1+ u u(22)The down-link data channel is given byy= dHHx+n(23) dis a measure of the expected SNR of the down-link channelEffects of quantization :x= HH q cwhere, the constantc=E[Tr( H HH )] is the quantization scheme on the channel response estimate HOpen Problems:Consideration of multi-cell interference and pilot contaminationAnalysis and impact of multi-antenna receiverBeamforming Techniques in Wireless Communications21 / 51 Beamforming in Multi-Antenna &Multi-Carrier SystemBeamforming Techniques in Wireless Communications22 / 51 Beamforming in MIMO-OFDM SystemIn [11], robust transceiver designs with statistical channeluncertainties are consideredThe channel matrices on each subcarrier have the followingrelationshipHk= Hk+ Hk(24)where, Hk= 2kHw 2kis the estimation error variance of each channel elementHwis the random matrix whose elements are with zero meanand unit variance[11] Chengwen Xing; et al.

10 , Robust Transceiver Design for MIMO-OFDM Systems Based onCluster Water-Filling, in IEEE Communications Letters, 2013 Beamforming Techniques in Wireless Communications23 / 51 Beamforming in MIMO-OFDM Systems[11] weighted MSE optimization problem of transceiver design isformulated asminPkK k=1Tr{Wk[PHk HHkK 1Pk HkPk+I] 1}(25) KPk= [ 2e,kTr(PkPHk) + 2n]I(26)K k=1Tr(PkPHk) pmax(27)where,Wkis the weighting matrix in thekthsubcarrierA series of auxiliary variablesPkare introducedThus enabling optimization problem to be decoupled into a series ofsubproblemsCluster water-filling is proposed to solve the robust transceiverdesignBeamforming Techniques in Wireless Communications24 / 51 Beamforming for Interference MitigationBeamforming Techniques in Wireless Communications25 / 51 Interfernce AlignmentInterference Alignment (IA) concepts have been well explored in[12],[13]Interference may arise due to multiuser, multicell, Inter-BlockInterference(IBI), Inter-Symbol Interference (ISI) signal space and interference space in disjoint subspacesKey idea is to maximize the space for the desired signalMinimize the interference space at the receiver which can besuppressed using proper filter (.)