1

Multiple-Antenna Systems
Jan Mietzner ([email protected], Room: Kaiser 4110)

1. Introduction

How is it possible to build (digital) wireless communication systems offering high data rates and small error rates ?

Multipleantenna techniques

Trade-off between spectral efciency (high data rates) and power efciency (small error rates), given xed bandwidth & transmission power

Tx

Rx

Example: Increase cardinality of modulation scheme Data rate , error rate Decrease rate of channel code Error rate , data rate and/ or in frequency domain
Spatial multiplexing techniques Spatial diversity techniques (Spacetime coding & diversity reception) Smart antennas (Beamforming)

Conventional transmitter & receiver techniques operate in time domain

Tradeo Multiplexing gain Diversity gain Coding gain Tradeo Antenna gain Interference suppression

Idea: Utilize multiple antennas at the transmitter and/ or the receiver Multiple-input multiple-output (MIMO) system Single-input multiple-output (SIMO) system Multiple-input single-output (MISO) system

Exploit spatial domain (in addition to time/ frequency domain)

Higher bit rates

Smaller error rates

Higher bit rates/ Smaller error rates

Better trade-off between spectral efciency and power efciency

Increased data rates by means of spatial multiplexing techniques Decreased error rates by means of spatial diversity techniques Improved signal-to-noise ratios (SNRs)/ signal-to-interference-plusnoise ratios (SINRs) by means of beamforming techniques

Benets of multiple antennas:

2. Basic Principles
2.1 Beamforming Techniques

Improved SNRs: Focus antenna patterns on desired angles of reception/ transmission, e.g., towards line-of-sight (LoS) or signicant scatterers Antenna gain

Goal: Improved SNRs or SINRs in multiuser scenarios Beamforming can be interpreted as linear ltering in the spatial domain Consider antenna array with N elements and directional antenna pattern receiving a radio-frequency (RF) signal from a certain direction

Improved SINRs: Steer nulls towards co-channel users Interference suppression

Beamforming/ smart antenna techniques thus enable space-division multiple access (SDMA), as an alternative to time-division or frequency-division multiple access (TDMA/ FDMA)

Due to antenna array geometry, impinging RF signal reaches antenna elements at different times (underlying baseband signal does not change)

SNR/ SINR gains can be utilized to decrease error rates or to increase data rates (by switching to a higher-order modulation scheme)

Adjust phases of RF signals to achieve constructive superposition

Corresponds to steering of antenna pattern towards desired direction Additional weighting of RF signals can shape antenna pattern
(N 1 degrees of freedom for placing maxima or nulls)

In practical systems directions of signicant scatterers must be estimated (e.g., MUSIC or ESPRIT algorithm); SINR can also be optimized without knowing the directions of all co-channel users (Capon beamformer)

Principle can also be utilized at the transmitter (reciprocity)

Transmitter Receiver

Beamforming techniques are well established since the 1960s (origins are in the eld of radar technology); however, intensive research for wireless communication systems started only in the 1990s

1 Beamformer

1 Beamformer

to detector

Literature: An exhaustive overview on smart antenna techniques for wireless communications can be found in [Godara97] Final remark: Beamforming can also be performed in baseband domain, if channel is known at transmitter and receiver (eigen-beamforming)

Information bit sequence

M Desired directions of transmission/reception

Phased array

Phased array

2.2 Spatial Multiplexing Techniques

2.3 Spatial Diversity Techniques

Goal: Increased data rates compared to single-antenna system Capacity of MIMO systems grows linearly with min{M, N } At the transmitter, the data sequence is split into M sub-sequences that are transmitted simultaneously using the same frequency band

Goal: Decreased error rates compared to single-antenna system Send/ receive multiple redundant versions of the same data sequence and perform appropriate combining (in baseband domain)

If the redundant signals undergo statistically independent fading,

Data rate increased by factor M (multiplexing gain)

it is unlikely that all signals simultaneously experience a deep fade

At the receiver, the sub-sequences are separated by means of interferencecancellation algorithm, e.g., linear zero-forcing (ZF)/ minimum-meansquared-error (MMSE) detector, maximum-likelihood (ML) detector, successive interference cancellation (SIC) detector, ...

Spatial diversity gain (typically, small antenna spacings sufcient)

Receive diversity: SIMO system with N receive antennas and linear combining of the received signals Various combining strategies, e.g., equal-gain combining (EGC), selection combining (SC), maximum-ratio combining (MRC), ... Well-established since the 1950s, see [Brennan59]

Typically, channel knowledge required solely at the receiver For a good error performance, typically N M required Intensive research started at the end of the 1990s Literature: [Foschini96] (Tutorials can be found in [Gesbert et al.03], [Paulraj et al.04])
Transmitter 1 2 Detection Algorithm Receiver

Transmit diversity: MISO system with M transmit antennas Appropriate pre-processing of transmitted redundant signals to enable coherent combining at receiver (space-time encoder/ decoder) Optionally, N > 1 receive antennas for enhanced performance Typically, channel knowledge required solely at the receiver Intensive research started at the end of the 1990s Well-known techniques are Alamoutis scheme for M = 2 transmit antennas [Alamouti98], space-time trellis codes [Tarokh et al.98], and orthogonal space-time block codes [Tarokh et al.99]

1 Demultiplexing

Information bit sequence

Estimated bit sequence

An abundance of transmitter/ receiver structures has been proposed (some offer additional coding gain)

M subsequences

Literature: An exhaustive overview of the benets of spatial diversity in wireless communication systems can be found in [Diggavi et al.04]

Transmitter

Receiver

Discrete-time channel model (contd): x [k ]: Transmitted symbol of transmit antenna , time index k ,
2 E{|x[k ]|2} =: x

1 Information bit sequence SpaceTime Encoder SpaceTime Decoder M Estimated bit sequence

E{x [k ]} = 0,

Redundant signals

3. Mathematical Details
3.1 System Model

Consider a MIMO system with M transmit and N receive antennas Assumptions: Frequency non-selective fading & square-root Nyquist lters at transmitter and receiver (pulse energy Eg := 1) Rayleigh fading (no LoS component), i.e., channel gains are zero-mean complex Gaussian random variables Block fading, i.e., channel gains are invariant over complete data block and change randomly from one block to the next

No intersymbol interference (ISI)

Discrete-time channel model: : Transmit antenna index (1 M ) : Receive antenna index (1 N ) k : Discrete time index (1 k NB, NB block length)

(Underlying information symbols are denoted as a[k ]) h, : Channel gain between th transmit & th receive antenna,
2 h, CN (0, h ) (i.i.d)

n [k ]: Additive white Gaussian noise (AWGN) sample at receive antenna , time index k ,
2 n [k ] CN (0, n ) (i.i.d)

(Amplitude |h, | is Rayleigh distributed)

y [k ]: Received symbol at receive antenna , time index k Matrix-vector model Transmitted vector: x[k ] := [ x1[k ], ..., xM [k ] ]T Noise vector: n[k ] := [ n1[k ], ..., nN [k ] ]T Received vector: y[k ] := [ y1[k ], ..., yN [k ] ]T Channel matrix:

h1,1 h1,M . . . . H := . ... . hN,1 hN,M

System model: y[k ] = H x[k ] + n[k ] (1)

10

3.2 Eigen-Beamforming

Transmit power allocation: In addition, the transmit power allocated to the parallel channels can be optimized, based on the instantaneous SNRs certain optimization criterion
2 | |2 x 2 n

Consider a quadratic MIMO system with M = N > 1 antennas Assume that the instantaneous realization of the channel matrix is perfectly known both at the transmitter and at the receiver

( = 1, ..., N ) and a

Eigenvalue decomposition of H:

3.3 Spatial Multiplexing

H

H := UU

(2)

: Diagonal (NN )-matrix containing eigenvalues 1, ..., N of H: 1 0 . ... . . = diag(1, ..., N ) = . . .

U: Unitary (NN )-matrix, i.e., UHU = IN

Consider a MIMO system with N M > 1 antennas (For N < M , the system is inherently rank-decient) Assume that the instantaneous realization of the channel matrix is known solely at the receiver

0 N

Linear ZF detection: Received vector y[k ] is post-processed as

Since H is perfectly known, transmitter and receiver can calculate the matrix U (e.g., using the Jacobian algorithm [Golub et al.96, Ch. 8.4])

zZF[k ] := (HHH)1 HHy[k ] =: H+y[k ]

(4)

(H+: Left-hand pseudo-inverse of H; for M = N and full rank use H1 )

Eigen-beamforming: Instead of x[k ], transmitter sends pre-processed vector x [k ] := Ux[k ] The received vector y [k ] is post-processed as U y [k ] =: y[k ]
H

zZF[k ] = H+y[k ] = H+ (Hx[k ] + n[k ]) = x[k ] + H+n[k ],

i.e., spatial interference completely removed; however, variance of the resulting noise samples may be signicantly enhanced
2 2 2 Linear MMSE detection: (assume x = ... = x =: x ) 1 M

y[k ] = UHy [k ] = UH(Hx [k ] + n[k ]) = UHHUx[k ] + UH n[k ] [k ] = x[k ] + n [k ] = UHUUH Ux[k ] + n y [k ] = x [k ] + n [k ] for all , = 1, ..., N
[k ] =: n

Received vector y[k ] is post-processed as

2 2 zMMSE [k ] := (HH H + n /x IM )1 HHy[k ]

(3)

(5)

Thus, assuming full rank (1 = 0, ..., N = 0) we have N parallel scalar channels without spatial interference (i.e., data rate enhanced by factor N compared to single-antenna system)
2 Noise samples n [k ] are still i.i.d. CN (0, n ), due to unitarity of U

Usually better performance than ZF detection, since better trade-off between spatial interference mitigation & noise enhancement
2 For high SNR values (n 0), both detectors become equivalent

Performance of ZF/ MMSE detection often quite poor, unless N M

11

12

ML detection:

3.4 Receive Diversity

2

[k ]|| ML [k ] := argminx x [k ] ||y[k ] Hx

for the transmitted vector x[k ]

(6)

[k ] For example, brute-force search over all possible hypotheses x For Q-ary modulation scheme, there are QM possibilities Optimal detection strategy (w.r.t. ML criterion), but very complex

Consider a SIMO system with N receive antennas Assume that the instantaneous realization of the (N1)-channel matrix is perfectly known at the receiver Received sample at receive antenna , time index k :

SIC detection: Good trade-off between complexity and performance Originally proposed in [Foschini96] for the well-known BLAST scheme (Bell-Labs Layered Space-Time Architecture) QR decomposition of H: (assume N = M )

y [k ] = h,1 x1[k ] + n [k ]
2 h,1 CN (0, h ) Amplitude |h,1| =: Rayleigh distributed 2 2 exp 2 2 h h
2 |h,1 |2 x 1 2 n

p( ) =
Instantaneous SNR

( 0),

(9)

=: Chi-square (2) distributed

H := QR R: Upper triangular (NN )-matrix:

(7)
where :=

Q: Unitary (NN )-matrix, i.e., QHQ = IN r1,1 r1,N . . . R= . . ... .

p ( ) =
2 2 x1 h 2 n

1 exp

( 0),

(10)

Large probability of small instantaneous SNRs

Idea: Combine received samples y1[k ], ..., yN [k ] to obtain more favorable SNR distribution at combiner output (comb ) Equal-gain combining (EGC): Add up all samples
N =1

0 rN,N (There are various algorithms for calculating the QR decomposition)

Received vector y[k ] is rst post-processed as zSIC [k ] := QH y[k ]

zcomb [k ] :=

[k ] zSIC [k ] := QHy[k ] = QH (Hx[k ]+ n[k ]) = Rx[k ]+ n

be detected

y [k ] =

N =1

h,1 x1[k ] +

N =1

n [k ]

(8)
2 hcomb CN (0, N h ),

=: hcomb 2 ncomb [k ] CN (0, N n ),

=: ncomb [k ]

Symbol xN [k ] is not affected by spatial interference and can directly Assuming that the detection of xN [k ] was correct, the inuence of

i.e., no gain!

Do it coherently (h,1 := ej )
zcomb [k ] := N =1

xN 1 [k ] can directly be detected, and so on ...

x N [k ] can be subtracted from the (N 1)th row of (8); then symbol

ej y [k ] =

=1

x1[k ] +

N =1

ej n [k ]

=: hcomb

Combiner-output SNR: comb = (

=: ncomb [k ] 2 2 2 ) x1 /(N n )

13

14

Selection combining (SC): Select branch with largest instant. SNR Combiner-output SNR: comb =

Example: BPSK, N = 1, ..., 4 receive branches

10
0

2 2 2 max { } x /n 1

= max { }
N

Maximum-ratio combining (MRC):

zcomb [k ] :=

N =1

h ,1 y [k ]

N =1

|h,1|

x 1 [k ] +
2

=1

h ,1 n [k ]

SER

10

N=1 receive branches N=2 receive branches N=3 receive branches N=4 receive branches Alamoutis scheme (M=2, N=1)

=: hcomb

=: ncomb [k ]

10

Combiner-output SNR: comb = (

2 2 |h,1 | )x1 /n

Maximizes combiner-output SNR; optimal w.r.t. ML criterion

10

Symbol error rates (SERs) with MRC: (without derivation ;-) )

10

: Average SNR per receive branch

Binary Phase-Shift Keying (BPSK) [Proakis01, Ch. 14]
0 2 4

SER( ) =

1 1 2N 1+
(Q1) Q

N 1 i=0

N 1 + i 1 1 + 2i 1+ i
[Simon et al.00]
N

6 8 10 12 14 Average SNR per branch (in dB)

16

18

20

(11)

Asymptotic slope (i.e., ) of the curves is N (diversity order N ) 3.5 Transmit Diversity

Q-ary Phase-Shift Keying (Q-PSK)

1 SER( ) =

sin2 d sin2 + sin2(/Q)

N

(12)

Consider a MISO system with M transmit antennas Assume that the instantaneous realization of the (1M )-channel matrix is perfectly known at the receiver, but not at the transmitter Transmit Diversity: Suitable pre-processing of transmitted data sequence required to allow for coherent combining at the receiver Example: Send identical signals over all transmit antennas

Q-ary Amplitude-Shift Keying (Q-ASK)

[Simon et al.00]

2(Q 1) 2 (Q2 1) sin2 d SER( ) = Q 0 (Q2 1) sin2 + 3 1 2 2(Q 1) sin2 4 1 d SER( ) = Q 0 2(Q 1) sin2 + 3 4 1 1 Q
2
4

(13)

Q-ary Quadrature-Amplitude Modulation (Q-QAM) [Simon et al.00]

2(Q 1) sin2 d 2(Q 1) sin2 + 3

(14)

Instead: Perform appropriate two-dimensional mapping/ encoding in time and space (i.e., over the transmit antennas)

No diversity gain! (corresponds to EGC without co-phasing)

15

16

Example: Alamoutis scheme for M = 2 transmit antennas (N = 1 receive antennas considered; can be extended to N > 1) Space-time mapping: Information symbols to be transmitted are processed in pairs [ a[k ], a[k + 1] ]; at time index k , symbol a[k ] is transmitted via the rst antenna and symbol a[k + 1] via the second antenna; at time index k +1, symbol a[k +1] is transmitted via the rst antenna and symbol a[k ] via the second antenna

Two parallel scalar channels for the symbols a[k ] and a[k +1]
(no spatial interference)

Corresponds to MRC with M = 1 transmit and N = 2 receive antennas;

however, using the same average transmit power, Alamoutis scheme exhibits a 3 dB loss compared to MRC

4. Literature
4.1 Cited References

A =

antenna 1

a [k ] a[k +1] time index k a [k +1] a [k ] time index k +1 antenna 2

L. C. Godara, Application of antenna arrays to mobile communications Part I: Performance improvement, feasibility, and system considerations; Part II: Beam-forming and direction-of-arrival considerations, Proc. IEEE, vol. 85, no. 7/8, pp. 10311060, 11951245, July/Aug. 1997.

(15)

(In terms of prior system model: A =: [ xT[k ], xT [k +1] ]T ) Received samples (time index k, k +1):

y1[k ] = h1,1 a[k ] + h1,2 a[k +1] + n1[k ] y1[k +1] = h1,1 a[k +1] + h1,2 a[k ] + n1[k +1]
Equivalent matrix-vector model (by taking the (.) of y1 [k +1])

G. J. Foschini, Layered space-time architecture for wireless communication in a fading environment when using multi-element antennas, Bell Syst. Tech. J., pp. 4159, Autumn 1996.

y1[k ]
y1 [k +1] =: yeq [k ]

h1,1 h1,2
h 1,2 h1,1 =: Heq

a [k ] a[k +1]
=: a[k ]

n 1 [k ] n 1 [k +1]
=: neq [k ]

D. Gesbert, M. Sha, D. Shiu, P. J. Smith, and A. Naguib, From theory to practice: An overview of MIMO space-time coded wireless systems, IEEE J. Select. Areas Commun., vol. 21, no. 3, pp. 281302, Apr. 2003.

Detection step at the receiver:

2 2 Heq is always orthogonal (!), while HH eq Heq = (|h1,1 | + |h1,2 | ) I2 H H zcomb [k ] := HH eq yeq [k ] = Heq Heq a[k ] + Heq neq [k ] =: neq [k ]

A. J. Paulraj, D. A. Gore, R. U. Nabar, and H. Boelcskei, An overview of MIMO communications A key to gigabit wireless, Proc. IEEE, vol. 92, no. 2, pp. 198218, Feb. 2004.

D. G. Brennan, Linear diversity combining techniques, Proc. IRE, vol. 47, pp. 10751102, June 1959, Reprint: Proc. IEEE, vol. 91, no. 2, pp. 331-356, Feb. 2003.

= (|h1,1|2 + |h1,2|2) a[k ] + neq[k ]

17

18

S. M. Alamouti, A simple transmit diversity technique for wireless communications, IEEE J. Select. Areas Commun., vol. 16, no. 8, pp. 1451 1458, Oct. 1998.

4.2 Books on Multiple-Antenna Systems

S. Haykin, Ed., Array Signal Processing. Englewood Cliffs (NJ): PrenticeHall, 1985.

V. Tarokh, N. Seshadri, and A. R. Calderbank, Space-time codes for high data rate wireless communication: Performance criterion and code construction, IEEE Trans. Inform. Theory, vol. 44, no. 2, pp. 744765, Mar. 1998.

A. Paulraj, R. Nabar, and D. Gore, Introduction to Space-Time Wireless Communications. Cambridge University Press, 2003.

B. Vucetic and J. Yuan, Space-Time Coding. John Wiley & Sons, 2003. E. G. Larsson and P. Stoica, Space-Time Block Coding for Wireless Communications. John Wiley & Sons, 2003.

V. Tarokh, H. Jafarkhani, and A. R. Calderbank, Space-time block codes from orthogonal designs, IEEE Trans. Inform. Theory, vol. 45, no. 5, pp. 14561467, July 1999.

T. Kaiser, A. Bourdoux, H. Boche, J. R. Fonollosa, J. Bach Andersen, and W. Utschick, Eds., Smart Antennas State of the Art. Hindawi Publishing Corp., 2004. New York:

S. N. Diggavi, N. Al-Dhahir, A. Stamoulis, and A. R. Calderbank, Great expectations: The value of spatial diversity in wireless networks, Proc. IEEE, vol. 92, no. 2, pp. 219270, Feb. 2004.

E. Biglieri and G. Taricco, Transmission and Reception with Multiple Antennas: Theoretical Foundations. lishers Inc., 2004. Hanover (MA) - Delft: now Pub-

G. H. Golub and C. F. van Loan, Matrix Computations, 3rd ed. more - London: The Johns Hopkins University Press, 1996.

Balti-

J. G. Proakis, Digital Communications, 4th ed. Hill, 2001.

New York: McGraw-

H. Jafarkhani, Space-Time Coding Theory and Practice. University Press, 2005.

Cambridge

M. K. Simon and M.-S. Alouini, Digital Communication over Fading Channels: A Unied Approach to Performance Analysis. John Wiley & Sons, 2000.