[Type the document title]

MODULE-2
Multiple Antenna Transmission and Reception

Spatial Diversity Overview

• Spatial diversity is exploited through 2 or more antennas which are separated by
enough distance so that fading is approximately decorrelated between them.
• No additional bandwidth or power is required for spatial diversity.
• When multiple antennas are used, 2 forms of gain are available which are referred to
as Array Gain and Diversity Gain.
• Array Gain does not rely on statistical diversity between different channels whereas
Diversity Gain is a product of statistical richness of channels and results from the
creation of multiple independent channels between transmitter and receiver.

Array Gain
• Array gain achieves its performance enhancement by coherently combining the
energy of each of the antenna to gain an advantage versus the noise signal on each
antenna which is uncorrelated and so does not add coherently.
• Even if channels are completely correlated, the SNR increases linearly with the
number of receive antennas.
• In correlated flat fading, each antenna receives a signal that can be characterized as.
• SNR on a single antenna is whereis noise power.
• If all receive antenna paths are added, the resulting signal is

• Assuming noise on each branch is uncorrelated, the combined SNR is

Diversity Gain
• The physical layer reliability is typically measured by the outage probability or
average bit error rate.
• Bit Error Probability (BEP)is given by .
where C1 and C2 are constants that depend on modulation type and is the received
SNR.
• With fading, SNR becomes random variable and so BEP is also a random variable.

1
 [Type the document title]

• Without diversity, the average BEP is .

• When Nt transmit antennas and Nr receive antennas are added to the system, the
diversity order is .
• With diversity, the average BEP is.
• If only array gain was possible, the average BEP decreases to .

Increasing the Data Rate with Spatial Diversity

• The Shannon capacity formula gives the maximum achievable data rate of a single
communication link in Additive White Gaussian Noise (AWGN)as

where C is the capacity or maximum error free data rate, B is the bandwidth of
channel and is SNR.
• Since antenna diversity increases the SNR linearly, diversity techniques increase the
capacity only logarithmically with respect to the number of antennas.
• Spatial multiplexing has the ability to achieve a linear increase in the data rate with
the number of antennas.
• The capacity can be increased as a multiple of mini.e. Capacity is limited by
minimum of number of antennas at either transmitter or receiver.

2
 [Type the document title]

Increased Coverage or Reduced Transmit Power

• If the increase in coverage range is , the coverage area improvementis .
• The required transmit power reduces todB while maintaining a diversity gain of .

Receive Diversity
1)Selection Combining (SC)
• It is the simplest type of combiner which estimates the instantaneous strengths of each
of the streams and selects the highest one.
• Its simplicity and reduced hardware and power requirements make it attractive for
narrowband channels.
• The outage probability is .
• Assuming uncorrelated receptions of the signal,

• For a Rayleigh fading channelwhereis the average received SNR at that location.
• The selection combining decreases the outage probability to

• The average received SNR for branch selection combining (SC) can be derived in
Rayleigh fading to be

• The average BEP can be derived by averaging (integrating) the appropriate BER
expression in AWGN against the exponential distribution.

3
 [Type the document title]

2)Maximal Ratio Combining (MRC)

• MRC combines the information from all the received branches in order to maximize
the ratio of signal to noise power.
• It works by weighting each branch with a complex factor and then adding
up the branches.
• The received signal on each branch can be written as .
• Assuming the fading is flat with a complex value ofon the ith branch, the combined
signal is .
• If the phase of combining coefficient for all branches, then SNR of y(t) is

where is the transmit signal energy.

• Maximizing by taking derivative with respect togives maximizing combining values
as i.e. each branch is multiplied by its SNR.
• The resulting SNR is .
• MRC ignores interference power.
• Equal Gain Combining (EGC) which only corrects phase usesand
for all combiner branches, achieves a post combining SNR of

• Advantage of MRC in frequency selective fading channels is that all the frequency
diversity can be utilized.
• An RF antenna selection algorithm selects the best average antenna.
• In a wideband channel, different coherence bands will have different SNRs and
hence MRC is used.

4
 [Type the document title]

Transmit Diversity
• Signals are sent from different transmit antennas and they interfere with one another
at the receiver.
1) Open Loop Transmit Diversity
• They are systems that do not require knowledge of the channel at the transmitter.
2) Closed Loop Transmit Diversity
• They are systems that require channel knowledge at transmitter.
1) Open-Loop Transmit Diversity

i) 2 X 1 Space-Frequency Block Coding

• The most popular open loop transmit diversity scheme is space-time coding where a
particular code known to receiver is applied at the transmitter.
• The receiver must know the channel to decode the space-time code.
• Space-Time Block Code (STBC) referred to as either Alamouti Code or the
Orthogonal Space-Time Block Code (OSTBC) came into existence in 1990s.
• This simple code is linear at both transmitter and receiver.
• STBCs can easily be adapted to a wide band fading channel using OFDM and they
use consecutive symbols in time.
• Space-Frequency Block Codes (SFBCs) are preferred to STBCs because they prefer
less delay.
• STBCs would require 2 OFDM symbols to be encoded and decoded which
significantly increases delay.
• The simplest SFBC corresponds to 2 transmit antennas and 1 receive antenna.
• If 2 symbols transmitted are S1 and S2, the Alamouti code sends the following over 2
sub carriers f1 and f2:

5
 [Type the document title]

• The 2X1 Alamouti SFBC is referred to as a rate1 code, since the data rate is neither
increased nor decreased; 2 symbols are sent over 2 adjacent sub carriers.
• Assuming a flat fading channel on each sub carrier, is the complex channel gain
from transmit antenna 1 to the receive antenna and is from transmit antenna 2.
• Assuming the channel is constant over 2 adjacent sub carriers i.e. .,the
received signal r(f) is
where n(.) is a sample of white Gaussian noise.
• Assuming the channel is known at the receiver

• Therefore,

• The resulting SNR can be computed as

• The total transmit energy per actual data symbol is .

• For SFBC, .

ii) 2X2 SFBC

• The channel description is represented as 2X2 matrix .
• The resulting signals at sub carriers f1 and f2 on antennas 1 and 2 can be represented
as

• Using the combining scheme

6
 [Type the document title]

• The resulting SNR is

• The decision statistics are obtained as

where is the interference from the ith transmit due to transmitting 2 simultaneous
data streams.

iii) 4X2 in LTE

• When 4 transmit antennas are available, a combination of SFBC and Frequency
Switched Transmit Diversity (FSTD) is used.
• The combination of SFBC and FSTD is a rate 1 diversity scheme i.e. 4 modulation
symbols are sent over 4 OFDM symbols using space-frequency encoder where the
columns correspond to the sub carrier index and rows to the transmit antenna.

• The 1st and 2nd symbols S1 and S2 are sent over antenna ports 0 and 2 on the first 2
OFDM sub carriers in the block.
• The 3rd and 4th symbols are sent using antenna port 1 and 3 on the other 2 sub carriers.

7
 [Type the document title]

2) Closed-Loop Transmit Diversity

i) Transmit Selection Diversity (TSD)

• It is the simplest form of transmit diversity where only a subset of the
available antennas is used at a given time.
• The selected subset typically corresponds to the best channels between the transmitter
and receiver.
• Hardware cost, complexity and spatial interference is reduced since only fewer
transmit signals are sent.
• In spite of maintaining diversity order same i.e. , TSD is not optimal in terms
of diversity gain.
• Wideband channels have multiple coherence bands so the gain from selecting the best
antenna averaged over all the coherence bands is likely to be small.
• The average SNR with single transmit antenna selection in asystem with independent
and identically distributed Rayleigh fading is .
• The feedback required for antenna TSD is low since index of the required antenna is
needed rather than the full Channel State Information (CSI).
• In case of single transmit antenna selection, only bits of feedback are needed for each
channel realization.

8
 [Type the document title]

• The required feedback for transmit antenna selection is aboutbits per channel
coherence time.
ii) Linear Diversity Pre coding
• Linear Pre coding is a general technique for improving the data rate or link reliability
by exploiting the CSI at the transmitter.
• Linear Pre coder at the transmitter and a linear Post coder at the receiver are applied
only to improve the link reliability.
• The received data vector z can be written as.
• The size of the transmitted vector x is Mx1 and the received vector y is Nrx1.
• The post coder matrix G is M x Nr, the channel matrix H is Nr x Nt, the pre coder
matrix F is Nt x M, the noise vector n is Nrx1 and received data vector z is Mx1.
• M is the number of spatial data “streams” sent in general case.
• For case of pure diversity pre coding only one data symbol is sent at a time where
M=1 and the SNR maximizing pre coder F and post coder G are the right and left
singular vectors of H corresponding to its largest singular value .
• The equivalent channel model after pre coding and post coding for a transmitted data
symbol x becomes y = hx + n.
• The received SNR iswhereis the noise variance.
• SNR is bounded as .
• denotes the Frobenius norm which is just the total sum of all the powers of each
spatial channel defined as .

• SNR for the case of STBC is given as .

• To employ linear diversity pre coding, CSI is required at the transmitter.

DOA based Beam steering

• Electromagnetic waves can be physically steered to create beam patterns at either
transmitter or receiver.
• At the transmitter, this causes energy to be sent predominantly in a desired direction
while only a small amount of residual energy is leaked in other directions.
• The most common and simple form is static pattern gain beam steering which is
known as sectoring.

9
 [Type the document title]

• Each DOA can be estimated using signal processing techniques such as MUSIC,
ESPRIT and MLE algorithms.
• From the acquired Direction of Arrival (DOA), a beam former extracts a weighting
vector for the antenna elements and uses it to transmit or receive the desired signal of
a specific user while suppressing the undesired interference signals.
• When the plane wave arrives at the d-spaced Uniform Linear Array (ULA) with
Angle of Arrival (AOA) θ, the wave at the first antenna element travels an additional
distance of dsinθ to arrive at the second element.
• The difference in propagation distance between the adjacent antenna elements can be
formulated as an arrival time delay τ = d/c sinθ.
• The signal arriving at the second antenna can be expressed in terms of the signal at the
first antenna element as

• For an antenna element with Nr elements all spaced by d, the resulting received signal
vector can be expressed as

where a(θ) is the array response vector.

• Considering a 3 element ULA with d=λ/2 spacing between the antenna elements and
assuming that the desired user’s signal is received with an AOA of θ1=0 and 2
interfering signals are received with AOAs of θ2=π/3 and θ3=-π/6, the array response
vectors are given by

• The beam forming weight vector should increase the

antenna gain in the direction of desired user while simultaneously minimizing the gain
in the directions of interferers.
• The weight vector should satisfy the criterion .
• A unique solution for the weight vector is obtained as

10
 [Type the document title]

• The null steering beam former (DOA based beam former) can be designed to
completely null out interfering signals only if the number of such signals is strictly
less than the number of antenna elements i.e. if the number of receive antennas is Nr,
then Nr-1 independent interferers can be cancelled.
• The disadvantage of DOA based beam former is that a null is placed in the direction
of the interferers and so the antenna gain is not maximized at the direction of the
desired user.

Linear Interference Suppression

1) By complete knowledge of Interference Channels
• Considering a single transmitter with Nt antennas trying to communicate to a receiver
with Nr >Nt antennas in the presence of one or moreinterfering transmitters each with
antennas, the total number of interfering sources is .
• With L = 1, Nt = 1 and Nr = 2, we have a total of 2 transmitted streams to a 2 antenna
receiver.
• The received signal model is y = Hx + n.
• Assuming the receiver knows not only its channel vector but the interfering channel
as well, a ZF receiverwould produce .
• If the transmitters are independent rather than co-located, it is known as uplink
multiuser MIMO or uplink SDMA since 2 users are supported simultaneously.

11
 [Type the document title]

2) By statistical knowledge of Interference Channels

• Allowing the transmitter to pre code its signal with a Ntx1 beam forming vector ,
the Nr dimensional received signal vector at the receiver is
where x is the desired symbol with energy
is the interference vector
n is the noise vector with covariance matrix
H is Nr x Nt channel gain matrix for the desired user
is the Nr x L channel gain matrix for the interferers
is the eigen vector corresponding to the largest eigen value
andwhere α is an arbitrary constant
is the interference plus noise covariance matrix
is the largest eigen value of A.
• The maximum output SINR is .
• The transmit power is focused on the largest eigen channelamong the
eigen channels in order to maximize post beam forming SINR and this beam forming
approach is termed as Optimum eigen beam former, interference aware beam forming
and/or Optimum Combiner (OC).
• If the interference terms are ignored,andand .
• In the absence of interference, the output SNR of optimum beam former with Nt> 1
can be upper and lower bounded as

12
 [Type the document title]

• To increase the system capacity using the acquired transmits CSI, up to

eigen channels can be used for transmitting multiple data streams.

Spatial Multiplexing
• Spatial multiplexing refers to breaking the incoming high rate data stream into M
parallel data streams.
• The standard mathematical model for spatial multiplexing is ..
• The size of the received vector y is Nrx1, the channel matrix H is Nr x Nt, the
transmit vector x is Nt x 1 and the noise n is Nr x 1.
• The transmit vector is normalized by Nt so that each symbol in x has average energy
.
• The channel matrix is of the form

• The entries in the channel matrix and noise vector are complex Gaussian and
independent and identically distributed (i.i.d.) with zero mean and diagonal
covariance matrices i.e. the spatial channels all experience uncorrelated Rayleigh
fading and Gaussian noise.
• Decoding Nt streams is possible when rank(H) ≥ Nt.

13
 [Type the document title]

Single user MIMO system

• The capacity or maximum data rate grows as
when the SNR is large.
• When SNR is low, the capacity grows linearly with .
• The average SNR of all Nt streams can be maintained without increasing the total
transmit power relative to SISO system.

1) Open Loop MIMO: Spatial Multiplexing

• The open loop techniques for spatial multiplexing attempt to suppress the interference
that results from all Nt streams being received by each of Nr antennas.

14
 [Type the document title]

i) Optimum Decoding: Maximum Likelihood Detection

• If the channel is unknown at the transmitter, the optimum decoder is the maximum
likelihood decoder which finds the most likely input vector via a minimum distance
criterion:
• To compute this, an exhaustive search must be done over allpossible input vectors
where M is the order of modulation.
• When optimum detection is achievable, the gain from transmitter channel knowledge
is fairly small.

ii) Linear Detectors

• Zero forcing (ZF) detector sets the receiver equal to the inverse of the channel
whenNt=Nr or .
• Zero forcing detector completely removes the spatial interference from the transmitted
signal andgives an estimated received vector .
• inverts the eigen values of H because of which the bad spatial sub channels can
severely amplify the noise in n.
• MMSE receiver attempts to strike a balance between spatial interference suppression
and noise enhancement by minimizing the distortion
• It is derived using the well knownorthogonality principle as .

15
 [Type the document title]

• As SNR grows large, the MMSE detector converges to ZF detector.

BLAST
• The earliest known spatial multiplexing receiver was invented and prototyped in Bell
Labs and is called Bell labs Layered Space Time (BLAST).
• BLAST consists of parallel “layers” supporting multiple simultaneous data streams.
• The layers (sub streams) are separated by interference cancellation techniques that
that decouple the overlapping data streams.
• Two types of BLAST are Diagonal BLAST (D-BLAST) and Vertical BLAST (V-
BLAST).

1) D-BLAST
• D-BLAST groups the transmitted symbols into “layers” which are then coded in time
independently of the other layers.
• These layers are then cycled to the different transmit antennas in a cyclic manner
resulting in each layer being transmitted in a diagonal of space and time.
• Each symbol achieves diversity in time via coding and in space since it rotates among
all the different antennas.
• The diagonal layered structure of D-BLAST can be detected by decoding one layer at
a time.

16
 [Type the document title]

• Each layer is detected by nulling the layers the layers that have not been detected and
cancelling the layers that have already been detected.
• The time domain coding helps compensate for errors or imperfections in the
cancellation and nulling process.
• The drawbacks of D-BLAST are that the decoding process is iterative and complex
and the diagonal layering structure wastes space time slots at the beginning and end of
D-BLAST.

2) V-BLAST
• V-BLAST is simpler than D-BLAST and reduces inefficiency and complexity.
• Each antenna simply transmits an independent symbol stream.
• The post detection SNR for the ith stream is
where is the ith row of ZF or MMSE receiver G.
• The essence of V-BLAST is to combine a linear receiver with ordered successive
interference cancellation.
• Instead of detecting all Nt streams in parallel, they are detected iteratively.
• First, the strongest symbol is detected and then subtracted from the composite
received signal.
• Next, the second strongest signal is detected which now sees effectively Nt-2
interfering streams.
• In general, the ith detected stream experiences interference from only Nt-1 transmit
antennas so that by the time weakest symbol stream is detected, the vast majority of
spatial interference has been removed.
• V-BLAST decreases the required SNR by about 4dB and have spectral efficiencies
above 20bps/Hz.

17
 [Type the document title]

2) Closed Loop MIMO

i) Singular Value Decomposition (SVD) Pre coding and Post coding
• Generalized eigen value decomposition of the channel matrix H is
where U and V are unitary and Σ is a diagonal matrix of singular values.
• The decision vector d should be close to input symbol vector b and can be written
systematically as

• has the same variance as n.

• The singular value approach does not result in noise enhancement.
• The complexity of finding the SVD of an Nt x Nr matrix is on the order of
if .

18
 [Type the document title]

ii) Linear Pre coding and Post coding

• The received symbol for the ith sub channel can be expressed as

• are the singular values of H, andare the pre coder and post coder weights and is
the noise per sub channel.
• The number of sub channels is bounded by .

19