Multiuser MIMO Channel Estimation

G.Indumathi et al. Int. Journal of Engineering Research and Applications www.ijera.com
ISSN: 2248-9622, Vol. 6, Issue 5, (Part - 7) May 2016, pp.52-56
www.ijera.com 52 | P a g e
Multiuser MIMO Channel Estimation
G.Indumathi1
, M.Indiradevi2
1,2
Electronics and Communication Engineering,MepcoSchlenk Engineering College,Sivakasi, India
ABSTRACT
In this paper, three beamforming design are considered for multi user MIMO system. First, transmit
beamformers are fixed and the receive (RX) beamformers are calculated. Transmit beamformer (TX-BF)is
projectedas a null space of appropriate channels. It reduces the interference for each user. Then the receiver
beamformer is determined which maximize the SNR. This beamforming design provides less computation time.
The second case is joint TX and RX beamformer for SNR maximization. In this transmitter and receiver
beamformer are calculated using extended alternating optimization (EAO) algorithm. The third one is joint
transmitter and receiver beamforming for SNR and SINR maximization using EAO algorithm. This algorithm
provides better error performance and sum rate performance. All the design cases are simulated by using
standard multipath channel model. Our simulation results illustrate that compared to the least square design and
zero forcing design, the joint TX and RX beamforming design using EAO algorithm provides faster
beamforming and improved error performance and sum rate.
Keywords: Beamforming, Interference channel, Massive MIMO, Optimization
I. INTRODUCTION
Massive MIMO is an antenna array systems
using huge amount of antenna.In Massive MIMO,
Base stations are equipped with very large number
of antenna, so that each BS antenna simultaneously
serves many tens of user equipment [1]. The benefits
of Massive MIMO include reduced latency, and
robustness to interference and intentional
jamming.Beamforming means transmitting and
receiving the signal in specific direction and
beamforming is used to control the directionality of
transmitter and receiver on a transducer array. It is
used to reduce the interference in communication
systems.Threebeamforing design are
considered,where the transmit and receive
beamformers are calculated.In the first case,
thetransmit (Tx) beamformers are fixed and the
receive (Rx) beamformers are calculated. The
secondcase is jointly optimizing the Transmitter and
Receiver beamformers for constrained SNR
maximization.The third one is joint Tx–Rx
beamforming for SNR and signal-to-interference-
plus-noise ratio maximization.All cases can include
a linear constellation precoder design for extracting
multipath diversity. Linear constellation precoder
design matrix improves the error performance and it
has less computational complexity.
[7], [8], MEA offers large increase in
capacity compared to the single antenna system. The
capacity of multielement antenna is the sum of
individual sub channel capacities and the Multi
element antenna (MEA) systems use diversity at
both transmitter and receiver in which the large
capacity growth occurs even if the transmitter has no
information of the channel. Fading correlation and
multi element capacity is determined using abstract
model. In this model the essential channel
Characteristics can be clearly illuminated. The
advantage of using this model is simple and intuitive
model [9].
The multi-user massive MIMO systems
exhibit a hidden joint sparsity structure in the user
channel matrices due to the shared local scatterer in
the physical propagation environments. Here
compressed sensing technique is used to reduce
training sequence as well as feedback overhead in
the channel state information at transmitter side
(CSIT) and a joint orthogonal matching pursuit
recovery algorithm is used to perform the CSIT
recovery performance with the capability of
exploiting the concealed joint sparsity in the user
channel matrices [1], [11].
In [12]Multiple antennas are used at the
base station and at the user terminals. A double level
of spatial multiplexing is used, from thatthe users
are spatially multiplexed (SDMA) and each user
receive the spatially multiplexed bit streams (SDM).
Multi user interference is a potential performance
limitation in this scheme. To avoid this, we used a
pre-filter with a null space constraint that
blockdiagonalize the channel matrices, which results
reducing the multi user interference (MUI). Each
terminal then only has to reduce its own interstream
interference, which does not require information
from the other users’ channel. Joint TX-RX requires
more channel knowledge at both sides and a prefilter
with a null space constraint is used to cancel
multiuser interference completely.
RESEARCH ARTICLE OPEN ACCESS

II. SYSTEM DESCRIPTION
1.1 System Model
There are K pairs of multi-antenna
terminalswhich are striving to share simultaneously
the spectrumin time and space. The channel is
modeled as 1) a tapped delayline (L + 1 taps)
according to the IEEE 802.11n propagationmodel or
2) a single-tap flat fading channel with a perfect
spatial correlation matrix. The Kusers all have tM
transmit antennas and rM receive antennas, and all
users utilize each of the Psubchannels. The transmit
beamformer for
th
i user at
th
P subcarrier are
written as
1M
i
t
C)p(v 
 and similarly the receiver
beamformer are
1M
i
r
C)p(u 
 for i{1……..K}
and p{0……P-1}.
The frequency selective channel from the
μth transmit antenna of the
th
i transmitting user to
the
th
v receive antenna of the
th
r receiving user is
denoted by the delay-time function )l(h
i,r
v where
v  {1……. rM }, }M.......1{ t and r
)K....1( .The received signal vector for the
th
i
user represented by
)p(n)p(s)p(v)p(H)p(s)p(v)p(H)p(y iii
K
ii
i.iiii,ii ',
'
'  

(1)
From Kronecker model
2/Ti
ll
2/1r
l
i,r
l )R(G)R(H  is the
th
l channel
tap matrix where   ,v
i,r
l
i,r
v ]H[)l(h and
tr NN
l CG 
 is a complex matrix with zero mean
and unit variance Gaussian entries. where
r
lR
represent receive spatial correlation matrix and
i
lR
represent transmit spatial correlation matrix. For a
special type of uniform linear antenna array, the
correlation of the fading between two antennas
spaced ,vd apart can be modeled by
)D(jR)D(R XY
l
XX
ll,r
,v  
(2)
Where   /d2D ,v and
 


d)(f))sin(Dcos()D(R l
XX
l
 


d)(f))sin(Dsin()D(R l
XY
l
Here  is the wavelength, and )(f l
 is
the probability density function for the power
azimuth spread (PAS).
III. BEAMFORMING DESIGN
3.1 ReceiverBeamforming design for SNR
maximization
In this transmit beamformer are fixed. The
transmit beamformeris given by
(3)
Where iv represent transmit beamformer for
th
i
user.The next step to determine iu that it maximizes
the signal to noise ratio of the
th
i user. Then
receiver beamformer are calculated using the given
equation
(4)
Finally calculating the SNR using transmit
and receive beamforming.SNR for
th
i user is given
by
3.2 Joint TX-BF and RX-BF for SNR
maximization
In this, joint Tx-BF and Rx-BF are
designed for the constrained SNR maximization by
using the extended alternating optimization (EAO)
algorithm for multi objective optimization.
Transmit beamformer for
st
1 user is
denoted by
121
H
11 v)GG(vJ 
(5)
Where
])vH......()vH[(C H
KK,1
H
22,11 
Then finally receiver beamformer are
calculated by using transmit and receiver
beamformer. The receiver beamformer for
st
1 user is
given by
)HvvH(wu
H
1,1
H
111,1max1 
(6)
Then SNR calculation for
st
1 user is given by
)H(Nv i,i1Ki 
||vH||
vH
u
ii,i
ii,i
i 
i
H
ini
2
i
H
ii
H
iiii
H
i
i
uu
uHvvHu
SNR


1,1
H
1111
H
1,11 H)]C(N[)]C(N[HG 
1,1
H
2121
H
1,12 H)]C(N[)]C(N[HG 

SNR calculation for
th
i user is denoted by
(7)
Where iv represent transmit beamformer
for
th
i user and iu represent receiver beamformer
for
th
i user and i,iH represent channel coefficients.
3.3 LS TX-BF Design
The TX-BF design with least square is
formulated as
H
iiK.....1ii ttmaxv 
The TX-BF design with LS reduces to
(8)
Where }ji{H , the cofactor of H is, denotes
the submatrix of H obtained by deleting row j and
column I of H and iv represent the transmit
beamformer for
th
i user , iH represent channel
coefficients. Then calculating the receiver
beamformer with the help of transmitbeamformer.
Receiver beamformer for
th
i user is given by
3.4 Zero forcing Transmitter beamforming
design
The transmit beamforming design for zero
forcing is given by
1HZF
)HH(Hv 
 (10)
Where H denotes the KMt  channel
matrix between BS and all users and tM represents
number of transmitter antenna. In zero forcing
beamforming the transmit beamformer is calculated
by
1HH
))s(H)s(H()s(H
1
)s(v 


(11)
Where H(s) represents concatenated channel vector
for the set of scheduled users and
1H
])s(H)s(H[(tr
P
1 
 andtr(.) is the trace
operator and P represent average transmit power.
Where
iu represent receiver beamformer for
th
i
user. After calculating receiver beamformer then
calculating the signal to noise ratio. SNR calculation
for
th
i user is given by
(9)
IV. RESULTS AND DISCUSSION
4.1 Performance measures
For simulation, the number of users are
assumed to be K=11 andnumber of transmitter
antenna ( tM ) per user is K+1and number of
receiver antenna ( rM ) per user is K. so that number
of transmitter antenna is 132 and number of receiver
antenna is 121.In simulation all user uses QPSK
modulation for measuring the BER performance.
Fig.1.BER Performance for receiver beamforming
Fig 1 shows BER Performance for receiver
beamforming design. From the graph, it is observed
that if the signal to noise ratio increases then the bit
error rate decreases. From these the signal to noise
H
1
H
1,1
H
111,111 uHvvHuSNR 
H
i
H
i,i
H
iii,iii uHvvHuSNR 




K
ji
1j
2
}ji{i2
i
k...1ii |)Hdet(|
|)Hdet(|
1
maxv
||vH||
vH
u
ii,i
ii,i
i 
i
H
i
2
ni
i
H
ii
H
iiii
H
i
i
uu
uHvvHu
SNR



ratio is maximum and the computation time is
minimum.
Fig. 2.Receiver beamforming design for SINR
From Fig 2 shows if the signal to
interference noise ratio increases then the error
decreases and if the average SINR is 13 db the error
will be minimum.
Fig.3.sum rate performance for receiver
beamforming
4.2 BER Performance for different beamforming
design
Beamforming means transmitting and
receiving the signal in specific direction and it is
used to reduce the interference and improve the
communication quality.
Fig. 4.Comparison of BER Performance
Fig 4 shows an BER performance for
different beamforming design using Extended
alternating optimization algorithm. From the graph,
it is observed that if the signal to noise ratio
increases then the bit error rate decreases and the
extended alternating optimization algorithm
provides better error performance than the least
square and zero forcing beamforming design and it
has less computation time compared to the least
square design and zero forcing transmit
beamforming design.
4.3 Sum rate Performance for different
beamforming design
Sum rate is a maximum aggregation of all
user data rate.From this graph it is observed that if
the signal to noise ratio increases then the sum rate
increases and the extended alternating optimization
algorithm provides better sum rate than the least
square and the zero forcing transmitter
beamforming design.
At 20 db SNR the extended alternating
optimization algorithm provides the sum rate of
30.96 bits/s/Hz.The computation time of extended
alternating optimization algorithm is observed to be
approximately five times faster than the least
square and zero forcing design for the same set of
specification.
Fig. 4.Comparison of sum rate performance
V. CONCLUSION
Three beamforming designare performed
for multiuser MIMO interference channels where
transmit and receive beamformers are obtained
iteratively. In first case the transmit beamformer are
fixed and receive beamformer are calculated. The
second case is jointly optimizing the transmit and
receive beamformer for SNR maximization. Then
the third case is joint transmit and receive
beamformer for SNR and SINR maximization using
Extended alternating optimization algorithm. The
proposed optimization algorithm provides better
sumrate and bit error rate performance and the
computation time of EAO algorithm is less
compared to the least square transmit beamforming
design. At maximum SNR the extended alternating
optimization algorithm provides better sum rate
performance. The computation time of least square

transmit beamforming is large compared to extended
alternating optimization algorithm and the execution
time of extended alternating optimization
(EAO)algorithm simulation is observed to be
approximately five times faster than the Least square
design for the same set of specification. At 20 db
SNR the extended alternating optimization
algorithm provides the sum rate is 30.98 bits/s/Hz.
REFERENCES
[1]. A. Adhikary, J. Nam, J.-Y. Ahn, and G.
Caire, “Joint spatial division and
multiplexing: The large-scale array
regime,” IEEE Trans. Inf. Theory, vol. 59,
no. 10, pp. 6441–6463, Oct. 2013.
[2]. S. Nguyen and A. Ghrayeb, “Precoding for
multicell MIMO systems with compressive
rank-q channel approximation,” in Proc.
IEEE PIMRC, London, U.K., Sep. 2013,
pp. 1227–1232.
[3]. J. P. Vila and P. Schniter, “Expectation–
maximization Gaussian-mixture
approximate message passing,” IEEETrans.
Signal Process., vol. 61, no. 19, pp. 4658–
4672, Oct. 2013.
[4]. H. Q. Ngo, E. G. Larsson, and T. L.
Marzetta, “The multicell multiuser MIMO
uplink with very large antenna arrays and a
finite-dimensional channel,”IEEE Trans.
Wireless Commun., vol. 61, no. 6, pp.
2350–2361, Jun. 2013.
[5]. J. Choi, D. J. Love, and P. Bidigare,
“Downlink training techniques for FDD
massive MIMO systems: Open-loop and
closed-loop training with memory,” IEEE
J. Sel. TopicsSignal Process., vol. 8, no. 5,
pp. 802–814, Oct. 2014.
[6]. T. L.Marzetta, “Noncooperative cellular
wireless with unlimited numbers of base
stations antennas, ” IEEE Trans. Wireless
Commun., vol. 9, no. 11, pp. 3590–3600,
Nov. 2010.
[7]. S. Rangan, “Generalized approximate
message passing for estimation with
random linear mixing,” in Proc. IEEE ISIT,
St. Petersburg, Russia, Aug. 2011, pp.
2168–2172.
[8]. W. U. Bajwa, J. Haupt, A. M. Sayeed, and
R. Nowak, “Compressed channel sensing:
A new approach to estimating sparse
multipath channels,” Proc. IEEE, vol. 98,
no. 6, pp. 1058–1076, Jun. 2010.
[9]. D. Shiu, G. J. Foschini, M. J. Gans, and J.
M. Kahn, “Fading correlation and its effect
on the capacity of multi-element antenna
systems,” IEEE Trans. Commun., vol. 48,
no. 3, pp. 502–513, Mar. 2000.
[10]. S. Ji, Y. Xue, and L. Carin, “Bayesian
compressive sensing,” IEEE Trans. Signal
Process., vol. 56, no. 6, pp. 2346–2356,
Jun. 2008.
[11]. X. Rao and V. K. N. Lau, “Distributed
compressive CSIT estimation and feedback
for FDD multi-user massive MIMO
systems,” IEEE Trans. Signal Process., vol.
62, no. 12, pp. 3261–3271, Jun. 2014.
[12]. A.Bourdoux and N. Khaled, “Joint TX–RX
optimisation for MIMO SDMA based on a
null-space constraint,” in Proc. IEEE VTC
Fall, 2002, vol. 1, pp. 171–174.
[13]. H. Sung, S.-H. Park, K.-J. Lee, and I. Lee,
“Linear precoder designs for K-user
interference channels,” IEEE Trans.
Wireless Commun., vol. 9, no. 1, pp. 291–
301, Jan. 2010.
[14]. Z. Liu, Y. Xin, and G. Giannakis, “Linear
constellation precoding for OFDM with
maximum multipath diversity and coding
gains,” IEEE Trans.Commun., vol. 51, no.
3, pp. 416–427, Mar. 2003.
[15]. S. Peters and R. Heath, “Cooperative
algorithms for MIMO interference
channels,” IEEE Trans. Veh. Technol., vol.
60, no. 1, pp. 206–218, Jan. 2011.

Multiuser MIMO Channel Estimation

More Related Content

What's hot (18)

Similar to Multiuser MIMO Channel Estimation (20)

Recently uploaded (20)

Multiuser MIMO Channel Estimation