Dr.

Anubhuti Khare, Manish Saxena, Vijendra singh mandloi / IOSR Journal of Engineering
(IOSRJEN) www.iosrjen.org ISSN : 2250-3021

Vol. 2 Issue 1, Jan.2012, pp.166-169
www.iosrjen.org 166 | P a g e
Performance Analysis of V-Blast Based MIMO-OFDM System
with Various Detection Techniques

Dr. Anubhuti Khare
1
, Manish Saxena
2
, Vijendra singh mandloi*
3

1. Dr. Anubhuti Khare, Reader, Department of Electronics and Communication, University Institute
Of Technology, Rajeev Gandhi Technical University, Bhopal.
2. Manish Saxena, Head of Electronics and Communication Department, Bansal Institute Of Science
And Technology Bhopal.
3*Vijendra singh mandloi, Student, Mtech (Digital Communication), Bansal Institute of Science
And Technology Bhopal.


Abstract
This paper presents the performance analysis of V-
BLAST based multiple input multiple output
orthogonal frequency division multiplexing (MIMO-
OFDM) system with respect to bit error rate per
signal to noise ratio (BER/SNR) for various detection
techniques viz zero forcing (ZF), minimum mean
square error (MMSE) and maximum likelihood (ML).
A 2X2 MIMO-OFDM system is used for the
performance evaluation. The simulation results shows
that the performance of V-BLAST based detection
techniques is much better than the conventional
methods.
Keywords- Vertical Bell Labs Layered Space-
Time (V-BLAST); multiple input multiple output
(MIMO); orthogonal frequency division multiplexing
(OFDM); bit error rate (BER); signal to noise ratio
(SNR).

I. INTRODUCTION
With the increasing demand for high speed
wireless transmission of data, the demand for vast
frequency spectrum arises. Since the available
frequency spectrum is scarce, future systems should
be characterized by significantly enhanced spectral
efficiency in order to increase link throughput and
network capacity. By using multiple antennas at both
transmitter and receiver (MIMO system) the
throughput can be increased by simultaneously
transmitting different streams of data on the different
transmit antennas but at the same carrier frequency
[1]. These parallel streams of data can be recovered at
the transmitter with spatial sampling and
corresponding signal process algorithm [2]. The
combination of the throughput enhancement of
MIMO system with the robustness of orthogonal
frequency division multiplexing (OFDM) against
frequency-selective fading caused by severe
multipath scattering and narrowband interference is
regarded as a very promising basis for future high

speed data communication. Thats why there is
high demand for a MIMO-OFDM system with high
efficiency and low complexity.

Among conventional detection techniques in
MIMO-OFDM communication such as zero forcing
(ZF) [3], minimum mean square error (MMSE) [4]
and maximum likelihood (ML) [5], ML shows better
performance at the cost of higher complexity [6].
Therefore we have applied V-BLAST algorithm on
all of the above detection techniques and then
compared their performance on the basis of bit error
rate per signal to noise ratio and found that the
detection techniques with V-BLAST algorithm shows
better performance than without V-BLAST. Even the
complexity of the system reduces by using V-BLAST
algorithm and the simulation results show that the
performance of V-BLAST based MMSE is close to
that of ML technique, with much lower complexity.
Therefore V-BLAST based detection techniques can
be preferred over conventional methods in MIMO-
OFDM system with high data transmission capacity
where high efficiency is required with lower
complexity. The analysis presented in this paper
shows the performance comparison of conventional
detection techniques in MIMO-OFDM
communication with that of V-BLAST based
techniques and hence shows that better performance
can be achieved by using V-BLAST based detection
techniques.
The remaining paper is organized as follows. In
section II we present the V-BLAST based system
model. Section III presents the performance
comparison of various detection techniques with their
V-BLAST counterpart. Simulation results are
presented in section IV followed by the conclusions
in section V.
II. V-BLAST BASED SYSTEM MODEL
A V-BLAST based 2X2 MIMO-OFDM system is
presented in figure (1). We considered N
T
transmitting antennas and N
R
receiving antennas. The
incoming signal bits are multiplexed and then
modulated with BPSK modulation [7]. The
modulated signal is processed through inverse
discreet Fourier transform (IDFT) to insert a cyclic
prefix and convert it to a radio frequency and then
transmit through channel [8].
Dr. Anubhuti Khare, Manish Saxena, Vijendra singh mandloi / IOSR Journal of Engineering
(IOSRJEN) www.iosrjen.org ISSN : 2250-3021

Vol. 2 Issue 1, Jan.2012, pp.166-169
www.iosrjen.org 167 | P a g e
At the receiver end, the signal is received through N
R
antennas and processed through discreet Fourier


Figure (1): A 2X2 MIMO OFDM system with V-BLAST detector

transform (DFT) and then passed through V-BLAST
detector. The received signal is passed through DFT
to convert it to frequency domain. Hence the equation
of received signal at first antenna is,
1
1 1,1 1,2 1
2
[ ] (1)
x
y h h n
x
(
= +
(


and the received signal at the second receive antenna
will be,
1
2 2,1 2,2 2
2
[ ] (2)
x
y h h n
x
(
= +
(


So, the equivalent equation for the received signal
can be written as,
(3) y Hx n = +
where x denotes the transmitted signal and n denotes
the additive white Gaussian noise(AWGN) added to
the signal.
The wideband MIMO channel is converted into a
set of N parallel narrow band channels from (3). Now
the V-BLAST algorithm is applied per subcarrier
basis [9]. In V-BLAST a filter matrix G is multiplied
to the signal and then detected layer by layer. The
channel with maximum SNR is then selected and the
effect of channel is nullified for that stream. Now the
nullified symbols are quantized with appropriate
constellation and the effect of detected stream is
cancelled from the received signal. A new channel
matrix is created then by replacing the columns
corresponding to the detected stream with zeros and
the whole process is repeated with the new channel
matrix until all the streams are detected.
The algorithm for the whole process can be given
as,
Finding nulling solution:
11 12
21 22
(4 )
g g
a
g g
| |
=
|
\ .
G
Ordering:
2 2
2
{|| || ,|| ( ) || } (4b)
1
= ( ) G G
( ) (4c) k min =
Nulling: Assuming k=2, then the nullified vector g
can be written as,
21
22
(4d)
g
g
| |
=
|
\ .
g
Nullifying the effect of channel from signal,
3
y = (4e)
T
g r
Quantizing:
2 2
( ) (4f) x Quant y =
Cancelling: Cancel the effect of detected stream from
the received signal,
21
2
22
(4g)
h
x
h
| |
=
|
\ .
r r
Finding new channel matrix:
11
12
0
(4h)
0
h
h
| |
=
|
\ .
H
Iteration: Now repeat from step 1 to get the first
stream.
III. COMPARISON
A. Zero Forcing
Zero Forcing is a linear detection technique. The
pseudo inverse of the signal is applied to the received
signal in order to make a decision about one user. So
the equation for filter matrix G for zero forcing will
be,
( )
-1
(5)
H H
G = H H H
In this way the received signal is detected by zero
forcing detectors. If V-BLAST algorithm is applied
on ZF detector, equation 4 will be applied on ZF filter
matrix. Figure 2 shows the performance comparison
of ZF and ZF-V-BLAST. As it is evident from the
figure, ZF with V-BLAST shows better performance
in comparison to normal ZF in terms of BER/SNR.


Multiplexor

V-BLAST
detector
IDFT
IDFT
DFT
DFT

n1
n2
Dr. Anubhuti Khare, Manish Saxena, Vijendra singh mandloi / IOSR Journal of Engineering
(IOSRJEN) www.iosrjen.org ISSN : 2250-3021

Vol. 2 Issue 1, Jan.2012, pp.166-169
www.iosrjen.org 168 | P a g e
B. Minimum Mean Square Error
MMSE is also a linear detection technique but
more reliable than ZF in case of noisy channel [10].
MMSE does not apply pseudo inverse of signal to
make decision about one user, instead it attenuates
them to noise level thereby reducing the diversity
order. From [4] the filter matrix for MMSE is,
1
(6)
r
H H t
N
N
I
SNR

| |
|
\ .
G = H H+ H
Now V-BLAST algorithm can be applied over above
filter matrix and results can be generated. Figure 3
shows the performance comparison of MMSE
detector and MMSE-V-BLAST detector in terms of
BER/SNR.
C. Maximum Likelihood
ML is a non-linear detection technique. The
BER/SNR results of ML are better than MMSE
detector but at the cost of additional complexity [6].
So ML is used in applications where high efficiency
is requires. Now if we apply V-BLAST algorithm on
ML, the performance will be better than ML detector
which is clear from figure (4).
IV. SIMULATION RESULTS
We have considered a 2X2 MIMO-OFDM
channel with BPSK modulation and various detection
techniques are applied along with V-BLAST
algorithm and their performance is compared on the
basis of bit error rate per signal to noise ratio
(BER/SNR). The simulation result shows that when
the detection techniques are applied with V-BLAST,
they show better results even for higher values of
SNR. Figure (5) shows the performance comparison
of all detection techniques along with V-BLAST. As
it is evident from the figure, we can obtain close to
ML performance with less complexity by applying V-
BLAST algorithm on linear detection techniques and
also if we require a system with performance better
than ML then we can also apply V-BLAST on ML.
V. CONCLUSION
We have analyzed the performance of V-BLAST on
various detection techniques and compared them on
the basis of BER/SNR and hence obtained better
performance without additional complexity. We can
also evaluate which system can be best suited for a
particular application on the basis of given
comparison. We can also compare the given
techniques on the basis of other parameters such as
symbol error rate (SER), computational time and
system configuration to obtain enhanced results.
VI. ACKNOWLEDGEMENT
Mrs. Anubhuti Khare one of the authors is indebted to
Director UIT RGPV Bhopal for giving permission for
sending the paper to journal. Manish Saxena is also
thankful to the Chairmen, Bansal Institute of Science
& Technology Bhopal for giving permission to send
the paper for publication. Last but not least Vijendra
Singh Mandloi is also thankful to the HOD Mr.
Manish Saxena Sir & Chairmen of BIST Bhopal for
giving permission to send the paper for publication.

REFERENCES
[1] S. Cui, A. J. Goldsmith, and A. Bahia (August,
2004). "Energy-efficiency of MIMO and
Cooperative MIMO in Sensor Networks". IEEE
journal Select. Areas of Communication. 22 (6):
10891098.
[2] Sam P. Alex and Lousy M.A. Jalloul,
"Performance Evaluation of MIMO in
IEEE802.16e/WiMAX", IEEE J. of Selected
Topics in Signal Processing, VOLUME 2,
NUMBER 2, April, 2008.
[3] R. Xu and F. C. M. Lau, Performance analysis
for MIMO systems using zero forcing detector
over fading channels, IEE Proc.
Communications, vol. 153, no. 1,2, pp. 74 80,
February 2006.
[4] Al-Dhahir, FIR Channel-Shortening Equalizers
for MIMO ISI Channels, IEEE Trans.
Communication vol. 49, pp.213-218, Feb. 2001.
[5] Inkyu Lee and John M. Cioffi, Design of
Equalized Maximum-Likelihood Receiver,
IEEE Communications Letters, VOLUME 2, No.
1, January 1998.
[6] Xu Zhu and Ross D. Murch, Performance
Analysis of Maximum Likelihood Detection in a
MIMO Antenna System, IEEE Transactions on
Communications, VOLUME 50, NUMBER 2,
FEBRUARY 2002.
[7] Proakis, John G. (1995). Digital
Communications. Singapore: McGraw Hill.
ISBN 0-07-113814-5.
[8] B. Yang, Z. Cao, and K. Letaief, Analysis of
Low-Complexity Windowed DFT-Based MMSE
Channel Estimation for OFDM Systems, IEEE
Trans. Communication., VOLUME 49, pp.
19771987, Nov. 2001.
[9] W. Y an, S. Sun and Z. Lei, A low complexity
VBLAST OFDM detection algorithm for
wireless LAN systems, IEEE Communication.
Letter, VOLUME 8 , pp. 347-376, June 2004.
[10] H. V. Poor and S. Verdu, \Probability of error in
MMSE multiuser detection," IEEE Transactions
on Information Theory, VOLUME 43, pp. 858-
871, May 1997.
Dr. Anubhuti Khare, Manish Saxena, Vijendra singh mandloi / IOSR Journal of Engineering
(IOSRJEN) www.iosrjen.org ISSN : 2250-3021

Vol. 2 Issue 1, Jan.2012, pp.166-169
www.iosrjen.org 169 | P a g e
TABLE I BER/SNR VALUES FOR VARIOUS DETECTION TECHNIQUES
BER/SNR SNR
ZF 0.3250 0.1400 0.0420 0.0110 0.0020
ZF-VBLAST 0.1470 0.0670 0.0140 0.0100 0.0010
MMSE 0.2370 0.0950 0.0210 0.0040 0.0000
MMSE-VBLAST 0.1070 0.0230 0.0100 0.0020 0.0000
ML 0.2160 0.0490 0.0050 0.0000 0.0000
ML-VBLAST 0.0600 0.0110 0.0000 0.0000 0.0000


Figure (2): BER/SNR plot for ZF and ZF-V-BLAST

Figure (3): BER/SNR plot for MMSE and MMSE-V-BLAST

Figure (4): BER/SNR plot for ML and ML-VBLAST

Figure (5): BER/SNR plot for all detection techniques.