JOURNAL OF TELECOMMUNICATIONS, VOLUME 1, ISSUE 1, FEBRUARY 2010

Optimization of MIMO detectors: Unleashing

the multiplexing gain
Nirmalendu Bikas Sinha1, S.Chakraborty1 , P. K. Sutradhar1, R.Bera2 ,And M.Mitra3

Abstract— Multiple Input Multiple Output (MIMO) systems have recently emerged as a key technology in wireless communication
systems for increasing both data rates and system performance. There are many schemes that can be applied to MIMO systems such
as space time block codes, space time trellis codes, and the Vertical Bell Labs Space-Time Architecture (V-BLAST). This paper
proposes a novel signal detector scheme called MIMO detectors to enhance the performance in MIMO channels. , we study the
general MIMO system, the general V-BLAST architecture with Maximum Likelihood (ML), Zero- Forcing (ZF), Minimum Mean-
Square Error (MMSE), and Ordered Successive Interference Cancellation (SIC ) detectors and simulate this structure in Rayleigh
fading channel. Also compares the performances of MIMO system with different modulation techniques in Fading and AWGN
channels. Base on frame error rates and bit error rates, we compare the performance and the computational complexity of these
schemes with other existence model.Simulations shown that V-BLAST implements a detection technique, i.e. SIC receiver, based on
ZF or MMSE combined with symbol cancellation and optimal ordering to improve the performance with lower complexity, although ML
receiver appears to have the best SER performance-BLAST achieves symbol error rates close to the ML scheme while retaining the
low-complexity nature of the V-BLAST.

Index Terms— MIMO, V-BLAST, ML, ZF, MMSE and SIC.

——————————
——————————

1. INTRODUCTION

Future wireless communication networks will need to During recent years, various space-time (ST) coding schemes
support extremely high data rates in order to meet the have been proposed to collect spatial diversity and/or
rapidly growing demand for broadband applications such achieve high rates. Among them, V-BLAST (Vertical Bell
as high quality audio and video. Existing wireless Labs Layered Space-Time) transmission has been widely
communication technologies cannot efficiently support adopted for its high spectral efficiency and low
broadband data rates, due to their sensitivity to fading. implementation complexity [4]. When maximum-likelihood
Recent research on wireless communication systems has (ML) detector is employed, V-BLAST systems also enjoy
shown that using MIMO at both transmitter and receiver receives diversity, but the decoding complexity is
offers the possibility of wireless communication at higher exponentially increased by the number of transmit-
data rates, enormous increase in performance and spectral antennas. Although some (near-) ML schemes (e.g., sphere-
efficiency compared to single antenna systems. The decoding (SD), semi-definite programming (SDP)) can be
information-theoretic capacity of MIMO channels was used to reduce the decoding complexity, at low signal to-
shown to grow linearly with the smaller of the numbers of noise ratio (SNR) or when a large number of transmit
transmit and receiver antennas in rich scattering antennas and/or high signal constellations are employed,
environments, and at sufficiently high signal-to-noise (SNR) the complexity of near-ML schemes is still high. Some
ratios [1].MIMO wireless systems are motivated by two suboptimal detectors have been developed, e.g., successive
ultimate goals of wireless communications: high-data-rate interference cancellations (SIC), decision feedback equalizer
and high-performance [2],[3]. (DFE), which are unable to collect receive diversity [5]. To
further reduce the complexity, one may apply linear
————————————————
detectors such as zero-forcing (ZF) and minimum mean-
• Prof. Nirmalendu Bikas Sinha, corresponding author is with the
Department of ECE and EIE , College of Engineering & Management,
square error (MMSE) equalizers. It is well-known that linear
Kolaghat, K.T.P.P Township, Purba- Medinipur, 721171, W.B., India. detectors have inferior performance relative to that of ML
• S.Chakraborty is with the Department of ECE, College of Engineering & detector. However, unlike ML detector, the expected
Management, Kolaghat, K.T.P.P Township, Purba- Medinipur, 721171, performance (e.g., diversity order) of linear equalizers has
W.B., India.
• P. K. Sutradhar is with the Department of ECE, College of Engineering & not been quantified directly. The mutual information of ZF
Management, Kolaghat, K.T.P.P Township, Purba- Medinipur, 721171, equalizer has been studied in [6] with channel state
W.B., India. information at the transmitter.
• Dr. R. Bera is with the S.M.I.T, SikkimManipal University, Majitar,
Rangpo, East Sikkim, 73713.
• Dr. M.Mitra is With the Bengal Engineering and science University,
Shibpur, Howrah, India .
 

JOURNAL OF TELECOMMUNICATIONS, VOLUME 1, ISSUE 1, FEBRUARY 2010
2



     … . . 2 ,
    … . 3
In this paper, we propose a modified V-BLAST system,
which introduces different delay offsets for each substreme



in the transmitter. At the receiver, we can employ ZF
strategy to recover information and the introduction of

   

delay offsets enables the requirement of Nr to be relaxed to

  
     … 4 ,   . . 5
Nr ≥ 1 (in the conventional V-BLAST, Nr ≥ Nt. Where, Nr

  
and Nt are the receiver and transmitter antennas

   
respectively. We will verify the performance improvement
by theoretical analysis and simulation results. From our
analysis, with ZF decoding, the diversity order can reach Mr
in the modified V-BLAST system. But the increase of the The above wireless channel is modulated as y=Hx+n. Where
diversity order is at the cost of the multiplexing gain. H is the channel matrix and n is the channel noise. For
The main goal of our paper is to study the MIMO transmit/receive beam forming with the diversity of order
detectors schemes and quantify the diversity orders MN, is considered as full diversity. On the other hand the
collected by linear equalizers for V-BLAST. Also optimize antenna gain is; Max = {M,N}≤antenna gain ≤MN
the ultimate detector and modulation technique that yields a
better error performance than general V-BLAST.
The rest of this paper is organized as follows. In Section 2,
the MIMO system model is introduced. Section 3 gives the
performances of MIMO system with different modulation
techniques in Fading and AWGN channels and Section 4
gives the performance analysis of the linear equalizers
optimize the ultimate detector.

2. MIMO SYSTEM MODEL

Consider a multi-antenna system with M transmit-antennas

and N receive-antennas. Furthermore, as a commonly used
structure for the MIMO system, V-BLAST shares some basic
modules with our general multiple an
TX
RX

S1(t) TX1
Rx1 X1(t)
Scattering

S2(t) TX2 medium

(Rayleigh
fading Rx2 X2(t)
/AWGN)
channel

SM(t) TXM

RxN XN(t)

Fig.1: Diagrams of a single user MIMO communication

system using M transmit and N receives antennas.

Let M and N be the number of transmit and receive

is given by
  ∑ 

   … … . 1

antennas, respectively. The received signal in the ith antenna 3. PERFORMANCES OF MIMO SYSTEM WITH
DIFFERENT PROPAGATION CONDITION
where i = 1, 2, 3…N, hij is the fading corresponding to the Communication in a slow flat Rayleigh fading channel with
path from transmit antenna j to receive antenna i. ni is the AWGN is not reliable as the channel frequently enters into
noise corresponding to receive antenna i. deep fades, (i.e., the channel attenuation is large). More
specifically, as seen in class, Rayleigh fading converts an
exponential dependency of bit error probability on the
JOURNAL OF TELECOMMUNICATIONS, VOLUME 1, ISSUE 1, FEBRUARY 2010
3
signal-to-noise ratio (SNR) into an inverse relationship. For Fig.2 shows the plots for the Rayleigh and Gaussian
BPSK, QPSK and 16 QAM the symbol error rate in an densities. Where μ is the mean, and σ is the standard
AWGN channel is
2)* )*
deviation.

!"#$%  & '(  - … 6 , !/#$%  & '(  - … … 7

+, +,

3 )*
! 1/2  & '( - … … . 8
4 10+,

Where, &
  5678
⁄1.414 … … … … 9



However, with Rayleigh fading the average probability of

bit error is

!;"#$%   · =1 > ?@A B … 10, !;/#$%   · =1 > ?@AB … 11

@ @

)G
!; 1/2  D · =1 > ?@ADEB … 12, F · )I   … . . 13
C @
HE

Using the model (Fig.1), the M ×N MIMO communication has Fig.2: Gaussian and Rayleigh probability densities.
M*N channels. This leads to the expressions for the received
signal-to-noise ratios (SNRs) of each of these communications
in terms of the MIMO channel.

1 )*
0
SER IN DIFF. MODULATION IN AWGN AND RAYLEIGH CHANNEL
10

JHK LM    … … . 14

2 +,
SYMBOL ERROR RATE

-2

Where, H  ∑R
S ∑Q O O . ……. (15)
10

Let TU, V be the probability density function of MIMO

-4
10
QPSK
system using BPSK are given by, QAM
BPSK

X X
)G
!LM  W W & '( - !U, V sin U\U\V … 16
-6 R-BPSK
10

H
R-QPSK

E E E
R-QAM
0 5 10 15 20 25 30 35 40
Eb/No, dB

Where,  are the channels that are set-up in the MIMO Fig.3 : SER for different modulation in AWGN and

function, Em is the signal power while +, is the noise power

communication system and α is Rayleigh distribution
Rayleigh channel.
spectral density. Fig.3 shows clearly this severe degradation in the

Rician and Rayleigh [If ]^V   0 H_`a5 bc_5 ] probability

probability of bit error for BPSK in a Rayleigh fading
channel for E(α 2) = 1. As it can be seen, the performance
density functions respectively. degradation is as severe as in the BPSK case. The overall

6 6
system performance of MIMO system is better than other

76  5
T d> e … … … … 17
systems for different modulations due to its improvement of

V 2V 
transmit and receive diversity. It is noticeable that the

When  ]^V   is very large Eq.(8) becomes a Gaussian

overall system performance of QPSK is better compare to

probability density function of mean A and variance V  .

other modulation schemes.For all SNR levels, MIMO system
has the least SER, and hence enhance the diversity gain
because the lower the error in the received signals, the
1 6 > ]
76 f 5
T i> j … … … … 18
higher is the detection. But the overall system performance of

g2hV  2V 
Rayleigh channel is better in comparison with other
channels with respect to probability density function as
well as SER vs. SNR.
JOURNAL OF TELECOMMUNICATIONS, VOLUME 1, ISSUE 1, FEBRUARY 2010
4

4. AVERAGE BER ANALYSIS

4.1 Maximum Likelihood (ML): BER for QPSK in ML,STBC,ZF and MMSE detector (Rayleigh channel)

Theoretically maximum likelihood (ML) detection algorithm 10

-1

is the optimum method of recovering the transmitted signal

at the receiver. ML receiver is a method that compares the -2

Bit Error Rate

10
received signals with all possible transmitted signal vector
which is modified by channel matrix H and estimates -3
10
transmit symbol vector x according to the Maximum
Likelihood principle , which is shown as:

k  arg op qros ,ot …ou v w`x6 > 

y x …….(19)
-4 ML(nTx=2, nRx=2)
10
STBC(nTx=2, nRx=2)
ZF(nTx=2, nRx=2)
-5 MMSE(nTx=2, nRx=2)
10

where,
is the estimated symbol vector. Although ML
0 5 10 15 20
Average Eb/No,dB

detection offers optimal error performance, it suffers from

complexity issues. It has exponential complexity in the sense Figure 4, Performance curves for different linear detectors
that the receiver has to consider |A|M possible symbols for (ML,STBC, ZF, MMSE ) in 2×2 MIMO - V-BLAST system in
an M transmitter antenna system with A is the modulation a slow fading channel.
constellation.

4.2 V-BLAST Zero Forcing (ZF) ZF characteristic: CONCLUSION

Simple linear receiver with low computational In this paper, we analyze the performance of linear detectors
complexity and suffers from noise enhancement. It works for MIMO V-BLAST systems in slow fading channels for

k   z { 
  A
… 20 , Where  A represents
best with high SNR. The solution of the ZF is given by: different modulations and different channels, which exhibit
the best trade-off between performance and complexity
the pseudo-inverse. among spatial multiplexing techniques. We show that
conventional linear equalizers can only collect diversity N r
4.3 V-BLAST Minimum Mean Square Error (MMSE): – N t +1 for MIMO V-BLAST systems though they have very
low complexity.By investigating and simulating each
The MMSE receiver suppresses both the interference and receiver concepts, it was shown that V-BLAST implements a
noise components, whereas the ZF receiver removes only detection technique, i.e. SIC receiver, based on ZF or MMSE
the interference components. This implies that the mean combined with symbol cancellation and optimal ordering to
square error between the transmitted symbols and the improve the performance, although ML receiver appears to
estimate of the receiver is minimized. Hence, MMSE is have the best SER performance. In this paper, the MIMO
superior to ZF in the presence of noise. Some of the principle is based on a rich multipath environment without
important characteristics of MMSE detector are simple linear a normal Line-of-Sight (LOS) that is the Rayleigh flat fading
receiver, superior performance to ZF and at Low SNR, channel, due to movement or other changes in the
MMSE becomes matched filter. Also at high SNR, MMSE environment, LOS situation can arise. So finally we

1 {
becomes Zero-Forcing. MMSE receiver gives a solution of: proposed that ML detector for MIMO-V-Blast in slow fading

k  | ·
 = }~    B ·  
. . 21
JHK
channel with QPSK modulation is the ultimate optimization
technique in the next generation broadband communication
system.

1 {
1
At low SNR, MMSE becomes ZF:

= }~    B   f   … 22
JHK JHK
ACKNOWLEDGEMENT

Authors would like to thank to the contribution of Makar

1 {
At high SNR, MMSE becomes ZF:

}€    B   f   {   … 23

Chand Snai , Sumit Kumar Gupta and Manish Sonal

k  | ·
 =
JHK
pursuing B.Tech in the Department of Electronics &
Communication Engineering at College of Engineering
JOURNAL OF TELECOMMUNICATIONS, VOLUME 1, ISSUE 1, FEBRUARY 2010
5
and Management, Kolaghat, under WBUT in 2010, W.B, Management, Kolaghat. W.B, India where he is currently an
Asst.Professor is with the department of Electronics & Communication
India for their extensive hard work and sincere support
Engineering & Electronics & Instrumentation Engineering. His current
in completing this project. research Interests are in the area of signal processing for high-speed
digital communications, signal detection, MIMO, multiuser
REFERENCES communications,Microwave /Millimeter wave based Broadband
Wireless Mobile Communication ,semiconductor Devices, Remote
[1] I. E. Telatar, “Capacity of multi-antenna gaussian Sensing, Digital Radar, RCS Imaging, and Wireless 4G communication.
channels," Eur. Trans.Tel., vol. 10, pp. 585-595, November- He has published large number of papers in different international
December 1999. Conference, proceedings and journals.He is currently serving as a
member in international journal editorial board and reviewer for
[2] G. J. Foschini and M. J. Gans, “On limits of wireless
Wireless communication and radar system in different international
communications in a fading environment when using
journals.
multiple antennas,” Wireless Personal Commun., vol. 6, pp.
311-335, Mar. 1998. Sourav Chakraborty is pursuing B.Tech in the Department of
[3] V. Tarokh, N. Seshadri, and A. R. Calderbank, “Space-time Electronics & Communication Engineering at
codes for high data rate wireless communication: College of Engineering and Management,
performance criterion and code construction,” IEEE Trans. Kolaghat, under WBUT in 2011, West Bengal,
Inf. Theory, vol. 44, no. 2, pp. 744–765, Mar. 1998. India. His areas of interest are in Microwave
[4] G. D. Golden, G. J. Foschini, R. A. Valenzuela, and P. W. /Millimeter wave based Broadband Wireless
Wolniansky, “Detection algorithm and initial laboratory Mobile Communication and digital electronics.
results using V-BLAST space-time communication He has published some papers in different
architecture,” Electron. Lett., vol. 35, no. 1,pp. 14-16, Jan. 7, international journals.
1999.
[5] D. Tse and P. Viswanath, Fundamentals of Wireless
Communications.Cambridge, 2005. Prosenjit Kumar Sutradhar is pursuing B.Tech in the Department of
Electronics & Communication Engineering at
[6] Alamouti, S. M., “A simple transmit diversity technique
College of Engineering and Management,
for wireless communications,” IEEE Journal Selected Kolaghat, under WBUT in 2011, West Bengal,
Areas on Communications, Vol. 16, 1451–1458, Oct. 1998. India. His areas of interest are in Microwave
[7] Hassell, C. Z. W., J. S. Thompson, B. Mulgrew, and P. M. /Millimeter wave based Broadband Wireless
Grant,“A comparison of detection algorithms including Mobile Communication and digital electronics.He
BLAST for wireless communication using multiple has published some papers in different
international journals.
antennas,” IEEE Personal, Indoor and Mobile Radio
Commun.’, Vol. 1, 698–703, Sept.2000.
[8] T. Pande, D. J. Love, and J. V. Krogmeier, “Reduced Dr. Rabindranath Bera is a professor and Dean (R&D), HOD in Sikkim
feedback MIMO-OFDM precoding and antenna selection,” Manipal University and Ex-reader of Calcutta
IEEE Trans. Signal Process.,vol. 55, pp. 2284–2239, May University, India. B.Tech, M.Tech and
2007. Ph.D.degrees from Institute of Radio-Physics and
[9] H. Zhang, Y. Li, V. Stoplman, and N. V. Waes, “A reduced Electronics, Calcutta University. His research
areas are in the field of Digital Radar, RCS
CSI feedback approach for precoded MIMO-OFDM
Imaging, Wireless 4G Communications,
systems,” IEEE Trans.Wireless Commun., vol. 6, pp. 55–58, Radiometric remote sensing. He has published
Jan. 2007. T. Pande, D. J. Love, and J. V. Krogmeier, large number of papers in different national and
“Reduced feedback MIMOOFDM precoding and antenna international Conference and journals.
selection,” IEEE Trans. Signal Process.,vol. 55, pp. 2284–
2239, May 2007. Dr. Monojit Mitra is an Assistant Professor in the Department of
Electronics & Telecommunication Engineering of
[10] N. Wang and S. D. Blostein, “Minimum BER transmit Bengal Engineering & Science University,
optimization for two-input multiple-output spatial Shibpur. He obtained his B.Tech, M.Tech & Ph. D
multiplexing,” Proc. IEEE GLOBECOM, vol. 6, Dec. 2005. .degrees from Calcutta University. His research
areas are in the field of Microwave &
Microelectronics, especially in the fabrication of
high frequency solid state devices like IMPATT.
Prof. Nirmalendu Bikas Sinha received the B.Sc (Honours in He has published large number of papers in
Physics), B. Tech, M. Tech degrees in Radio- different national and international journals. He
Physics and Electronics from Calcutta University, has handled sponsored research projects of DOE
and DRDO. He is a member of IETE (I) and
Calcutta,India,in1996,1999 and 2001, respectively. Institution of Engineers (I)society.
He is currently working towards the Ph.D degree
in Electronics and Telecommunication
Engineering at BESU. Since 2003, he has been
associated with the College of Engineering and