Engineering, 2010, 2, **-**
doi:10.4236/eng.2010.21004 Published Online January 2010 (
Copyright © 2010 SciRes. ENGINEERING
‘Fading is Our Friend!’: A Performance Comparison of
WiMAX-MIMO/MISO/SISO Communication Systems
Arya ROHAN1,2, Nicolas DAILLY1, Palash KUSUMWAL1,2
1ESIEE-Amiens, Ecole Supérieure d’Ingénieurs en Electrotechnique et Electronique, France
2NIT-Trichy, National Institute of Technology, Trichy, India
Received August 24, 2009; revised September 22, 2009; accepted September 28, 2009
Research work for some time now has shown that fading wireless channels present enormous advantages if
properly exploited through a Multiple-Input Multiple-Output (MIMO) communication model. In this paper,
we demonstrate the advantages of implementing the MIMO communication model by investigating three
communication techniques, namely, Single-Input Single-Output (SISO), Multiple-Input Single-Output
(MISO) and MIMO for WiMAX communication systems. The performances of these communication tech-
niques are analyzed and compared for three scenarios - rural environment, TGV (high-speed train) environ-
ment and urban environment by using the models to investigate several communication parameters.
Keywords: MIMO, MISO, SISO, Wimax, Space-Time Coding, Alamouti Space-Frequency Block Coding
1. Introduction
In wireless communication systems, MIMO, pronounced
my-moh or mee-moh, refers to a link for which the trans-
mitting end as well as the receiving end is equipped with
multiple antenna elements. Like MIMO, MISO is an-
other smart antenna technology, but characterized by
multiple antennas only at the transmitting end. To under-
stand smart antenna technology, it is best to consider an
example in which, say, you are in a room. Someone in
the room is talking to you and, as he speaks, he begins
moving around the room. Your ears and brain have the
ability to track where the user's speech is originating
from as he moves throughout the room. This is similar to
how smart antenna systems operate. They locate the us-
ers, track them, and provide optimal RF signals to them
as they move throughout the base station's coverage area.
MIMO is rapidly becoming the face of smart antenna
technology. On the other hand, SISO, which has a single
antenna at both transmitting and receiving ends, is the
simplest and cheapest to implement among the three and
has been in use since the birth of radio technology.
MIMO promises to resolve the bottlenecks of traffic ca-
pacity in the forthcoming high speed wireless broadband
wireless internet access networks like Worldwide Inter-
operability for Microwave Access (WiMAX), 3G-Long
Term Evolution (see [3]) and beyond.
In this paper we have limited our analysis to the Wi-
MAX system and/or mobile-WiMAX system, which
were based on the IEEE 802.16-2004 standard and IEEE
802.16e-2005 standard respectively. In essence, WiMAX
is a 4G technology for a state-of-the-art ‘’last mile’’
telecommunication infrastructure (see [4,9]). WiMAX is
poised to replace a number of existing broadband tele-
communication infrastructure for wireless local loop,
while mobile-WiMAX can replace cellular networks.
There are several ways to implement MIMO systems,
such as, BLAST described by G. J. Foschini (see [1,2]),
space-time coding (see [5–7,10,11,13]) and more. How-
ever, we have stuck to the Alamouti space-time block
code proposed by Siavash Alamouti in 1998 (see [12]).
This code achieves transmit diversity by correlating the
transmit symbols spatially across the two transmit an-
tennas, and temporally across two consecutive time in-
tervals. The only condition is that the channel should
remain stationary over two consecutive symbols. Al-
though the Alamouti code achieves the same rate as
SISO, it attains maximum diversity for two transmit an-
tennas. The greatest advantage it offers is that the coding
and the decoding mechanisms it symbolizes are re-
markably simple and equally effective. The code also
provides the lowest probability of error and implementa-
tion complexity among all MIMO implementation tech-
niques. At the receiver end we use Maximum Likelihood
(ML) detection technique which largely does an exhaus-
tive search among all received signals in order to find the
optimum received signal (see [7]). Importantly, the per-
formance of the Alamouti code depends on an accurate
estimation of the channel between the transmitter and the
receiver. Transmission of training symbols is used to
perform channel estimation (discussed in Section 2).
In this paper we have analyzed the performance of a
WiMAX-MIMO system (by means of Bit Error Rate
(BER)) vis-à-vis WiMAX-MISO and WiMAX-SISO
systems using the MATLAB simulation tool. Analyses
for the following scenarios have been performed.
Rural environment: We consider an environment with
no obstacles, hence no fading takes place. Also, the
transmitter and the receiver are in zero relative motion
Train à Grande Vitesse (TGV)/High-speed train envi-
ronment: We consider a doppler fading environment.
The transmitter is stationary and the receiver is sitting
in a TGV moving at its top speed of 574.8 km/hr.
Urban environment: We consider a static multipath
environment with a LOS link between transmitter and
receiver. Again, the transmitter and receiver are in
zero relative motion.
We will also compare the behavior of the three models
to varying SNR and number of input bits in all the three
environments. This paper is structured as follows. Sec-
tion 2 introduces the Alamouti space-frequency block
code; it’s encoding and decoding scheme and how it dif-
fers from the well-known Alamouti space-time block
code. In Section 3, we present our WiMAX- MIMO/
MISO/SISO communication models with a detailed de-
scription of the complete layout of each model. Section 4
is devoted to the results obtained from computer simula-
tions for different analyses performed to compare the
communication performances of the three models and
also proposes a hybrid model (part SISO - part MISO -
part MIMO) which can be implemented for rural envi-
ronments. Finally, Section 5 presents the conclusion.
2. Alamouti Space-Time Block Code
In this paper we have focused on the Alamouti coding
scheme, precisely the Alamouti space-frequency block
code which is a slight variant of the Alamouti space-time
block code (see [11]).
For implementing MIMO for WiMAX systems, we
have employed the desired diversity differently at the
reception and transmission. The reception employs
Alamouti space-time block code (STBC) while the trans-
mission employs an Alamouti space-frequency block
code (SFBC). The motivation behind such a variation is
that STBC requires the channel to be stationary over two
consecutive OFDMA symbols (also see [14–17]). How-
ever, in a fast-fading radio channel, this is
Figure 1. SFBC encoding scheme.
not always true. In SFBC, the coding is implemented
across two consecutive sub-carriers in the frequency do-
main and thus within the OFDMA symbol. This elimi-
nates the aforementioned handicap posed by STBC.
Figure 1 illustrates the encoding scheme for SFBC’s.
As clearly visible, the mapping scheme is designed in
such a way that the first antenna transmits the entire
symbol stream without any modification, also facilitating
the system to act as a SISO system provided antenna two
is switched off. However, it is assumed that two adjacent
sub-carriers in the frequency domain experience corre-
lated fading. This assumption holds in channels where
the delay spread is low enough for the resulting coher-
ence bandwidth to exceed twice the sub-channel spacing.
Also, this is the reason why SFBC cannot be used for
For transmitted symbols X1 and X2 the receiver an-
tenna obtains the received symbol r1 and r2 for a 2x1
MISO system as
 hz
 
 
 
where Z is the Additive White Gaussian Noise (AWGN)
and h1 and h2 are the channel coefficients.
The optimal estimates for h1 and h2 can be obtained by
linear processing at the receiver, and are given by
 
 
 
 
11 2222 12
zxzxz xz
These channel estimates can then be used to detect the
next pair of code symbols. After the next code symbols
are decoded, the channel estimate can be updated using
Copyright © 2010 SciRes. ENGINEERING
Copyright © 2010 SciRes. ENGINEERING
those decoded symbols (see [9]). When the channel
variation is slow, the receiver improves stability of the
decoding algorithm by averaging old and new channel
The decision for the two transmitted symbols or in
other words the estimate of the two transmitted symbols
according to ML estimation, ~X1 and ~X2, is given by:
hr hr
hr hr
Figure 2. Real Time SISO communication model.
Similarly, the above scheme can be extended to 2 re-
ceiver antennas and hence to Nr receiver antennas.
Communication Model
This section presents the communication model block
diagram of all the three models and talks about the nu-
ances of each of them. It is interesting to note that the
MIMO and the MISO models have a far more complex
implementation than the simple SISO model (Figures 5,
6 & 7).
Figure 3. Real Time MIMO communication model.
SISO communication systems are vulnerable to envi-
ronments characterized by problems caused by multipath
effects. Figure 2 illustrates a real-time model of a SISO
system. On the other hand, the MISO transmission strat-
egy maximizes the received SNR by adding up the re-
ceived signal from all transmit antennas in-phase and by
allocating more power to the transmit antennas. MISO
wireless communication system exhibits transmitter di-
versity. Some of the transmitter diversity techniques in-
clude frequency weighting, antenna hopping, delay di-
versity and channel coding (see [8]). The real-time model
of a MISO system is similar to the one in Figure 3; how-
ever, there is only one antenna at the receiver end. The
MIMO system exhibits both transmitter diversity and
receiver diversity. While the transmitter diversity tech-
niques have already been discussed, some of the receiver
diversity techniques include selection diversity, antenna
diversity, maximal ratio combining and equal gain com-
bining (see [11]). Figure 3 illustrates a real-time model
of a MIMO system.
Figure 4. 200×200 black&white input image.
We have used an in-built MATLAB function ‘qam-
mod (data, index) to perform 16-QAM and similarly
‘qamdemod (data, index)' to perform 16-QAdeM. The
SFBC encoder and decoder are designed in accordance
with the equations already talked about in Section 2. The
characteristics of the models for the three environments
that we have used for our analysis are given below.
Rural environment: There is no ‘Multipath Rayleigh
filter ’block in the block diagram and the output of
‘peak power clipping’ goes to the ‘add Gaussian
noise’ block.
The advantages of using MIMO systems are increased
spectral efficiency, throughput, coverage, capacity, better
BER and resistivity to fading effects to name a few.
However, the greatest challenge it faces is the necessity
of complex DSP circuitry and the fact that its promise of
better communication performance hold true most only
for scattering-rich environments.
TGV environment: The ‘Multipath Rayleigh filter’
block is replaced by the ‘Doppler fading filter’ block.
We have used an in-built MATLAB function ‘’rayl-
eighchan (sampling freq, Doppler spread, ’path-de-
lays’, ’path-gains’)’’ to realize the doppler fading en-
vironment by setting Doppler spread as maximum of
1300 Hz, corresponding to 574.8 km/hr. Path delay is
set to zero.
Figures 5, 6 & 7 illustrate the block diagram for urban
environment for the WiMAX-SISO, WiMAX-MISO and
WiMAX-MIMO communication model discussed in this
paper. We have taken a 200x200 black & white image
(Figure 4) as the input to the communication system.
Urban environment: We again use the same in-built
MATLAB function mentioned in TGV environment to
realize the multipath fading environment. We have
taken a total of 15 multipath delays generated randomly.
Figure 5. WiMAX-SISO communication model for urban environment.
Figure 6. WiMAX-MISO communication model for urban evironment.
Figure 7. WiMAX-MIMO communication model for urban environment.
opyright © 2010 SciRes. ENGINEERING
Copyright © 2010 SciRes. ENGINEERING
Figure 8. Comparison of SISO, MISO & MIMO model for rural, TGV and urban environments.
The output is compared with the input data and the
BER is calculated. Also, importantly, it should be noted
that we have not used channel state information at the
transmitter (CSIT) for our analysis.
4. Simulation Results
4.1. SISO / MISO / MIMO Comparison in Rural,
Urban and High Speed Environments.
This section contains the results of simulations carried
out to compare the communication performance (BER)
of the three models in the aforementioned three envi-
ronments. Also, we analyze the performance of the three
models by varying the SNR and the number of input bits
for all the three environments.
For the first part of our analysis, we have fixed the
Eb/N0 value to be 3dB and compared the behavior of
the three communication models in all the three envi-
ronments. We have used the image in Figure 4 as the
input data.
Figure 8 illustrates the results thus obtained. Here ‘q’
Figure 9. Bervs SNR for SISO, MISO & MIMO models for rural, TGV and urban environments.
Rural environment
TGV environment
Urban environment
is the number of wrong bits when compared to the input
data. The following conclusions can be drawn from the
simulation results.
The image is least distorted for the MIMO model fol-
lowed by the MISO model and finally the SISO model
for all the three environments. Also, there is a great
difference between the MISO/MIMO model and the
SISO model especially for the TGV and urban envi-
ronment. However, there is not much difference be-
tween the MIMO model and the MISO model for any
environment since the only difference between them is
the inherent receiver diversity in the MIMO model.
Another notable point is that the Bit- Error Rate (BER)
for MIMO model decreases for urban and TGV envi-
ronment as compared to rural environment. However,
for the other two models, i.e. MISO model and SISO
model, the BER increases for urban and TGV envi-
ronment as compared to rural environment. However,
unlike in the SISO model, the increase is a slight one
for the MISO model. This result vindicates the point
that a fading environment improves the performance
of the MIMO model. Hence, fading is our friend!
opyright © 2010 SciRes. ENGINEERING
Rural environment
TGV environment
Urban environment
Figure 10. Ber vs. number of bits for SISO, MISO & MIMO models for rural, TGV and urban environments.
For the second part of our analysis, we have varied the
Eb/N0 value, hence the SNR value, and found out the
Bit-Error Rate (BER) value for all the three communica-
tion models each time. Again, we have used the image in
Figure 4 as our input data. We have then plotted the BER
vs. SNR curve for all the three models in the same graph.
Also, we have performed the analysis for all the three
environments. Figure 9 illustrates the results obtained
from the aforementioned simulation process. The fol-
lowing conclusions can be drawn.
The BER is highest for the SISO model and the least
for the MIMO model for all the three environments, as
can be inferred from the graph and the average BER
values from the tables.
It can also be inferred from the graphs that, as we in-
crease the SNR values, the BER decreases for the
SISO model in all the three environments with the ex-
ception of urban environment. However, for MISO
and MIMO models, as we increase the SNR value, the
BER decreases till a certain SNR value and then be-
come steady at a fixed value irrespective of the SNR.
The exception of the SISO model curve in urban en-
vironment can be attributed to the inability of the
model to counter multipath effects.
Another point to be noted is that, the average BER for
MIMO model decreases for urban and TGV environ-
ment on comparison to rural environment. But, this is
not the scenario for SISO model. Hence, we can infer
opyright © 2010 SciRes. ENGINEERING
that multipath and similar fading environments truly
help MIMO in its communication performance.
Variance is not the greatest parameter to be compared
for this analysis.
4.2. Influence of Data Size on Transmission
For the third part of our analysis, we have fixed the Eb/N0
value to 3 dB and varied the number of input bits and
found out the Bit-Error Rate (BER) for all the three
communication models each time. Here we have used
randomly generated input data for this analysis. We have
then plotted the BER vs. number of bits curve for all the
three models in the same graph. Again, we performed the
analysis for all the three environments. Figure 10 illus-
trates the results obtained from the aforementioned si-
mulation process. The following conclusions can be
The BER is highest for the SISO model and the least
for the MIMO model for all the three environments, as
can be inferred from the graph and the average BER
values from the tables.
Also, it can be inferred from the graphs that as we
increase the number of bits, the BER values remain
more or less constant for all the three models with the
MIMO model providing the best results in terms of
stability of the curve. Hence, we can infer that the
communication performance of all the three models is
irrespective of the number of bits.
The variance of MIMO model is the least for all the
three environments as compared to the SISO and
MISO model. This states that the MIMO model pro-
vides us with the maximum stability in the communi-
cation performance for the aforementioned analysis.
Also, it is interesting to note that the average BER for
the MIMO model decreases for urban and TGV envi-
ronments on comparison to rural environment. But,
this is not the case for SISO. This again vindicates the
point that multipath and other fading environments
prove to be favorable for MIMO’s communication
Finally, we can infer that fading environments prove
to be a big downfall for the SISO model as the BER
increases enormously for TGV and urban environ-
4.3. Hybrid Model for Mobile Power Saving
The purpose of this final part is to propose a hybrid
model where the mobile will switch between SIMO,
MISO and MIMO communications models, depending of
radio transmission performances. The aim is to obtain an
optimal radio performance over power consumption ratio.
Using multiple antennas at the MS will utilize much
power, which is a great source of concern as the MS has
limited battery power. Such a hybrid model will activate
antennas as a function of radio conditions.
To investigate this hybrid model, we have considered
the transmitted power and the received power for various
Eb/N0 values. Then, for each power value thus calculated
we have found the distance between the transmitter (base
station-BS) and the receiver (mobile terminal-MS) by
using the relation given below (also see [18])
The above expression is valid for flat-terrain mobile
communication environments and hence can be applied
for our rural environment. The values for Pr (received
power) and Pt (transmitted power) are found from the
model using MATLAB. We assume Gt and Gr to
be19dBi and 10dBi respectively, while ht and hr are as-
sumed to be 50m and 1.5m respectively. All the as-
Figure 11. Ber vs distance for SISO, MISO &MIMO. Figure 12. Ber vs distance for hybrid model.
Copyright © 2010 SciRes. ENGINEERING
sumed values are in accordance with the WiMAX stan-
dards (see [4,9]). We have then plotted the BER vs. dis-
tance curve for all the three models in the same graph, as
illustrate in Figure 11.
As a result, a hybrid model scenario for rural environ-
ment is proposed in Figure 12. Depending of the MS-BS
distance, the BER and the power consumption, the mo-
bile choose the optimal system: SISO, MISO or MIMO.
5. Conclusions
This paper shows the use of Alamouti space-frequency
block codes, a slight variant of the well-known Alamouti
space-time block code, to design MISO and MIMO
communication models for WiMAX systems for three
environments, namely, rural, TGV/high-speed train and
urban environment. The performances of the three mod-
els (SISO, MISO & MIMO) are compared for all the
three environments with MIMO model clearly surpassing
the other two models in every environment. This paper
also notifies the improvement in the performance of
MIMO systems in fading environments and also how
such environments prove to be a downfall for the other
models. The simulation results obtained from BER vs.
number of bits analysis confirm that the MIMO model
offers the maximum stability even if we have large input
data bits. While those obtained for the BER vs. SNR
analysis emphasize on the growing need for implement-
ing MIMO enhanced communications systems (in this
paper WiMAX system) especially for fading environ-
ments similar to urban and TGV (high-speed train) envi-
ronment discussed in this paper. Such a step, if taken,
will not only increase the coverage area of the commu-
nication system, but also allow for uninterrupted com-
munication service to be possible even at the edges of the
hexagonal cell. However, such an implementation would
increase the power consumption at the user end. To
counter this problem we have proposed our hybrid model,
as of now for the rural environment, where-in, the com-
munication system can switch from a SISO to MISO to
MIMO depending upon the communication parameters
(here BER). Hence this will ensure controlled power
consumption as well as good communication perform-
ance. However, more research in this direction needs to
be done especially at the various network layers (MAC
layer). Also, such a system needs to be expanded to the
urban environments as well.
. References
[1] G. J. Foschini, “Layered space-time architecture for
wireless communication in a fading environment when
using multi-element antennas,” Bell Laboratories Tech-
nical Journal, Vol. 2, pp. 41–59, October 1996.
[2] P. W. Wolniansky, G. J. Foschini, G. D. Golden, and R.
A. Valenzuela, “V-BLAST: An architecture for realizing
very high data rates over the rich-scattering wireless
channel,” URSI International Symposium on Signals,
Systems, and Electronics, 1998.
[3] E. Dahlman, A. Furuskar, Y. Jading, M. Lindstrom, and S.
Parkvall, “Key features if the LTE radio interface,” Erics-
son Review, No. 2, 2008.
[4] IEEE 802.16a Standard and Wimax Igniting Broadband
Wireless Access, White paper, WiMAX Forum, Septem-
ber 2003.
[5] D. Gesbert, University of Oslo, “Lecture to 5-th year
NTNU students,” September 2003.
[6] D. Gesbert, M. Shafi, D. Shiu, P. Smith, and A. Naguib,
“From theory to practice: An overview of MIMO space-
time coded wireless systems,” IEEE Journal on selected
areas of communication (JSAC), April 2003.
[7] M. Jankiraman, “Space-time codes and MIMO systems,
“Artech house Publishers, July 2004.
[8] D. Tse, P. Vishwanath, “Fundamentals of wireless com-
munication,” Cambridge University Press, June 2005.
[9] S. Ahson and M. Ilyas, WiMAX: Technologies, Per-
formance Analysis, and QoS, CRC, 1 Edition, pp. 41–96,
September 2007.
[10] G. Tsoulos, “MIMO system technology for wireless com-
munications,” CRC, 1 Edition, March 2006.
[11] B. Vucetic, J. H. Yuan, “Space-time coding,” Wiley, pp.
49–148, 245-264, June 2003.
[12] S. M. Alamouti, “A simple transmit diversity scheme for
wireless communications, IEEE Journal on selected areas
of communication (JSAC),” Vol. 16, pp 1451–1458, Oc-
tober 1998.
[13] V. Tarokh, H. Jafarkhani, and A. R. Calderbank,
“Space-time block codes from orthogonal design,” IEEE
Transactions on Information theory, Vol. 45, pp.
1456–1467, July 1999.
[14] H. Bolcskei and A. Paulraj, “Space-frequency codes for
broadband fading channels,” Proceedings IEEE Interna-
tional Symposium on Information Theory, pp. 219, 2001.
[15] B. Muquet, E. Biglieri, A. Goldsmith, and H. Sari, MIMO
Techniques for Mobile Wimax Systems, SEQUANS
Communications White Paper.
[16] Mai Tran, Angela Doufexi, and Andrew Nix, Mobile
WiMAX MIMO Performance Analysis: Downlink and
Uplink, Centre for communication research, University of
Bristol, 2008, IEEE.
[17] A. A. Hutter, S. Mekrazi, B. N. Getu, and F. Platbrood,
“Alamouti-based space-frequency coding for OFDM,
Wireless Personal Communication Review, Vol. 35, pp.
173–185, October 2005.
[18] T. S. Rappaport, Wireless Communication Principles and
Practices, Prentice Hall PTR, July, 2002.
opyright © 2010 SciRes. ENGINEERING