LIST OF ACRONYMS
AWGN Additive White Gaussian Noise
BER Bit Error Rate
CSI Channel State Information
FDM Frequency Division Multiplexing
ICI Inter-Carrier Interference
ISI Inter-Symbol Interference
MLD Maximum Likelihood Decoding
M-PSK M-ary Phase-Shift Keying
MMSE Minimum Mean Square Error
MRC Maximum Ratio Combining
MRT Maximum Ratio Transmit
OFDM Orthogonal Frequency Division Multiplexing
PDF Probability Density Function
SISO Single Input Single Output
SNR Signal Noise Ratio
STBC Space-Time Block Code
TABLE OF CONTENT
LIST OF ACRONYMS i
TABLE OF CONTENT i
ACKNOWLEDGEMENTS iii
LIST OF FIGURES iv
i
vi
LIST OF TABLES vi
ABSTRACT 1
CHAPTER 1 2
MOBILE RADIO CHANNEL CHARACTERISTICS 2
1.1 Introduction 2
1.2 AWGN 3
1.3 Path loss 5

3.5 ISI in OFDM system 34
3.6 ICI in OFDM system 38
3.7 PAPR in OFDM system 41
ii
3.7.1 Clipping 43
3.7.2 Selected mapping 44
3.7.3 Partial Transmit Sequences 45
3.8 Conclusion 46
CHAPTER 4 46
COMBINED OFDM AND TRANSMIT DIVERSITY SYSTEMS 46
4.1 Introduction 47
4.2 OFDM combined with transmitter diversity 47
4.2.1 Delay approach 47
4.2.2 Permutation approach 49
4.2.3 Space-time coding approach 51
4.2.3.1 System description 51
4.2.3.2 Maximum likelihood detection 54
4.3 Conclusion 62
CONCLUSION 63
APPENDIX 63
REFERENCES 71
ACKNOWLEDGEMENTS
First of all, I would sincerely like to thank my supervisor, Doctor Tran Xuan
Nam for many discussion hours, valuable advice, and his continuous
iii
guidance.
I would also like to acknowledgement Associate Professor Nguyen Quoc
Binh for many useful and interesting information about wireless
communication. Thanks to lecturers in Military Technical Academy providing
me with full knowledge during 5 years.

Figure 3.6: OFDM symbol without cyclic prefix 36
Figure 3.7: OFDM symbol with cyclic prefix 36
Figure 3.8: OFDM-QPSK with Delay spread 37
Figure 3.9: Transmitted signal inserted guard interval 38
Figure 3.10: OFDM signal with cyclic prefix 39
Figure 3.11: Frequency offset error 40
Figure 3.12: Time error 40
Figure 3.13: PAPR in OFDM system 41
Figure 3.14: IBO and OBO 42
Figure 3.15: An example illustrates the clipped signal 43
Figure 3.16: Transmitter with clipping and filtering 44
Figure 3.17: Selected mapping 44
Figure 3.18: Partial Transmit Sequences 45
Figure 4.1: Delay transmit diversity 48
Figure 4.2: Permutation approach 50
Figure 4.3: Space time coding approach 51
Figure 4.6: STBC-OFDM over selective Rayleigh fading channel 58
Figure 4.7: The original image 58
Figure 4.8: Received images over flat fading channel using STBC-OFDM
60
v
Figure 4.9: Received images over flat and frequency selective fading
channel 62

LIST OF TABLES
Table 2.1: Alamouti parameters with BPSK constellation 26
Table 3.1: OFDM parameters for simulation 31
Table 3.2: OFDM parameters for simulation in channel with delay
spread 37
Table 4.1: Simulation parameters of STBC-OFDM system 55

1
This chapter introduces a combined approach of OFDM and transmit
diversity techniques to obtain both path and transmit diversities. Matlab
simulation is used to evaluate efficiency of the combined STBC-OFDM
approach.
CHAPTER 1
MOBILE RADIO CHANNEL CHARACTERISTICS
1.1 Introduction
In an ideal radio channel, received signal consists of only a single direct
path so it can be recovered perfectly at the receiver. In real channel, wireless
communication channel suffers from many impairments such as the thermal
2
noise, often modeled as Additive White Gaussian Noise (AWGN), path loss
in power, shadowing effects due to the presence of fixed obstacles in the
radio path, fading due to the effect of multi-path propagation, and Doppler
effect due to movement of mobile units. Consequently, signal copies undergo
different attenuations, distortions, delays and phase shifts. An example of
multi-path propagation in a wireless channel is illustrated in Figure 1.1. Due
to these problems, the overall system performance is degraded significantly.
Figure 1.1: An example of multi-path propagation in a wireless channel
1.2 AWGN
In practice, transmission is always effected by noise. The appearance of
noise reduces ability in detecting exact transmitted signal, so transmission
efficiency is reduced, too. Noise is resulted from many different sources, such
as thermal noise, noise of electronic devices, man-made noise and other
sources. Superposition of many independent processes, noise can be modeled
as a Gaussian distributed random process with white spectral density. The
popular noise model in communication system is Additive White Gaussian
3
Noise. This is a very good model for the physical reality as long as the

The noise is a zero mean Gaussian random process. This means that the
output of every noise measurement is a zero mean Gaussian random
variable that does not depend on the time instant when the measurement is
done.
Autocorrelation function of white noise which is described in Figure 1.2 (b),
is the inverse Fourier transform of the power spectral density given by
1 2
( ) { ( )} ( ).
j f
n n n
R F G f G f e df
π τ
τ
∞
−
−∞
= =
∫
0
( )
2
N
δ τ
=
(a)
(b)
4
It is seen that, the autocorrelation of white noise is a Dirac delta function It is
weighted by a factor
0

P
p d W m
d
π
=
where
( )p d
is power density at distance d,
t
P
is power density of the
isotropic radiator, and d is the distance between source and viewed point.
Since
2
4 d
π
is the area of sphere, the power extracted at receiver antenna
which is described in Figure 1.3, can be written as
2
( )
4
t
r r r
P
p p d A A
d
π
= =
Figure 1.3: An illustration of power density on sphere
Power density at the receiver when the transmitter antenna has gain

r
A
into equation (1.6), we can express the
receiver signal power in equivalent form
2
4
r t t r
P PG G
d
λ
π
 
=
 ÷
 

The path loss P
L
which expresses signal attenuation in decibels across entire
communication link, is defined as the difference between the transmitted
signal and the received signal, as shown by
2
10 10 10
4
10log 10log ( ) 10log
t
L t r
r
P d
P G G

proportional to the speed of mobile unit. Let us assume that, we have a signal
with a frequency
c
f
transmitted between the transmitter and the receiver and
a mobile receiver moving with a velocity v. Also, we define θ as the angle
between the motion direction of the mobile unit and the arrival direction of
7
the signal. In this case, the frequency change of the signal is known as the
Doppler shift and denoted by
d
f
, is given by
. cos
d c
v
f f
c
θ
=
where
d
f
is the Doppler shift, v is the velocity of the mobile unit, c is velocity
of light,
c
f
is frequency carrier, and θ is angle between the motion direction
of the mobile and the arrival direction of the signal. Since different paths
arrive from different angles, a variety of Doppler shifts corresponding to

and surround objects. Incoming waves arrive from many different directions
with different propagation. These signals are combined at the receiver
antenna. Consequently, signals can vary widely in amplitude and phase. An
illustration of multi-path signal is expressed in Figure 1.6. Base on channel
parameters and characteristics of signal to be transmitted, fading channels can
be classified as follows.
1.6.1 Flat fading versus frequency selective fading
Frequency selectivity is also an important characteristic of fading
channels. If all the spectral components of the transmitted signal are affected
in a similar manner, the fading is said to be nonselective fading or flat fading.
This is case for narrowband systems, in which the transmitted signal
bandwidth is much smaller than coherence bandwidth
c
B
. This bandwidth
9
measures frequency range over which fading process is correlated. In
addition the coherence bandwidth is related to the maximum delay
spread
max
τ
by
max
1
c
B
τ
;
where
max

m
f
by
s
T
1
c
m
f
τ
;
where
m
f
is maximum Doppler spread,
c
τ
is the coherence time
The fading is said to be slow if the symbol time duration
s
T
is smaller than
the channel’s coherence time
c
τ
, slow fading often modeled as time invariant
channels over a number of symbol intervals. Moreover, channel parameters
can be estimated with different estimation techniques. Otherwise it is
considered to be fast. In general, it is difficult to estimate channel parameters
in a fast fading channel. Figure 1.8 illustrates frequency selective fading and

As illustrated in Figure 2.1, at the receiver, the independent copies are
combined to get a good decision. Frequency diversity is used to combat
frequency selective fading.
13
2.2.2 Time diversity
Another approach to achieve diversity, which is illustrated in Figure 2.2, is
transmitting the desired signal in M different time slots.
Figure 2.2: Time diversity
The intervals between transmissions of same symbol should be at least the
coherence time so that different copies of same symbol undergo independent
fading. We notice that sending same symbol M times is applying the
( ,1)M

repetition code. Error control coding together with interleaving can be an
effective way to combat time selecting fading.
2.2.3 Space diversity
Figure 2.3 illustrates space diversity method. This method has been a
popular technique in wireless communications. Space diversity is also called
antenna diversity. It is typically implemented by using multiple antennas or
antenna arrays arranged together in space for transmission and reception. The
multiple antennas are separated physically by a proper distance so that the
individual signals are uncorrelated. Typically, a separation of a few
wavelengths is enough to obtain uncorrelated signals. Space diversity can be
employed to combat both frequency selective fading and time selective
fading.
14
Figure 2.3: Space diversity
2.3 Diversity combining methods
The idea of receive diversity is to combine copies of transmitted signal
which undergo independent fading. In general, the performance of

1 2 1
( ), ( ) ( )
M
r t r t r t
−
=r
2.3.1 Selection combining
Selection combining is a simple diversity combining method. As shown in
15
Figure 2.4. Consider a receive diversity system with
R
n
receive antennas. In
this method, the signal with the strongest signal-to-noise ratio (SNR) at
every symbol interval is selected as the output. In practically, the signal
with the highest sum of the signal and noise power
( )S N+
is used, since it is
difficult to measure the SNR.
Figure 2.4: Selection combining
2.3.2 Switched Combining
In a switched combining diversity system, the receiver scans all the
diversity branches and selects a particular branch with the SNR above a
certain predetermined threshold. This signal is selected as the output,
until its SNR drops below the threshold. When this happens, the
receiver starts scanning again and switches to another branch. This scheme is
also called scanning diversity. A scheme of switched combining is shown in
Figure 2.5.
16
Figure 2.5: Switched combining

and
i
ϕ
be the amplitude and phase of the received signal
i
r
,
respectively. Assuming that each receive antenna has the same average noise
power, the weighting factor
i
α
can be represented as
.
i
j
i i
A e
φ
α
−
=
This method is called optimum combining since it can maximize the
17
output SNR. It is shown that the maximum output SNR is equal to the sum of
the instantaneous SNRs of the individual signals.
In scheme as shown in Figure 2.6, each individual signal must be co-phased,
weighted with its corresponding amplitude and then summed. This scheme
requires the knowledge of channel fading amplitude and signal phases. So, it
can be used in conjunction with coherent detection, but it is not practical for
non-coherent detection.