Performance improvement of mimo mc cdma system using equalization, beamforming and relay

PERFORMANCE IMPROVEMENT OF MIMO
MC CDMA SYSTEM USING EQUALIZATION,
BEAMFORMING AND RELAY
THESIS
Submitted by
N.TAMILARASAN
In partial fulfilment for the award of the degree
of
DOCTOR OF PHILOSOPHY
in
ELECTRONICS AND COMMUNICATION ENGINEERING
DEPARTMENT OF ELECTRONICS AND COMMUNICATION ENGINEERING
PONDICHERRY ENGINEERING COLLEGE
PONDICHERRY UNIVERSITY
PUDUCHERRY- 605 014
INDIA
JULY 2015

ii
Dr. L.Nithyanandan
Professor
Department of Electronics and Communication Engineering
Pondicherry Engineering College
Puducherry – 605 014.
CERTIFICATE
Certified that this thesis entitled “PERFORMANCE IMPROVEMENT
OF MIMO MC CDMA SYSTEM USING EQUALIZATION,
BEAMFORMING AND RELAY” submitted for the award of the degree of
DOCTOR OF PHILOSOPHY in ELECTRONICS AND COMMUNICATION
ENGINEERING of the Pondicherry University, Puducherry is a record of original
research work done by Shri. N. TAMILARASAN during the period of study under
my supervision and that the thesis has not previously formed the basis for the award
to the candidate of any Degree, Diploma, Associateship, Fellowship or other similar
titles. This thesis represents independent work on the part of the candidate.
Supervisor
(Dr. L. Nithyanandan)
Date:
Place: Puducherry

iii
ABSTRACT
In wireless communication system, signals are transmitted through
electromagnetic wave propagation in the atmosphere. The presence of reflectors in
the surrounding of the transmitter and the receiver creates multiple paths through
which signals are transmitted. The signals transmitted through these multiple paths
often interfere with each other and receiver receives the superposition of multiple
copies of the transmitted signal which elongates the symbol and results in inter
symbol interference (ISI). Each signal copy will experience different attenuation,
delay and phase shift while traveling from the source to the destination. This can
result in either constructive or destructive interference which is referred as fading.
To mitigate channel fading, diversity techniques which transmit and process
multiple copies of the same signal through different fading channel is often
employed. At the receiver, these different copies of the signal are effectively
combined to reconstruct the complete transmitted signal.
The multi carrier Code division multiple access (MC CDMA) combined
with themulti input and multi output (MIMO) technique has become a core
technology for future mobile radio communication system to cater millions and
millions of users with enriched services. Multipath propagation, mobility of
transmitter, receiver and local scattering cause the signal to spread in frequency,
time and angle, resulting in ISI in the received signal. Possible potential gain in
spectral efficiency is challenged by the receiver’s ability to accurately detect the
symbol in the presence of ISI. Eventhough MC CDMA mitigates the problem of
time dispersion, still it is necessary to effectively remove the amplitude/phase shift
caused by channel and to solve this problem,channel estimation based equalization
in the receiver, beam forming (BF) in the transmitter and relay in channel can be
invoked.

iv
The ISI arising in the multipath fading channels has been recognized as
the major obstacle for high speed data transmission over wireless channel.
Equalization technique is an effective means for reducing the influence of ISI.
Several combining and equalization techniques for orthogonal frequency division
multiplexing (OFDM), MC CDMA and MIMO signals have appeared in the
literature. The equal gain combining(EGC), maximum ratio combining (MRC) and
orthogonality restoring combining (ORC) suitable for single carrier communication,
if used in multi carrier system, reduces the orthogonality of user creating additional
interference in the receiver. Among the various equalization methods for OFDM,
MC CDMA and MIMO, minimum mean square error (MMSE) is considered to be a
good solution for data recovery inspite of its complexity. MMSE can effectively
utilize the diversity of the frequency selective channel and reduce the ISI. The
performance of proposed MIMO MC CDMA system with MMSE equalization is
compared with MC CDMA and MIMO MC CDMA without equalization. Further
the performance of the system is tested with different antenna configuration,
subcarriers (SCs) and different modulations.
The MMSE equalization alone is not an efficient way of reduction of ISI,
as the equalization is not carried out with the knowledge of channel impairments. To
strengthen the effect of equalization for reducing ISI, channel estimation is used to
estimate the amplitude/phase shift caused by the wireless channel impairments. A
modified pilot channel estimation (MPCE) is proposed for MIMO MC CDMA
system in which the number and position of pilots are varied dynamically based on
the channel condition. The advantage of MPCE based MMSE equalization is the
reduction of ISI with optimal pilots so as to reduce spectral loss. The MPCE based
MMSE equalization removes the effect of the wireless channel and ensures effective
symbol demodulation.
Apart from MPCE based MMSE equalization in the receiver, improved
transmit BF(ITBF) is incorporated as a preventive measure where array of antennas
are “directed” at a desired target or source by adjusting the relative gain and phase of
the array elements. By this, the antenna pattern, or beam, can be made to point in a
favoured direction and to attenuate other directions in order to reduce the effect of

v
interference source. BF improves reception quality and increases throughput in a
MIMO MC CDMA communication system. ITBF is made more effective with the
knowledge of channel information obtained from the MPCE.
To further improve the system performance, novel relays (NR) are
utilized for contiguous coverage of areas with high traﬃc density. A traditional
approach would be to increase the base station density. However, the network cost
scales linearly with the amount of deployed BSs whereas the amount of users and
the average revenue per user will not increase accordingly. Therefore, cost eﬃcient
alternative deployment concepts are needed. One promising alternative deployment
is a novel relay with MPCE based equalization and ITBF to extend the high
throughput coverage of next generation mobile networks.
To summarize, this work proposes different techniques in the transmitter,
receiver and the channel to enhance the performance of MIMO MC CDMA system
through ISI reduction so that apart from increasing the capacity, quality is also
ensured.

vi
ACKNOWLEDGEMENT
With immense pleasure and deep sense of gratitude, I would like to place
on record my thanks to all those who have contributed to the successful completion
of this research work.
I am duty bound to express my deep indebtness to my supervisor and
mentor, Dr. L. Nithyanandan, Professor, Department of Electronics and
Communication Engineering, for his unflinching support, selfless motivation and
contagious enthusiasm from the inception of the research to the culmination stage.
As my supervisor, he has constantly forced me to remain focused on achieving my
goal. His observations and comments helped me to establish the overall direction of
the research and to move forward with investigation in depth.
I express my ingenious, sincere requital to Doctoral committee member
Dr. P. Dananjayan, Professor, Department of Electronics and Communication
Engineering, Pondicherry Engineering College for his immense help, unswerving
motivation and lively deliberation inspite of his busy schedule. I express my
profound gratitude and allegiance to Doctoral committee member Dr. K. Porsezian,
Professor, Department of Physics, Pondicherry University, for his magnanimous
benignity and benevolence, which has enthused me to work harder and achieve the
goal.
I am grateful to Dr. D. Govindarajulu, Principal, Pondicherry
Engineering College for his whole hearted support and for permitting me to make
use of the facilities in the college for the research work. I deem it a privilege to
record my sincere thanks to Dr. V. Prithiviraj, Former Principal, Pondicherry
Engineering College, who has taught me to think ahead and encouraged me.
I take this opportunity to thank Dr. G.Nagarajan, Professor and Head,
Department of Electronics and Communication Engineering and Dr. E Srinivasan,

vii
Former Head of the Department for allowing me to utilize the facilities in the
department for my work as and when needed. I am also thankful for their
encouragement throughout my research work.
I also thank, the Chairman and Principal of Shri Krishna College of
Engineering and Technology, Puducherry for their support and encouragement and
love given by them immensely during my research work.
With deep sense of gratitude, I wish to express my sincere thanks to my
God in heaven Thiru J. Narayanasamy and my Goddess in earth Mrs. Lakshmi
Kandham Narayanasamy for the motivation provided.
My profound gratitude is to my wife, Mrs. R. Ranjani Tamilarasan and
my child T.R. Tharanerajaan for their patience, kind heartedness and moral
support offered during the crucial period of my research
N. TAMILARASAN

viii
TABLE OF CONTENTS
CHAPTER NO. TITLE PAGE NO.
ABSTRACT iii
ACKNOWLEDGEMENT vi
LIST OF TABLES xiii
LIST OF FIGURES xiv
LIST OF ABBREVIATIONS xviii
LIST OF SYMBOLS xxi
1 INTRODUCTION 1
1.1 GENERAL 1
1.2 NEED FOR MIMO MC CDMA SYSTEM 2
1.2.1 CDMA 3
1.2.2 OFDM 4
1.2.3 Multicarrier CDMA 7
1.2.4 MIMO 8
1.3 LITERATURE SURVEY 9
1.4 OBJECTIVE OF THE RESEARCH WORK 17
1.5 THESIS ORGANIZATION 17
2 PERFORMANCE OF MIMO MC CDMA SYSTEM
WITH MMSE EQUALIZATION 19
2.1 INTRODUCTION 19
2.2 DIFFERENT TYPES OF EQUALIZATION
TECHNIQUE 20
2.2.1 Equal Gain Combining (EGC) 20
2.2.2 Maximum Ratio Combining (MRC) 20
2.2.3 Orthogonality restoring combining (ORC) 21
2.2.4 Threshold ORC (TORC) 21

ix
2.2.5 Partial Equalization(PE) 22
2.2.6 Pre Equalization 22
2.2.7 Combined Equalization 23
2.2.8 Maximum Likelihood Sequence
Estimator (MLSE) 23
2.2.9 Maximum Likelihood Symbol by Symbol
Estimation (MLSSE) 24
2.2.10 Maximum Likelihood Detection 25
2.2.11 Minimum Mean Square Equalizer (MMSE) 25
2.3 MIMO MC CDMA SYSTEM WITH MMSE
EQUALIZATION 27
2.4 SIMULATION RESULTS AND DISCUSSION 31
2.4.1 Performance of the System with MMSE
Equalization (BPSK) 32
2.4.2 Performance of the System with MMSE
Equalization (QPSK) 35
2.5 SUMMARY 37
WITH MPCE BASED MMSE EQUALIZATION 38
3.1 INTRODUCTION 38
3.2 TYPES OF CHANNEL ESTIMATION 39
3.2.1 Pilot Based Channel Estimation 39
3.2.2 Blind Channel Estimation 41
3.2.3 Semi Blind Channel Estimation 42
3.3 CHANNEL ESTIMATION FOR OFDM SYSTEM 42
3.4 PILOT BASED CHANNEL ESTIMATION FOR
MC CDMA SYSTEM 43

x
3.5 MODIFIED PILOT CHANNEL ESTIMATION
(MPCE) BASED MMSE EQUALIZATION FOR
MIMO MC CDMA SYSTEM 44
3.6.1 Performance of the System with MPCE
based MMSE Equalization (BPSK) 49
3.6.2 Performance of the system with MPCE
based MMSE Equalization (QPSK) 51
3.6.3 Performance of the system with MPCE based
MMSE Equalization (16 QAM) 54
3.7 SUMMARY 58
4 PERFORMANCE OF MIMO MC CDMA
SYSTEM WITH MPCE BASED MMSE
EQUALIZATION AND ITBF 59
4.1 INTRODUCTION 59
4.2 TYPES OF BEAMFORMING 61
4.2.1 Fixed Beamforming 61
4.2.2 Adaptive Beamforming 63
4.3 MIMO MC CDMA SYSTEM WITH MPCE
BASED MMSE EQUALIZATION AND ITBF 63
based MMSE and ITBF (QPSK) 66
based MMSE and ITBF (16 QAM) 67
4.5 SUMMARY 71

xi
WITH MPCE BASED MMSE EQUALIZATION,
ITBF AND NR 72
5.1 INTRODUCTION 72
5.2 CLASSIFICATION OF RELAY SYSTEM 73
5.2.1 Single Relay System 73
5.2.2 Multiple Relay System 73
5.2.3 MIMO Multi Relay System 74
5.3 ONE WAY AND TWOWAY RELAY SYSTEM 75
5.3.1 One Way Half Duplex Mode 75
5.3.2 Two Way Half Duplex Mode 77
5.4 FULL DUPLEX TRANSMISSION 80
5.5 CLASSIFICATION OF RELAY SCHEMES 80
5.5.1 Amplify and Forward (AF) Relay 80
5.5.2 Decode and Forward (DF) Relay 81
5.5.3 Compress and Forward (CF) Relay 82
5.6 MIMO OFDM SYSTEM WITH RELAY 82
5.7 MIMO MC CDMA SYSTEM WITH MPCE BASED
MMSE EQUALIZATION, ITBF AND NR 83
5.8.1 Performance of the System with MPCE based
MMSE, ITBF and NR (QPSK) 88
5.8.2 Performance of the System with MPCE based
MMSE, ITBFand NR (16QAM) 90
5.9 SUMMARY 95

xii
6 CONCLUSION 96
6.1 RESEARCH CONTRIBUTIONS 97
6.2 SCOPE FOR FURTHER WORK 99
REFERENCES 100
LIST OF PUBLICATIONS 126
VITAE 127

xiii
LIST OF TABLES
TABLE NO. TITLE PAGE NO.
2.1 Simulation parameter of the system with MMSE equalization 31
2.2 Impact of MMSE equalization (BPSK) at target BER of 10-3
34
2.3 Impact of MMSE equalization (QPSK) at target BER of 10-3
37
3.1 Simulation parameters of the system with MPCE based
MMSE equalization 48
3.2 Impact of MPCE based MMSE equalization (BPSK) at
target BER of 10-3
51
3.3 Impact of MPCE based MMSE equalization (QPSK) at
target BER of 10-3
53
3.4 Impact of MPCE based MMSE equalization (16QAM) at
target BER of 10-3
57
4.1 Simulation parameters of the system with MPCE based
MMSE and ITBF 66
4.2 Impact of MPCE based MMSE and ITBF at target BER
of 10-3
(QPSK) 67
4.3 Impact of MPCE based MMSE and ITBF at target BER
of 10-3
(16 QAM) 71
5.1 Simulation parameter of the system with MPCE based
MMSE, ITBF and NR 88
5.2 Impact of MPCE based MMSE, ITBF and NR (QPSK) at
the target BER of 10-3
89
5.3 Impact of MPCE based MMSE, ITBF and NR (16 QAM) at
95

xiv
LIST OF FIGURES
FIGURE NO. TITLE PAGE NO.
1.1 Code division multiple access 4
1.2 Comparison of the bandwidth utilization for
FDM and OFDM 5
1.3 OFDM transmitter 6
1.4 OFDM receiver 6
1.5 A simple MC CDMA transmitter 7
1.6 A simple MC CDMA receiver 8
1.7 Multiple Input Multiple Output antenna system 8
2.1 Example of ISI on received pulse in a Communication system 19
2.2 Noise Enhancement in ORC 21
2.3 Noise Enhancement in MMSE 26
2.4 Simplified diagram of MMSE equalizer 28
2.5 Block diagram of MIMO MC CDMA Transmitter 29
2.6 Block diagram of MIMO MC CDMA Receiver with MMSE
Equalization 29
2.7 Performance of the system with/without MMSE
(BPSK, 16 SCs) 32
(BPSK, 64 SCs) 33
(BPSK, 128 SCs) 33
(QPSK, 16 SCs) 35
(QPSK, 64 SCs) 36
(QPSK, 128 SCs) 36

xv
3.1 Channel estimation procedure 39
3.2(a) Block type pilot arrangement 40
3.2(b) Comb type pilot arrangement 40
3.2(c) Irregular pilot arrangement 41
3.3 Block diagram of OFDM transceiver with pilot based
channel estimation 43
3.4 Block diagram of MC CDMA transmitter with pilot signal 44
3.5 Block diagram of MC CDMA receiver with pilot based
channel estimation 44
3.6 (a) Flow chart of MPCE (Delay spread)
(b) MPCE (Doppler spread) 45
3.7 Block diagram of modified pilot insertion based transmitter 46
3.8 Block diagram of MPCE based MMSE equalization 46
3.9 MPCE estimation procedure 47
3.10 Performance of the system with MPCE based MMSE
(BPSK, 16 SCs) 49
(BPSK, 64 SCs) 50
(BPSK, 128 SCs) 50
(QPSK, 16 SCs) 52
(QPSK, 64 SCs) 52
(QPSK, 128 SCs) 53
(2x2, 16 SCs, 16 QAM) 54
(2x2, 64 SCs, 16 QAM) 55

xvi
(2x2, 128 SCs, 16 QAM) 55
(4x4, 16 SCs, 16 QAM) 56
(4x4, 64 SCs, 16 QAM) 56
(4x4, 128 SCs, 16 QAM) 57
4.1 Switched Beamforming 61
4.2 Delay and sum Beamforming 62
4.3 Beam space Beamforming 62
4.4 Flowchart of ITBF 64
4.5 Block diagram of MIMO MC CDMA system with
MPCE based MMSE and ITBF 65
4.6 Performance of 4 x 4 antennas configuration (128 SCs, QPSK) 67
4.7 Performance of 2 x 2 antennas configuration (16 SCs, 16 QAM) 68
4.9 Performance of 2 x 2 antennas configuration (128 SCs,16 QAM)69
4.12 Performance of 4 x 4 antennas configuration (128 SCs, 16 QAM)70
5.1 Simple three node relay 73
5.2(a) Multi relay without direct link to D 74
5.2(b) Multi relay with direct link to D 74
5.3(a) MIMO relay with direct link to D 74
5.3(b) MIMO Multi relay without direct link to D 75
5.4(a) One way half duplex module 1 76
5.4(b) One way half duplex module II 76
5.4(c) One way half duplex module III 76
5.4(d) One way half duplex module IV( Forwarding) 77

xvii
5.5(a) Two way half duplex module I 78
5.5(b) Two way half duplex module II 78
5.5(c) Two way half duplex module III 79
5.5(d) Two way half duplex module IV 79
5.6 Full duplex module 80
5.7 Comparison of AF and DF relay 82
5.8 Block Diagram of MIMO OFDM transmitter and receiver
with Relay 83
5.9 Block diagram of MIMO MC CDMA system with MPCE
based MMSE, ITBF and NR 84
5.10 MIMO MCCDMA with three Node AF relay 85
5.11 Performance of 4x1x 4 antenna configuration for 128 SCs
(QPSK) 88
5.12 Performance of 4 x 2 x 4 antenna configuration for 128 SCs
(QPSK) 89
5.13 Performance of 2 x1x 2 antenna configuration for 64 SCs
(16 QAM) 91
(16 QAM) 91
(16 QAM) 92
(16 QAM) 92
(16 QAM) 93
5.18 Performance of 4 x 2 x 4 antenna configuration for 64 SCs
(16 QAM) 93
(16 QAM) 94
(16 QAM) 94

xviii
LIST OF ABBREVIATIONS
AF Amplify and Forward
ANC Analog Network Coding
AWGN Additive White Gaussian Noise
BC Broadcast Channel
BEP Bit Error Probability
BER Bit Error Rate
BF Beam Forming
BPSK Binary Phase Shift keying
CDMA Code Division Multiple Access
CP Cyclic Prefix
CSI Channel State Information
D Destination
DF Decode and Forward
DFT Discrete Fourier Transform
DS Direct Sequence
DVB Digital Video Broadcasting
Eb/No Energy per Bits to Spectral Noise Density
EDGE Enhanced Data Rates for GSM Evolution
EF Estimate and Forward
EGC Equal Gain Combining
FDM Frequency Division Multiplexing
FDMA Frequency Division Multiple Access
FFT Fast Fourier Transform
GPRS General Packet Radio Service
GSM Global System for Mobile Communication
HDTV High Definition Television
ICE Ideal Channel Estimation
IFFT Inverse Fast Fourier Transform
IP Internet Packet

xix
ISI Inter Symbol Interference
ITBF Improved Transmit BF
ITU(T) International Telecommunication Union (Technical)
LS Least Square
MAI Multi Access Interference
MAP Maximum a Posteriori
MC CDMA Multicarrier Code Division Multiple Access
MIMO Multi Input and Multi Output
MISO Multiple Input Single Output
ML Maximum likelihood
MLSE Maximum Likelihood Sequence Estimator
MLSSE Maximum Likelihood Symbol by Symbol Estimation
MMS Multimedia Messaging Service
MMSE Minimum Mean Square Error
MMSEC Minimum Mean Square Error Combining
MPCE Modified Pilot Channel Estimation
MRC Maximum Ratio Combining
MSNR Maximum Signal to Noise Ratio
MT Multi Tone
NR Novel relay
OF Observe and Forward
OFDM Orthogonal Frequency Division Multiplexing
ORC Orthogonality Restoring Combining
P/S Parallel to Serial
PE Partial equalization
PN Pseudo Noise
QF Quantize and Forward
QPSK Phase Shift Keying
R Relay
RBF Random BF
S Source
S/P Serial to Parallel

xx
SC Subcarrier
SIMO Single Input Multiple Output
SINR Maximum Signal to Interference Noise Ratio
SISO Single Input Single Output
SNR Signal to Noise Ratio
STC Space Time Coding
TDD Time Division Duplexing
TDMA Time Division Multiple Access
TORC Threshold ORC
WCDMA Wideband Code Division Multiple Access
ZF Zero Forcing
1D One dimensional
1G First Generation
2D Two dimensional
2G Second Generation
3G Third Generation
4G Fourth Generation

xxi
LIST OF SYMBOLS
B(n)
Beam form vector
bk(i) ith
transmitted data bit of k th
user
cn Spreading code
d Data sequence
d(k)
Data symbol of user k
dµ Data symbols vectors
Dn Desired signal
E{.} Expectation
Eb Bit energy
f Frequency
F(n)
Amplified signal vector
Gl,l lth
diagonal element of the equalizer matrix G
Gm Subcarrier weighting coefficient
H Diagonal matrix
𝐻 𝑚
∗
mth
channel coefficient
𝐻𝑠
𝑛
Channel matrix between source and relay
𝐻(k) Channel transfer function of k th
user
Hd Channel matrix between source and destination
hi Channel coefficient of ith
transmit antenna
Hl,l lth
diagonal element of the channel matrix H
hn Subcarrier flat fading gain
Hr Channel matrix between relay and destination
ηr(t) Noise vector of the relay at time t
In Multi user interference
𝐼 𝑁 𝐶
Identity matrix
M Number of data symbols transmitted
N Processing gain
N l lth
element of the noise vector n

xxii
N0 Single sided power spectral density
Nb Number of symbols in nth
Subcarrier
Nc Subcarrier
Nd Destination or Receiving antenna
Nr Relay antenna
Ns Source ortransmitting antenna
P(n)
Precoding matrix for source signal
P{.} Probability
r Received vector
r(t) Received signal at time t
Rl lth
element of the received vector r
S l lth
element of the vector S
S(n)
(t) Signal vector at time t
S(t) Transmitter signal at time t
T
Transpose
T Bit duration
Ts Symbol duration
uTs(t) Rectangular waveform with unit amplitude
X(k) Transmitted data samples
x(n)
(t) Precoded signal vector at time t
x(t) Channel precoding matrix for source
xr(t2) Amplified signal of relay at t2
Y(k) Received signal
yd(t) Received signal at the destination at time t
YP(k) Received pilot signals
yr(t) Received signal at the relay at time t
(.)* Complex conjugation
(n)
nth
Subcarrier
||.|| Norm of a vector
∆2
(.,.) Squared Euclidean distance
αn Equalizer co efficient
β Equalization parameter with value in the range of [-1,1]

xxiii
η(t) Additive white Gaussian noise
ηd(t) Noise vector of the destination at time t
ηn Noise component
λ Threshold
σ2
Variance of the noise
φn Subcarrier fading phase
ωn Radian frequency of nth
subcarrier

1
CHAPTER 1
INTRODUCTION
1.1 GENERAL
Wireless technologies have evolved remarkably since Guglielmo Marconi
first demonstrated radio transmission in the year 1896 by controlling an electric bell
through a remote arrangement. The radio transmission has been very much useful
for the people to communicate with each other and has timely access to information
regardless of the location of individuals. This is supported by the advances in the
wireless networking technology and portable computing devices with reduction in
the size of physical device leading to the rapid development in mobile
communication infrastructure.
The first generation (1G) has been an analog based FDMA (Frequency
division multiple access) system for voice transmission only. The 2G cellular
systems uses time division multiple access (TDMA) based digital modulation
schemes with source and error correction coding techniques to improve user
capacity, voice quality and spectrum efficiency. The second generation (2G) systems
include Global System for Mobile communications (GSM), IS-54, IS-136, IS-95 and
Japan Digital Cellular. These systems were able to perfectly provide basic services
and when demand for variety of wideband services increased the evolution towards
the 3G started. The first step in this direction has been commonly accepted as 2.5G
with the development of new technologies based on the classical GSM (e.g., General
Packet Radio Service (GPRS) and Enhanced Data Rates for Global Evolution
(EDGE)).The real third generation (3G) mobile system have been developed to offer
both low and high bandwidth services like telephony, Internet access and
multimedia at any time and from anywhere through a single device. The first 3G

2
system based on the Wideband Code Division Multiple Access (WCDMA)
transmission protocol offer wideband services, such as wireless Internet services
(with peak rate of 384 Kbps) and video transmissions (with data rate up to 2Mbps).
As 3G technology and its first evolution 3.5G congested in few years,
telecommunication research community started working for the fourth generation
(4G) wireless communication. 4G solution is based on full IP convergence system to
meet the needs of high performance applications like wireless broadband access,
Multimedia Messaging Service, video chat, mobile TV, High Definition Television
(HDTV) content and Digital Video Broadcasting (DVB).The goal is to have data rates
in the range of 100-500 Mbps with premium quality and high security. In order to
realize this, it is necessary to tackle the hostile physical properties of wireless channel
in the form of rapid time variation, fading (large scale and small scale), multipath
propagation and channel interference. The interference created by user of different
cells using the same resource is called co-channel interference and interference
resulting from users of same cell is called adjacent cell interference. Counter measures
were employed to combat these impairments through spreading, space diversity and
frequency diversity. Therefore MIMO MC CDMA system is standardized by
international telecommunication unit (ITU(T)) for 4G and beyond 4G system.
1.2 NEED FOR MIMO MC CDMA SYSTEM
The performance of current wireless communication system degrades due to
excessive multi path propagation and Doppler spread. Due to delay spread bit arrives
the receiver at different times because of the different paths taken causing bit
overlap and interferes with each other resulting in ISI, which leads to distortion of
signals causing bit error at the receiver. As the data rate is increased the transmission
time for individual pulse decreases, become narrow and effect of channel delay
spread increases. ISI has been recognized as the major obstacle for high speed data
transmission over multipath wireless channel.
OFDM, a multi carrier modulation scheme splits high data rate streams into a
number of lower rate streams that are transmitted simultaneously over a number of
SC. Since the symbol duration increases in the lower rate parallel SCs, the relative

3
amount of dispersion in time caused by multipath delay spread decreases. Although
OFDM is robust to frequency selective fading, it has severe disadvantages in SC
synchronization and sensitivity to frequency offset estimation. The other main issue
of OFDM is the presence of a large number of SCs which exhibits a non constant
nature in its envelope.
Future generation is to cater more users with enriched services, CDMA has a
strong candidate to support multimedia mobile services as opposed in conventional
schemes such as TDMA/FDMA which also has the ability to cope up with the
asynchronous nature of the multimedia traffic. By employing Rake receivers CDMA
systems can coherently combine the multipath components due to the hostile frequency
selective channel and the processing gain due to spreading provides robustness to the
multi user interference. The use of conventional CDMA does not seem to be realistic
when the data rates go up to a hundred megabits per second due to severe ISI(inter
symbol interference) and the difficulty in synchronizing a fast sequence.
Further to enhance the performance, the MIMO antennas at both ends of a
wireless link helps to realize high data rate through spatial multiplexing and
drastically improve the spectral efficiency of the system and provide link reliability
in rich scattering environments through spatial diversity. In addition, the capacity of
MIMO system increases linearly with the number of transmit receive antenna pairs
without increasing transmitted power. Even though multicarrier modulation and
MIMO mitigate ISI, effectiveness of this suppression is limited and both of these two
techniques cannot work perfectly in fast fading environment [1-3]. To overcome the
problem, this work proposes equalization and channel estimation in the receiver BF in
the transmitter and relay in the network to reduce the ISI, cater more number of users
and improve the performance.
1.2.1 CDMA
In CDMA the information signals of different users are modulated by
orthogonal or non orthogonal spreading codes. The resulting spread signals
simultaneously occupy the same time and bandwidth, as shown in Fig.1.1. The

4
receiver uses the spreading code structure to separate out the different users. This
method has the feature that the unwanted signals with different code will get spread
even more by the process and appear like noise to the receiver. Some of the
properties that have made CDMA useful are: signal hiding, non interference with
existing systems, anti jam, interference rejection, information security and multipath
tolerance. Though CDMA ensure effective utilization of bandwidth, it is not capable
of transmitting data at rates of around 5 to 10 Mbps. Various studies indicate that
some sort of multi carrier modulation (MCM) scheme can only support higher data
rates and be suitable for mobile paths.
Fig.1.1 Code division multiple access
1.2.2 OFDM
OFDM is the basic MCM scheme that has been extremely used for high data
rate application. In OFDM, the available spectrum is divided into many carriers,
each one being modulated by a low rate data stream. Like FDMA, in OFDM also
[2-4] the multiple user access is achieved by subdividing the available bandwidth
into multiple narrow band channels, which are allocated to users. However, OFDM
uses the spectrum much more efficiently by spacing the channels much closer
together with no overhead as in FDMA. This is achieved by making all the carriers
orthogonal to one another, thus preventing interference between the closely spaced
carriers as shown in Fig. 1.2.
Time
Frequency

5
Fig.1.2 Comparison of the bandwidth utilization for FDM and OFDM
Each carrier in an OFDM signal has a very narrow bandwidth
(i.e. 1 kHz), thus the resulting symbol rate is low. This result in the signal having a
high tolerance to multipath delay spread, as the delay spread must be long enough to
cause significant inter symbol interference (e.g. > 100 ms). Fig.1.3 and 1.4 show a
simple OFDM transmitter and OFDM receiver respectively.
OFDM has a serious limitation of being sensitive to nonlinear amplification,
frequency and phase errors in SCs. Moreover the mobile community may not be
interested in loosing the merits of CDMA and therefore possibilities of mixing
CDMA and MCM have been initiated. Researchers have analyzed three different
combinations of OFDM and CDMA; multi carrier (MC) CDMA, MC direct
sequence (DS) CDMA and multi tone (MT) CDMA [5], out of which MC CDMA is
chosen as possible access scheme for beyond 3G mobile communication systems.
Frequency
Conventional Frequency Division Multiplex (FDM)
Frequency
Saving of bandwidth
Orthogonal Frequency Division Multiplex (OFDM)

6
Fig.1.3 OFDM transmitter
Fig.1.4 OFDM receiver
S/P
Data Symbols
1/ √T
(1/ √T) e
j 2π t / T
(1/ √T) e
j 2π (N-1) t / T
OFDM Signal
P/S
∫T
∫T
∫T
1/ √T
OFDM Signal (1/ √T) e
– j 2π / T
(1/ √T) e
- j 2π (N-1) t / T
Data Symbols

7
1.2.3 Multicarrier CDMA
The MC CDMA transmitter spreads the original data stream in the frequency
domain over different SCs [6] using a given spreading code as depicted in Fig. 1.5.
In this system the SCs convey the same information at one time. The binary data dk
are first multiplied by the spreading code (PN binary sequences) ck(i) and the bit
duration T of the data is much longer than the PN sequence chip duration Tc.
Fig.1.5 A simple MC CDMA transmitter
Normally N=T/TC may be equal to 32, 64, 128 or higher. After the necessary
encoding, the sequence that corresponds to one data bit is modulated in the band of
N modulators. A very efficient way of realizing such a modulation scheme is
through the inverse fast fourier transform (IFFT). The distance between SC is equal
to the chip rate or its multiple. The guard time/ cyclic prefix should be introduced
for each symbol in order to combat the inter symbol interference caused by the
channel and filters in the system. At the output of IFFT the pulse shaping filter
followed by a modulator shifts the whole spectrum to the desirable frequency band.
Another solution is to apply the offset IFFT that eliminates the necessity of the
above mentioned frequency shift. In the MC CDMA system, SC frequencies are
usually chosen to be orthogonal to each other, i.e., SC frequencies attempt to satisfy
the following condition:
∑dk (t)
ck (1)
ck (N)
exp (j 2πf1t)
exp (j 2πfNt)
sk (t)

8
0
cos( ) cos( ) 0,
Tc
i jt t dt for i j    (1.1)
where TC is the chip duration, ωi and ωj are ith
and jth
carrier frequencies respectively.
The basic structure of the MC CDMA mobile station receiver shown in
Fig. 1.6 contains the same blocks as that of OFDM receiver. However, the output of
fast fourier transform (FFT) is fed to k branches (one branch for each user)
containing an equalizer and a despreading block to equalize the individual channel
and receive the signal using the user‟s unique PN code.
Fig.1.6 A simple MC CDMA receiver
1.2.4 MIMO
To improve channel throughput and robustness of a radio link, multiple
antennas are employed at both the transmitter and the receiver which is known as
MIMO shown in Fig.1.7. A MIMO system with similar count of antennas at both the
transmitter and the receiver is able to multiply the system throughput linearly with
every additional antenna. MIMO often employs Spatial Multiplexing to enable
signal (coded and modulated data stream) to be transmitted across different spatial
domains.
Fig.1.7 Multiple Input Multiple Output antenna system
Tx Rx
∫T (.)
∑
∫T (.)
exp (-j 2πf1t)
exp (-j 2πfNt)
ck (1)
ck (N)
r(t)
d k, n

9
The use of multiple antennas in the receiver and/or in the transmitter
basically provides the following advantages [7].
 Array gain: Array gain is an increase in average received SNR obtained
by coherently combined signals transmitted from multiple antennas or
received by multiple antennas. In order to get array gain channel
knowledge at the transmitter or receiver is required.
 Diversity gain: Diversity gain refers to improved link reliability in
space, time, frequency or polarization. Diversity gain can be exploited in
the case of independently fading multiple channels experienced by
different antennas in order to mitigate the effect of multipath fading
 Spatial multiplexing gain: Multiplexing gain corresponds to the
increased data rate due to the ability to transmit multiple parallel data
streams over MIMO channel without increasing the bandwidth or total
transmit power. The capability to support multiple streams depends on
the rank of the channel matrix H.
As a result, the MIMO technology can be used to achieve improved system
performance, such as increased spectral efficiency, increased user throughput and
better coverage.
1.3 LITERATURE SURVEY
Today‟s wireless communication requires high data rate and good quality,
which is possible only if the received signal is not affected by channel impairments.
The diversity combining schemes are applied to combine the multipath received
signals of a diversity reception device into a single improved signal. These schemes
are applied at the receiver to mitigate the effect of channel fading on the received
signal to obtain the signal with maximum signal to noise ratio (SNR).

10
Various types of combating techniques for channel impairments are proposed
in the literature based on noise, interference and capacity for OFDM, MC CDMA
and MIMO system. The signals of different SCs are appropriately weighted and
summed using EGC [8-11], MRC [12-15], ORC or zero forcing (ZF) [16-18] and
threshold based ORC (TORC) [19]. The MRC technique represents the optimal
choice when the system is noise limited; in contrast, when the system is interference
limited, the ORC can completely eliminate MAI in the downlink transmission; yet, it
causes noise amplification in the receiver [20]. The MMSE [21-26] criterion may
also be used to derive the equalizer coefficients, whereas an even more powerful
optimization criterion is the minimum bit error ratio (MBER) criterion [27].
A lot of research studies have been dedicated to the error performance
evaluation of MC CDMA systems [28-31], and the approximation of the bit error
rate (BER) for MC CDMA with MRC, EGC, and MMSEC has been derived [28].
However, it is based on the law of large numbers, i.e. the spreading code length must
be sufficiently large. In [29], the authors have derived the exact BER of a
synchronous MC CDMA system with MRC. The BER is calculated based on a
moment generating function method with no assumption on the distribution of
multiple access interference. The BER performance of MC CDMA with MRC and
EGC has been evaluated over a Rayleigh fading channel with correlated envelopes
and phases [30]. A lower bound and a tight approximation on the BER of MC
CDMA with ORC is presented in [31].Although MRC, EGC and ORC only requires
the channel state information (CSI), the MMSE and MBER equalizers are more
complex because they exploit additional knowledge of SNR.
Partial equalization (PE) [32-34] is another technique to reduce the ISI,
where the signals are partially equalized to avoid noise and interference in the
receiver instead of complete equalization. A different approach has been proposed in
[35,36] in the form of combined equalization. This technique applies pre
equalization at the transmitter in conjunction with post equalization at the receiver
which requires channel knowledge at both the transmitter and receiver ends
increasing the complexity of the system. In [37] combined partial equalization was
discussed where partial equalization was considered at the transmitter and TORC at

11
the receiver. From the literature survey it is observed that MMSE equalization is
good choice inspite of its complexity. Many researchers concentrate only on
equalization but not on channel estimation. CSI is a prime factor for equalization to
reduce the amplitude and phase distortion.
Differential demodulation without channel estimation was used in [1] which
results in SNR loss of upto 3 dB. The channel tracking method using kalman filter
has been analyzed which results in better estimation accuracy but suffers from
computational complexity and delay [38, 39]. Various authors have reported the
effect of channel estimation error and phase noise [40-43]. Some authors have tried
channel estimation using known data at the receiver. It involves sending pilots
(symbols which are known to the receiver) with the data symbols so that the channel
can be estimated and therefore the data at the receiver.
Training sequence and optimal training based channel estimation is discussed
for MIMO system [44,45] in which with increasing number of transmitting and
receiving antennas, the length of the training sequence needs to be increased, leaving
less space for data transmission and reduced overall spectral efficiency. However the
optimal choice of training sequence is quite challenging. In [46, 47] the channel
estimation is totally avoided by introducing differential modulation at the expense of
a performance penalty, however the spectral efficiency of MIMO system depends
upon the channel estimation.
Most channel estimation methods for OFDM transmission systems have
been developed under the assumption of a slow fading channel, where the channel
transfer function is assumed stationary within one OFDM data block. In addition the
channel transfer function for the previous OFDM data block is used as the transfer
function for the present data block. In practice, the channel transfer function of a
wide band radio channel may have significant changes even within one OFDM data
block. Therefore, it is preferable to estimate channel characteristics based on the
pilot signal in each individual OFDM data block [48-50].

12
To increase the spectral efficiency less number of pilots operating in time
varying channels [51], dynamic pilot symbol arrangement with wiener filter [52] is
proposed. Channel estimation technique for MIMO OFDM is studied [53, 54],
concentrating only on OFDM but not on MIMO system. To avoid increasing the
length of pilot, optimal/training based channel estimation is discussed for MIMO
OFDM system [55-58], and MIMO CDMA [59] but real challenge is in arriving the
optimal number of pilots Incorporating pilots decreases the spectral efficiency of the
system and they usually require averaging over many symbols before obtaining a
good estimate which make the scheme inefficient for high speed data
communication.
The limitations in training based estimation techniques motivated interest in
the spectrally efficient blind approach [60-71]. They utilize certain underlying
mathematical information about the kind of data being transmitted. Existing blind
methods can be broadly categorized as statistical or deterministic: The former
methods rely on assumptions on the statistics of the input sequence [60], while the
latter make no such assumption [61]. In the first category, i.e., statistical approach,
blind channel estimation using second order statistics can potentially achieve
superior performance for a given time averaging interval than approaches based on
higher order statistics [62]. The second category, i.e., deterministic method, is
generally favored when the input statistics are unknown, or there may not be
sufficient time samples to estimate them. Several interesting deterministic methods
were discussed in [63]; however most of them are exclusively for SISO or single
carrier transmissions.
Amid second order statistics based blind approaches, subspace based
estimation are attractive since estimates can often be obtained in a simple form by
optimizing a quadratic cost function [64]. Without employing any precoder, a
subspace based method is proposed for OFDM systems by utilizing the redundancy
introduced by the cyclic prefix (CP) [65], and it is further extended for MIMO
OFDM systems in [66]. Virtual carriers (VCs) are SCs that are set to zero with no
information being transmitted, which can be used for channel estimation. Such a

13
scheme is proposed for OFDM systems [67] and further extended to MIMO OFDM
systems in [68].
Variants of the statistics based methods for example by inserting zero
padding instead of CP for each OFDM block or by introducing the so called
repetition index and remodulation on the received signal [69-71]. However, the
number of required time samples is still implicitly proportional to the size of the
IFFT in the OFDM modulator. It is also noted that deterministic approaches still
need to accumulate data samples to algebraically obtain channel estimates, and their
performance in noise improves as the number of samples increases. Therefore, as the
dimension of the parameter space is increased in the MIMO OFDM context, the
number of samples required for deterministic methods to achieve an acceptable level
of performance will also inevitably be increased. Eventhough these methods are
bandwidth efficient but are notoriously slow to converge, extremely computationally
intensive and hence are impractical to implement in real time systems.
Channel estimation based on semi blind techniques are proposed in order to
combine the advantage of pilot based channel estimation and semi blind approach
[72-76]. These methods use pilots to obtain initial CSI and blind approach methods
are used to improve the system performance. However obtaining optimal value of
pilot signals are difficult. In [73] authors develop a semi blind channel estimation
method for MIMO OFDM systems based on a specific and non redundant precoding
scheme, say, circular precoding, since the circular precoding allows channel
estimation at the receiver and simplifies the encoding scheme at the transmitter. In
[76], an efficient semi blind sparse channel estimation approach for MIMO OFDM
systems was proposed. Through the analysis of the second order statistics of the
received signal that passes through a sparse channel, a blind algorithm for obtaining
a constraint on the sparse channel vector with respect to the MSTs is developed. By
combining this blind constraint with a training based sparse LS criterion, a semi
blind solution to the estimation of the effective channel is then obtained.

14
It is possible to use the CSI for BF also, a representative MIMO technique
which is a promising solution that could be utilized to enhance throughput [77-80].
There are many available beam design algorithms for MIMO interference channels
such as interference alignment based algorithms [81-83] and sum rate targeted
algorithms [81-86]. However, most of these algorithms assume perfect CSI at
transmitters and receivers, whereas the assumption of perfect CSI is unrealistic in
practical wireless communication systems since perfect CSI is hard to obtain
practically due to channel estimation error, limited feedback or other limitations
[87]. Thus, the CSI error should be incorporated into the beam design to yield better
performance, and this is typically done under robust beam design frameworks.
One approach for multi antenna channels [88-90] is based on quantization of
the instantaneous CSI at the receiver followed by its conveyance to the transmitter
using a low rate feedback channel. This is a limited feedback approach, there are
also other approaches [91-93] which are based on feedback of partial (or statistical)
channel information. These methods, in general, do not perform well as the ones
using instantaneous feedback since they do not track the rapid fluctuations of the
channel. There are more general feedback schemes [94-96] that dynamically adapt
to the distribution of the channel, but these methods are generally too complicated to
implement in practice. In [95-97] and [98-104], the MIMO BF vector is quantized at
the receiver using a fixed codebook available at both the transmitter and the
receiver. There are two challenges in designing a MIMO BF system using quantized
feedback. One is how to design a good codebook, and the other is to decide on a BF
weight vector in a codebook efficiently. The unitary DFT matrix codebook and the
Grassmannian codebook were proposed in [97] and [89], respectively. In [101], an
iterative search algorithm that has linear complexity on average with respect to the
number of transmit antennas was proposed. Random BF(RBF) has been widely
studied due to its simplicity and lower feedback overhead, although its performance
is poorer than that of other orthogonal BF schemes that utilize full channel
information [105-108]. Since RBF uses only SINR feedback rather than the full
channel information, the feedback overhead can be efficiently reduced.

15
In MIMO OFDM, BF is performed independently on each SCs, which is
referred to as subcarrier wise beamforming [109,110], to exploit the spatial
diversity. Unfortunately, subcarrier wise beamforming induces high computational
complexity since a discrete fourier transform (DFT) operation is required for each
antenna element [111,112]. In addition, the optimized transmit weight vector for
every SC has to be fed back to the transmitter, which leads to the growth of feedback
overhead linearly in proportion to the number of SC. To reduce the overall
complexity and the feedback overhead, several schemes have been proposed for
OFDM systems with single or multiple antennas in the literature [111-115] and can
be classified as pre DFT beamforming or symbol wise beamforming which increase
the complexity of implementation because of BF operation in both transmitter and
receiver. The distributed BF schemes presented in [116-118] assume that the
transmitter, the receiver and the relay nodes all use a single antenna. As a result,
these schemes do not benefit from spatial processing at the nodes.
Relays can be deployed in wireless networks to extend coverage and improve
throughput[119]. Relays can be either full duplex or half duplex. Full duplex relays
can transmit and receive at the same time while half duplex relays cannot. Because
full duplex relays are difficult to implement, practical systems focus on half duplex
relays [120,121]. Relays are also classified based on how they process the received
signal. The most popular relay types are decode and forward (DF) [122-124] and
amplify and forward (AF) [125-127]. DF relays have higher computational
complexity due to the requirement of decoding the signals and are helpful only if
they can decode successfully the signals. Transparent to the modulation and coding
of the signals, AF relays can be used flexibly in heterogeneous networks comprising
of many nodes of different complexity or standards [114]. AF relays provide better
diversity performance and in some cases higher throughput (for example, when the
source relay channel is weak) than DF relays in single antenna half duplex relay
systems [128,129]. AF relays may be attractive in practice due their lower
complexity and faster signal processing.
MIMO techniques can be incorporated into relay systems by equipping each
node with multiple antennas to provide multiplexing and diversity gains [130].

16
Transceiver designs for traditional one way MIMO relay systems were proposed in
[131-135,155]. The optimal two way relay strategies were studied in [136,157] to
maximize achievable sum rate and to achieve optimal diversity multiplexing
tradeoff. For an analog network coding (ANC) based MIMO two way relay system
consisting of a multi antenna relay node and two single antenna source nodes, the
optimal relay BF was studied in [137,156] to characterize the achievable capacity
region, and also to minimize the mean squared error [138]. In [139] and [140], the
authors considered jointly optimizing the source and relay pre coders to maximize
sum rate for MIMO two way relay systems where all nodes are equipped with
multiple antennas. The joint optimization problem was also investigated in [141]
based on the MMSE criterion, and an iterative algorithm as well as a heuristic
algorithm based on channel parallelization was proposed therein. In [142] and [143],
the authors addressed joint source and relay optimization for multiple relay systems.
All of the above mentioned works assume perfect CSI is available. In practice, CSI
has to be estimated using pilot symbols, and channel estimation errors are inevitable
due to noise and the time varying nature of wireless channels. Hence, it is desirable
to provide robustness against channel uncertainties when designing transceiver.
Further Capacity bounds for MIMO relay channels are studied in the
literature [144,136]. Automatic repeat request techniques for MIMO relays are
proposed to achieve the optimal diversity multiplexing delay tradeoff [145] and to
improve throughput and outage probability performances [146]. MIMO relay with
multiuser linear precoding schemes in fixed relay networks have been considered in
[147,148] where CSI is assumed to be known at the transmitter. Relay gain matrix
optimization was studied in [131,132] to maximize the source destination capacity
and linear transceiver designs were proposed based on the MMSE criterion [135].
Full duplex relay discussed for MC CDMA system [149,150] is difficult to
implement due to the generation of self interference and different types of relaying
technique for MIMO OFDMA system is discussed in [151-154]. However, most of
MIMO relay techniques require high complex equalizers and also error free
forwarding at the relay to enjoy the diversity, which is impractical and thus may not
be suitable for low complexity communication environments.

17
1.4 OBJECTIVE OF THE RESEARCH WORK
Globally the demand for wireless application is increasing every year which
necessitates the design of progressive and perfect transmission of multimedia data over
wireless multipath channels. Even though, MIMO MC CDMA is chosen as the access
technique for realizing the same, quality offered by this technique fails to meet the
demands of future wireless communication system. To improve the performance of the
system, this work proposes the MPCE based MMSE equalization, ITBF and NR to
meet the quality of wireless communication such as high transmission rate, more
transmission range and transmission reliability.
1.5 THESIS ORGANIZATION
Chapter 1 provides an overview of CDMA, OFDM and MIMO techniques.
The objective of the research work, state of art survey relevant to the topic of
research and author‟s contribution in the related areas are also given. Organization
of the thesis is presented at the end of this chapter.
Chapter 2 discusses the different combining and equalization techniques. The
simulated BER performance of MIMO MC CDMA system with/without proposed
MMSE equalization under ideal channel estimation(ICE) for different antenna
configuration, SC and modulation are explained and performance of the proposed
system is compared with the existing MC CDMA system.
Chapter 3 deals with different channel estimation techniques and importance
of channel estimation for OFDM, MIMO OFDM and MIMO MC CDMA system.
Simulations are carried out for MIMO MC CDMA with MPCE based MMSE for
different antenna configuration, SC and modulation and the results are discussed in
detail.
Chapter 4 highlights BF techniques. BER Performance of MPCE based
MMSE equalization and ITBF for proposed MIMO MC CDMA is simulated and
compared with different modulation.

18
Different types of relaying performance improvement through proposed NR
technique is discussed in chapter 5. Impact of MPCE based MMSE equalization and
ITBF with NR on BER for MIMO MC CDMA is investigated through simulation.
Chapter 6 concludes the thesis by emphasizing the major aspects of the
study. A summary of major contribution and scope for further work is also included
in this chapter.

19
CHAPTER 2
PERFORMANCE OF MIMO MC CDMA SYSTEM WITH MMSE
EQUALIZATION
2.1 INTRODUCTION
Wireless communication system requires signal processing technique that
improves the link performance in hostile mobile radio environment. The mobile
radio channel is particularly dynamic due to multipath propagation and Doppler
spread. From Fig.2.1 it can be noted that rectangular multilevel pulses passing
through the channel will spread in time and the pulse for each symbol smears into
adjacent time slots due to multipath time delay spread causing ISI resulting in an
irreducible BER floor for mobile system. However, even if a mobile channel is not
frequency selective, the time varying Doppler spread due to motion creates an
irreducible BER that can be realized over a frequency selective channel. Hence ISI
has been recognized as the major obstacle to high speed data transmission over
wireless channel.
Fig.2.1 Example of ISI on received pulse in a Communication system
0
0 t
t
Ts
Sampling points
Individual pulse response
(Sum of pulse)
Inter symbol Interference
Sampling points
t
t0
0
Ts
Individual pulse response
0
0
1 0 0 0
1 0 0 0
Input Waveform
Sampling points
t
t

20
The use of OFDM and cyclic prefix mitigates the effect of time dispersion.
However, it is still necessary to remove the amplitude and phase shift caused by
multipath effect on the channel. The effect of ISI can be counteracted by different
measures such as time or frequency domain equalization. The function of equalizer
within a receiver is to compensate for the average range of expected channel
amplitude and delay variations, also it may be adaptive as the channel is generally
unknown and time varying.
2.2 DIFFERENT TYPES OF EQUALIZATION TECHNIQUE
Many different detection schemes have been proposed in the literature for
OFDM and MC CDMA signal detection [158,159,179]. Every detection technique is
characterized by its specific combining parameters. The basic detection technique
treat the signals from other users as interference and their combining parameters are
functions of multipath fading channel.
2.2.1 Equal Gain Combining (EGC)
The EGC weights equally each sub channel contribution and compensate
only the phases. In this case SCs are processed with the same amplification and
correlation is performed by the means of Walsh matrix. The EGC method would be
a suitable approach for an additive white Gaussian noise (AWGN) channel where all
transfer factors are equal but in a fading channel the transfer factors of the SCs have
different amplitudes [160-162]. For this reason the orthogonality of the Hadamard
codes is violated and interference between the information symbols in the received
signal creates large error floor.
2.2.2 Maximum Ratio Combining (MRC)
In this method the signals from all the SCs are weighted according to their
individual signal power to noise power ratios and then summed. The individual
signal should be cophased before being summed which generally requires individual
receiver and phasing circuit [163-170]. MRC produces an output SNR equal to sum
of the individual SNRs. Thus it has an advantage of producing an output with an

21
acceptable SNR even when none of the individual signals are themselves acceptable.
But this method performs well only when the signals are quite above noise level. In
this technique, if fading occurs in any one of the branches, a maximum gain is
applied to it which enhances the noise level and totally induce non orthogonality
between the codes.
2.2.3 Orthogonality Restoring Combining (ORC)
While EGC may be desirable for simplicity and MRC for combating noise,
neither of these techniques significantly exploit the coding of the SC. Good choice
is to restore the orthogonality between the sequence through ORC, which implies a
total cancellation of multiuser interference; but this method enhances the noise
because the sub channels with low SNR have higher weights as shown in
Fig.2.2.[171-173].
Fig.2.2 Noise Enhancement in ORC
2.2.4 Threshold Orc (TORC)
With TORC (or Controlled Equalization), an attempt at restoring the
orthogonality between users is made by normalizing the amplitudes of the SCs. As
the orthogonality of the users are encoded in the phase of the SCs, it is primarily
beneficial in the downlink where phase distortion for all users may be corrected and
in this, SCs above a certain threshold will be equalized and retained. This
constraint is added to prevent the amplification of SCs with small amplitudes that
may be dominated by a noise component. A threshold for amplitude of the transfer
factor is introduced to suppress the weak SCs. The bit error rate depends on the
Noise Enhancement
f
Channel Noise
f
Equalizer
f
Information
f
Information
and Noise
f

22
noise level and on interference due to SCs below the threshold. Both effects show a
different behavior. Interference depends on the probability of rejecting a SC and
dearly rises with threshold. The interference is especially responsible for the error
floor at high SNR. The noise level is reduced with rising threshold because fewer
weak SCs are accepted. At moderate and low thresholds the average noise level is
increased hence it is difficult to determine the optimum threshold in fast fading.
2.2.5 Partial Equalization (PE)
An equalizer with partial channel amplitude compensation is similar in
structure to a ORC. The only difference is that it does not force the multi user
interference to zero, instead the channel phase is fully compensated but the channel
amplitude is partially compensated [174]. In this technique coefficient Gm depends
on a parameter β as 𝐺 𝑚 =
𝐻 𝑚
∗
|𝐻 𝑚
∗ |1+𝛽, , −1 ≤ 𝛽 ≤ 1, where 𝐻 𝑚
∗
is the mth
channel co
efficient. Note that EGC, MRC and ORC can be viewed as particular cases for
which the parameter β assumes the values 0, −1 and 1 respectively. The key idea is
that if MRC and ORC are optimum in the extreme cases of noise limited and
interference limited system respectively then for each intermediate situation there
exists an optimum value of the parameter β which minimizes the mean bit error
probability (BEP) averaged over fading. The problem is to find the appropriate β
that minimizes the average bit error probability of the system and this system
compensates channel amplitude partially.
2.2.6 Pre Equalization
If information about the actual channel is a prior known at the transmitter,
pre equalization can be applied at the transmitter such that the signal at the receiver
appears non distorted and an estimation of the channel at the receiver is not
necessary. Information about the channel state can, for example, be made available
in Time division duplexing (TDD) schemes if the TDD slots are short enough such
that the channel of an uplink and a subsequent downlink slot can be considered as
constant and the transceiver can use the channel state information obtained from
previously received data.

23
An application scenario of pre equalization in a TDD mobile radio system
would be that the terminal station send pilot symbol in the uplink which are used in
the base station for channel estimation and detection of the uplink data symbol. The
estimated channel state is used for pre equalization of the downlink data to be
transmitted to the terminal station. Thus, no channel estimation is necessary in the
terminal station which reduces its complexity. Only the base station has to estimate
the channel, i.e. the complexity can be shifted to the base station.
A further application scenario of pre equalization in a TDD mobile radio
system would be that the base station send pilot symbol in the downlink to the
terminal station, which performs channel estimation. In the uplink, the terminal
station applies pre equalization with the intention to get quasi orthogonal user
signals at the base station receiver antenna. This results in a high spectral efficiency
in the uplink, since MAI can be avoided. Moreover, complex uplink channel
estimation is not necessary. The accuracy of pre equalization can be increased by
using prediction of the channel state in the transmitter where channel state
information from the past is filtered.
2.2.7 Combined Equalization
With combined equalization CSI is available at both the transmitter and
receiver. This enables pre equalization to be applied at the transmitter together with
post equalization at the receiver. The knowledge about the transmission channel can
be made available, for example, by exploiting TDD to gather CSI at transmitter
needed for pre equalization, and by performing channel estimation at Rx in order to
obtain CSI needed for post equalization.
2.2.8 Maximum Likelihood Sequence Estimator (MLSE)
The optimal equalizer with highest probability correctly detects the
transmitted sequence which is the maximum likelihood sequence estimator (MLSE).
It minimizes the sequence error probability, i.e. the data symbol vector error
probability, which is equivalent to maximizing the conditional probability  |dP r

24
that d
was transmitted given the received vector r. The estimate of d obtained with
MLSE is
 |P r dd=arg max d

(2.1)
with arg denoting the argument of the function. If the noise lN is additive white
Gaussian,
Equation (2.1) is equivalent to finding the data symbol vector d
that
minimizes the
Squared Euclidean distance between the received and all possible transmitted
sequences.
2 2
( , ) || ||r r A   d d
(2.2)
The most likely transmitted data vector is
2
min ( , )r dd=arg d

(2.3)
MLSE requires the evaluation of MK
squared Euclidean distances for the
estimation of the data symbol vector d

which will increase the complexity of the
system, if the number of user and data is increased.
2.2.9 Maximum Likelihood Symbol by Symbol Estimation (MLSSE)
The Maximum Likelihood Symbol by Symbol Estimation (MLSSE)
minimizes the symbol error probability, which is equivalent to maximizing the
conditional probability  ( )
|k
P d r that ( )k
d was transmitted given the received
sequence r.
The estimate of
( )ˆ k
d obtained by MLSSE is
 ( )
( ) ( )ˆ |k
k k
d P d

d
=arg max r
(2.4)

25
If the noise lN is additive white Gaussian the most likely transmitted data
symbol is
( )
( )
( ) 2
2
1ˆ exp ( , )k
k
k
d with same
reelization of
d r





 
  
 

d
d
=arg max d
(2.5)
The increased complexity with MLSSE compared to MLSE can be observed
in the comparison of (2.5) with (2.3). An advantage of MLSSE compared to MLSE
is that MLSSE inherently generates reliability information for detected data symbols
that can be exploited in a subsequent soft decision channel decoder.
2.2.10 Maximum Likelihood Detection
The optimum multi user detection technique exploits the maximum a
posteriori (MAP) criterion or the maximum likelihood criterion. In the previous
section, two optimum maximum likelihood detection algorithms are shown, namely
the MLSE, which optimally estimates the transmitted data sequence
(0) (1) ( 1)
( , ,...... )k T
d d d 
d and the MLSSE, which optimally estimates the transmitted
data symbol d(k)
. It is straightforward that both algorithms can be extended to a MAP
sequence estimator and to a MAP symbol by symbol estimator by taking into
account the a priori probability of the transmitted sequence and symbol respectively.
When all possible transmitted sequences and symbols respectively are equally
probable a priori, the estimator based on the MAP criterion and the one based on the
maximum likelihood criterion are identical. The possible transmitted data symbol
vectors are , 0,..... 1K
M   d , where MK
is the number of possible transmitted
data symbol vectors and M is the number of possible realizations of d(k)
.
2.2.11 Minimum Mean Square Equalizer (MMSE)
The ORC equalizer cancels all ISI but can lead to considerable noise
enhancement. The MMSE technique minimizes the expected mean squared error
between the transmitted symbol and the symbol detected at the equalizer output,

26
thereby providing better balance between ISI mitigation and noise enhancement as
shown in Fig.2.3 [175-178].
Fig.2.3 Noise Enhancement in MMSE
Equalization according to the MMSE criterion minimizes the mean square
value of the Error between the transmitted signal and the output of the equalizer.
The mean square error
,l l l l lS G R  
(2.6)
 2
l lJ E 
(2.7)
The mean square can be minimized by applying the orthogonality principle,
stating that the mean square error lJ is minimum if the equalizer coefficient ,l lG is
chosen such that the error l is orthogonal to the received signal
1
( ) ( )
0
k
k k
l
k
R H s n


 
i.e   0l lE R 

The equalization coefficient based on the MMSE criterion is
,
, 2 2
,
l l
l l
l l
H
G
H 




(2.8)
The computation of the MMSE equalization coefficients requires knowledge
about the actual variance of the noise 2
 . For SNR→ ∞, the MMSE equalizer
becomes identical to the ZF equalizer. To overcome the additional complexity for
the estimation of 2
 , a low complex sub optimum MMSE equalization can be
Noise Enhancement less than ZF
f
Information
f
Channel
f
Equalizer
f
Information
and NoiseNoise
f

27
realized. With sub optimum MMSE equalization, the equalization coefficients are
designed such that they perform optimally only in the most critical cases for which
successful transmission should be guaranteed. The variance 2
 is set equal to a
threshold λ at which the optimal MMSE equalization guarantees the maximum
acceptable BER. The equalization coefficient with sub optimal MMSE equalization
results in and requires only information about ,l lH
. The value λ has to be
determined during the system design. Among the various equalization methods for
MIMO and MC CDMA, MMSE is considered to be a good solution for data
recovery since it can effectively reduce the ISI and utilize the diversity of the
frequency selective channel.
,
, 2
,
l l
l l
l l
H
G
H 




(2.9)
2.3 MIMO MC CDMA SYSTEM WITH MMSE EQUALIZATION
From the literature survey it is observed that equalization under ICE is
carried out for OFDM and MC CDMA using single antenna. It has been observed
that in comparison with conventional single antenna system, significant capacity
gain is achievable when multiple antennas are used both at the transmitter and
receiver. Conventional equalizer techniques like EGC, MRC and ORC are suitable
for single carrier communication only. OFDM is the multicarrier technique used in
MIMO MC CDMA system and it depends on the orthogonality among SCs for
bandwidth saving and proper detection. Incorporating multiple antennas with
multiple paths causes different phase/amplitude shifts and loss of orthogonality.
EGC, MRC and ORC equalizer cannot be used for MIMO MC CDMA system as
these equalizer reduces the orthogonality of user thus creating additional
interference in the receiver and using ORC for MIMO will also increase noise in the
receiver. Increased noise/Interference weakens wireless signal which will affect the
quality of the link. In this work MMSE equalizer is proposed for MIMO MC
CDMA system, the simplified diagram of which is given in Fig.2.4. It is to be noted,
the novelty of this equalizer is that it considers different antenna signals as different
links basically and further different path of different links are considered while
equalizing the signal, which proves to be effective in counteracting the
amplitude/phase shift.

28
Fig.2.4 Simplified diagram of MMSE equalizer
Figs. 2.5 and 2.6 show the proposed model of MIMO MC CDMA transmitter
and receiver. In this system the signal from the users are multiplexed at the
transmitter and transmitted as one entity through the radio channel. In the
transmitter, data is first modulated using Binary phase shift keying (BPSK) or
quadrature phase shift keying (QPSK) modulation, multiplied by the spreading
sequence of the user and then modulated using an OFDM modulator. In the receiver,
the signals are received through multiple antennas and demodulated using FFT.
After OFDM demodulation the amplitude and phase distortions are adjusted using
MMSE equalization where the ISI is reduced improving the performance of the
system and the user data symbols are recovered by despreading with corresponding
spreading codes.
Consider a MC CDMA system having Nc SCs and s d
N N MIMO system,
the transmitted signal after modulation can be expressed as
2
1 1
( ) ( ) ( )cos( )
s c
b
sc s
N N
E
k n T s nN T
i k n
s t b i c u t iT t

  
    (2.10)
where Eb and Ts are the bit energy and symbol duration respectively, uTs(t)
represents a rectangular waveform with unit amplitude and pulse duration Ts,
bk(i) is the ith
transmitted data bit cn is the spreading code, S
N is the source(S) (or
transmitting) antenna, ωn = 2πf0+ 2π(n - 1)∆f is the frequency in radians of the nth
SC, and the frequency spacing is ∆f= 1/Ts. The received signal r(t) through
destination(D)(or receiving )antenna d
N is given by
2
1 1
. ( )cos( )( ) ( ) ( ) s
d c
b
c s T s n n
N N
E
n k nN T
i k n
u t iT tr t t h b i c  

  
     (2.11)
Transmitter
Receiver
with
MMSE

29
Fig. 2.5 Block diagram of MIMO MC CDMA transmitter
Fig. 2.6 Block diagram of MIMO MC CDMA receiver with MMSE equalization
MappingData CP
Insertion
IFFT P/S
PN
Sequence
S/P
PN
Sequence
Detector
CP
Removal
FFTP/S
MMSE
Equalization
Estimated
Data
S/P
29

30
where hn is the SC flat fading gain, φn is the SC fading phase and η(t) is AWGN with
single sided power spectral density N0. After phase compensation, the receiver
performs amplitude correction described by ( 1,.., )n cn N  , called equalizer
coefficient. In the literature, different equalizer coefficient expressions for MMSE
have been proposed for MC CDMA system [171-173]. After demodulation the
received signal on the nth
SC is given by [176]
0
( )cos( )
sT
n n n n n ny r t t dt D I      
(2.12)
where Dn is the desired signal and In is the multi user interference Components and
n
 is the noise component with zero mean and variance 0
/ 4s
N T .Denoting 1, ,n nc  ,
' ' ' '
1 2,[ , ..., ]cN    , 1 2[ , ,... ] .c
T
Ny y y y After equalization and despreading, the
decision variable 'T
U y  . Where
1
1
' y
yy b
R R 

(2.13)
Let 1, 2,
[ ..., ]c
T
N
    , 1, 2,[ ..., ]u
T
Nb b b b , d
C is a c uN N matrix with kth
column
being the spreading code for kth
user, and H is a diagonal matrix with nth
diagonal
element equals to n
h .
2
b s
c
E T
dN
y HC b  
(2.14)
Then the Gaussian distribution of the matrix 1b yR is
 1 1 2
b s
y c
E T
b dNR E b HC b   
  
 12
b s
c d
E T
N E b HC b
(2.15)

31
The matrix yy
R is
2 4
b s o s
cc
T
d d
E T N T
NN HC C H I
(2.16)
where cNI is an identity with c cN N .Substituting (2.15) and (2.16) into (2.13), the
set of equalizer coefficients for MMSE scheme for downlink as
 
12
2
1
.
c
c c
cb s b o
T
d d
N
N N
n N nE T E N
n
HC C H I h


  (2.17)
2.4 SIMULATION RESULTS AND DISCUSSION
Simulation of MC CDMA and MIMO MC CDMA system are performed
using MATLAB version 9.1 with BPSK/QPSK modulation and the simulation
parameter is shown in Table 2.1. The system is tested using 16, 64 and 128 SCs. To
understand the impact of MIMO and equalization, simulations are carried out for
2x2, 3x3 and 4x4 transmitting and receiving antenna respectively with and without
MMSE equalization
Table 2.1 Simulation Parameter of the system with MMSE Equalization
Spreading Codes Walsh Hadamard Code
Number of SCs 16/ 64 /128
Channel Rayleigh fading
Modulation BPSK / QPSK
Antennas 2x2/ 3x3/ 4x4
Equalization Technique MMSE

32
2.4.1 Performance of the system with MMSE equalization (BPSK)
Figs. 2.7, 2.8 and 2.9 show the BER performance of MIMO MC CDMA
system with MMSE and without MMSE for BPSK modulation with different
antenna configuration and SCs. In Fig. 2.7 the legend graph illustrates the BER of
MIMO MC CDMA without using MMSE and the line graph shows the BER of
MIMO MC CDMA using MMSE for 2x2 and 4x4 antennas. It can be inferred from
the graph equalization improves the performance due to the effective compensation
of amplitude/phase shifts caused by different links/different path and subsequent
reduction of ISI. It can also be noted that the increase in number of antennas
improves the performance due to the spatial diversity and multiplexing gain.
Fig.2.7 Performance of the system with/without MMSE (BPSK, 16 SCs)
0 5 10 15
10
-4
10
-3
10
-2
10
-1
10
0
E
b
/N
o
(dB)
BER
MC-CDMA
MIMO MC-CDMA (2 x 2)
MIMO MC-CDMA with MMSE (2 x 2)

33
0 2 4 6 8 10 12 14 16
10
-4
10
-3
10
-2
10
-1
10
0
Eb
/No
(dB)
BER
MC-CDMA
MIMO MC-CDMA w ith MMSE(4 x 4)
0 2 4 6 8 10 12 14 16
10
-4
10
-3
10
-2
10
-1
10
0
Eb
/No
(dB)
BER
MC-CDMA

34
Table 2.2 Impact of MMSE Equalization (BPSK) at target BER of 10-3
Antenna
con.
16 SC 64 SC 128 SC
without
Equalization
Eb/No (dB)
MMSE
Equalization
Eb/No (dB)
without
Equalization
Eb/No (dB)
MMSE
Equalization
Eb/No (dB)
without
Equalization
Eb/No (dB)
MMSE
Equalization
Eb/No (dB)
2 x 2 14.8 12 14 11.6 13.8 11
3 x 3 14 11 13.5 10.2 12.8 10
4 x 4 11 9 10.3 8 9.8 7.5
Figs.2.8 and 2.9 show the BER performance of the system using 64 and
128 SCs respectively. Comparing Figs. 2.7, 2.8 and 2.9, it can be inferred that the
increase in number of SCs improve the performance due to the frequency diversity
offered by the SCs. The impact of frequency and spatial diversity is depicted in the
Table 2.2. From the Table it is observed that for 16 SCs with 2x2 antenna
configuration the Energy per bits to spectral noise density (Eb/No) requirement
without equalization is 14.8 dB which is reduced to 14 dB and 13.8 dB for 64 SCs
and 128 SCs respectively. Similar behavior is also observed in 3x3 and 4x4 antenna
configuration for the target BER of 10-3
. Further the performance is enhanced by
equalization technique incorporated in the receiver, for example, Eb/No of 16 SCs
(4x4) with equalization is reduced by 2.8 dB than the system without equalization
due to reduction of ISI. The reduction of Eb/No is also observed in 64 SCs and
128 SCs i.e 2.4 dB for 64 SCs and 2.8 dB for 128 SCs.
The performance of the system is also tested with different antenna
configuration such as 2x2, 3x3 and 4x4. For 2x2(16SCs) antenna configuration the
Eb/No is 12 dB which is reduced to 11dB for 3x3 and 9 dB for 4x4 antenna
configuration. Comparing 2x2 with 4x4 antenna configuration the Eb/No is 3 dB less
at target BER of 10-3
.

35
2.4.2 Performance of the System with MMSE Equalization (QPSK)
The performance of the MIMO MC CDMA system is tested using QPSK
modulation and the numerical result is shown in the Table 2.3 when comparing
Figs. 2.7, 2.8 and 2.9 with Figs 2.10, 2.11 and 2.12 the latter figures shows slightly
lesser performance because of higher order modulation. From the Table 2.3 it is
observed that for 16 SCs (3x3) with equalization the Eb/No is 12.8 dB which
requires 2.6 dB less Eb/No than the system without equalization. For 64 SCs and
128 SCs too the Eb/No is 2.9 dB and 2.5 dB lesser respectively. The improved
performance can be obtained through different antenna diversity. For example in
2x2 (128 SCs) the Eb/No requirement with equalization is 12 dB which is reduced to
11.3 dB for 3x3 and 9.8 dB for 4x4 antenna configuration. The same improvement is
also obtained for 16 SCs and 64 SCs. From the above simulation results and
discussion, it is clear that the BER of the MIMO MC CDMA using MMSE for
BPSK and QPSK modulation is considerably reduced than the system without
equalization due to reduction of ISI in the receiver.
Fig.2.10 Performance of the system with/without MMSE (QPSK, 16 SCs)
0 2 4 6 8 10 12 14 16
10
-4
10
-3
10
-2
10
-1
10
0
Eb
/No
(dB)
BER
MC-CDMA

36
0 2 4 6 8 10 12 14 16
10
-4
10
-3
10
-2
10
-1
10
0
Eb
/No
(dB)
BER
MC-CDMA
0 2 4 6 8 10 12 14 16
10
-4
10
-3
10
-2
10
-1
10
0
Eb
/No
(dB)
BER
MC-CDMA
MIMO MC-CDMA (2 x 2 )

37
Table 2.3. Impact of MMSE Equalization (QPSK) at target BER of 10-3
Antenna
configuration
16 SCs 64 SCs 128 SCs
without
Equalization
Eb/No (dB)
MMSE
Equalization
Eb/No (dB)
without
Equalization
Eb/No (dB)
MMSE
Equalization
Eb/No (dB)
without
Equalization
Eb/No (dB
MMSE
Equalization
Eb/No (dB)
2 x 2 16 12.8 15 12.5 14.5 12
3 x 3 14.8 12.2 14.5 11.6 13.8 11.3
4 x 4 13 10.5 12.5 10.2 12 9.8
2.5 SUMMARY
In this chapter MMSE equalization with ICE is proposed for MIMO MC
CDMA system in frequency selective fading channels and tested. Simulation is
carried out for MC CDMA system, MIMO MC CDMA system without equalization
and MIMO MC CDMA system with MMSE equalization. The proposed system
shows improved performance compared to other existing technique due to the
restoration of orthogonality and reduction of ISI through MMSE equalization. It is
inferred that the system performance can further be improved by increasing the
frequency and spatial diversity through the increase in the number of SCs and the
antennas. From the simulation results it is concluded that the performance of MIMO
MC CDMA system with MMSE equalization increases by 23.45% and 18.3% for
BPSK and QPSK respectively compared to the system without equalization.

38
CHAPTER 3
PERFORMANCE OF MIMO MC CDMA SYSTEM WITH
MPCE BASED MMSE EQUALIZATION
3.1 INTRODUCTION
In wireless system, transmitted information reaches the receiver through a
radio channel. A channel can describe everything from the source to the sink of a
radio signal which includes the physical medium (free space, fiber, waveguides etc.)
between the transmitter and the receiver through which the signal propagates. The
signals from the transmitter is usually protected by the channel coding and
interleaved against fading phenomenon after which the binary signal is modulated
and transmitted over multipath fading channel. Due to the multipath channel signals
are distorted in a variety of ways like AWGN, amplitude, frequency and phase
distortion, resulting in ISI. In order to recover the original transmitted signal at the
receiver, it is essential to be aware of channel characteristics like reflection,
refraction and diffraction to ensure successful equalization and is possible with the
help of channel estimation. The general channel estimation procedure is shown in
Fig 3.1. Channel estimation is defined as the process of characterizing the effect of
physical channel on the input sequence. Channel estimation can be avoided by using
differential modulation technique, however such system result in low data rate and
there is a penalty of 3-4 dB less SNR [180-184]. In some cases channel estimation at
the user side can be avoided if the base station performs the channel estimation and
sends a pre distorted signal [185]. However, for fast varying channels, the pre
distorted signal might not bear the current channel distortion, causing system
degradation. Hence, system with channel estimation is needed for the future high
data rate system.

39
The CSI can be obtained through training based, blind and semi blind channel
estimation techniques. The blind channel estimation is carried out by evaluating the
statistical information of the channel and certain properties of the transmitted signals
[186]. It has no overhead but is applicable only to the slow fading channel due to its
need for a long data record. In training based channel estimation algorithm, training
symbols or pilot tones that are known prior to the receiver are multiplexed along with
the data stream for channel estimation [187]. Semi blind channel technique is a
combination of blind and training technique utilizing pilots and other natural
constraints to perform channel estimation.
Fig.3.1 Channel estimation procedure
3.2 TYPES OF CHANNEL ESTIMATION
This section discusses in detail the different channel estimation techniques.
3.2.1 Pilot Based Channel Estimation
The most preferred and used method to estimate the channel and the offset in
frequency is to use pilot symbols. Pilot symbols are known to the transmitter and
receiver in advance. The basic idea with pilot symbols is that there is a strong
correlation between the pilot symbol fading and the fading of data symbols that are sent
Channel
Estimated channel
model
Channel
Estimation
Technique
Transmitted
Sequence
x(n)
Estimated
signal
𝑌(n)
Error Signal
e(n)
Actual Received
signal Y(n)

40
close to the pilot symbol in time and SC. Different types of pilot arrangements are
discussed in literature such as block type pilot, group type pilot and irregular pilot
distribution [188-191]. In block type pilot based channel estimation shown in
Fig.3.2 (a), channel estimation symbols are transmitted periodically, in which all SCs
are used as pilots. If the channel is constant during the block, there will be no channel
estimation error since the pilots are sent in all carriers.
In comb type pilot based channel estimation shown in Fig.3.2 (b), an
efficient interpolation technique is necessary in order to estimate channel at data SC
by using the channel information at pilot SCs. In Fig.3.2 (a) and 3.2 (b) the pilots are
arranged in time and frequency domain respectively. In order to increase the channel
capacity irregular pilot arrangements [2, 192] discussed by authors is shown in the
Fig. 3.2 (c).
Fig.3.2 (a) Block type pilot arrangement
Fig.3.2 (b) Comb type pilot arrangement
Time
F
r
e
q
e
n
c
y
F
r
e
q
e
n
c
y
Time

41
Fig.3.2 (c) Irregular pilot arrangement
3.2.2 Blind Channel Estimation
In pilot or training based channel estimation the transmitter sends a signal
known prior to the receiver and is used for estimating the channel condition.
However, most wireless devices will be battery powered and the transmission of
training signals will seriously affect the robustness of such devices. Moreover,
training increases the overhead of the transmitted signal, reducing the net data
transmission rate. Thus, it is sensible to use blind channel estimation method that
possibly reduces the amount of training required significantly. Typically some
special property of the transmitted signal is exploited for blind channel estimation.
Blind equalization method provides attractive solution since they do not
require any known transmitted data for channel estimation and equalization purposes
[193-196] instead they use statistical and structural properties of the communication
signals. Channel identification or equalization requires information about both the
channel amplitude and phase responses that can be acquired from the received
signal. The higher order statistical properties gives phase information however
higher order statistics and spectra may not provide a feasible approach for
constructing practical equalizers. They have a large variance and consequently large
sample sets are needed in order to obtain reliable channel estimates. Even though
this is a drawback, can be used, in fast fading channel with high data rates and low
computational complexity requirement.
F
r
e
q
e
n
c
y
Time

42
3.2.3 Semi Blind Channel Estimation
Semi blind technique uses both pilot and natural constraints to estimate the
channel. These methods use pilots to obtain an initial channel estimate and improve
it by using a variety of prior information. Thus, in addition to the pilots, semi blind
methods use the cyclic prefix [197-199], time and frequency correlation [200],
Gaussian assumption on transmitted data [201] and virtual carriers [202] for channel
estimation and subsequent data detection.
3.3 CHANNEL ESTIMATION FOR OFDM SYSTEM
There are two main problems in designing channel estimators for wireless
OFDM system. The first problem is the arrangement of pilot information and the
second problem is the design of an estimator with both low complexity and good
channel tracking ability. In general, fading channel of the OFDM system can be
viewed as a two dimensional (2D) signal (time and frequency). The optimal channel
estimator in terms of mean square error is based on 2D Wiener filter interpolation.
Unfortunately, such a 2D estimator structure is too complex for practical
implementation due to need of power delay profile and as Doppler spread is not
known in wireless environment[252-254]. The combination of high data rate and
low bit error rate in OFDM system necessitates the use of estimators that have both
low complexity and high accuracy, where the two constraints work against each
other and a good tradeoff is needed. The one dimensional (1D) channel estimations
are usually adopted in OFDM system to accomplish the tradeoff between complexity
and accuracy [203-209].
The two basic 1D channel estimations are block type pilot channel estimation
and comb type pilot channel estimation in which the pilots are inserted in the
frequency and in the time domain as shown in Fig.3.3. The estimations for the block
type pilot arrangement can be based on least square (LS), MMSE and modified
MMSE. The estimations for the comb type pilot arrangement includes the LS
estimator with 1D interpolation, the maximum likelihood (ML) estimator, and the
parametric channel modeling based estimator. Other channel estimation such as

43
estimators based on simplified 2D interpolations, the estimators based on iterative
filtering and decoding, estimators for the OFDM system with multiple transmit and
receive antennas are studied in [210-213].
Fig.3.3 Block diagram of OFDM transceiver with pilot based channel
estimation
3.4 PILOT BASED CHANNEL ESTIMATION FOR MC CDMA SYSTEM
The block diagram of MC CDMA system incorporating channel estimation
and equalization is shown in Figs.3.4 and 3.5. In the transmitter after modulation the
data stream is spread by the wideband PN sequence to form the chip level transmit
matrix. The pilot signals are added for estimating channel response in the receiver
and modulated by the IFFT. Subsequently the output samples of the IFFT are
parallel to serial (P/S) converted to form the baseband signal which is added with
cyclic prefix (CP).
Upon receiving the signal, CP is removed and FFT is performed and the pilot
symbols are used for obtaining transfer function of the channel which in turn is used
for reducing the amplitude and phase distortion using equalization. With the
knowledge of estimated channel, data symbols are recovered after despreading the
signal with help of PN sequence.
AWGN
Data Out
Channel
Estimation
Mapping
Serial to
parallel
Pilot
insert
IFFT
Add Cyclic
Prefix
Parallel
to serial
Channel
Demap
Serial to
Parallel
Remove
Cyclic
Prefix
FFT
Pilot
Extract
Parallel
to Serial
Data In

44
Fig.3.4 Block diagram of MC CDMA transmitter with pilot signal
Fig.3.5 Block diagram of MC CDMA receiver with pilot based channel
estimation
3.5 MODIFIED PILOT CHANNEL ESTIMATION (MPCE) BASED
MMSE EQUALIZATION FOR MIMO MC CDMA SYSTEM
From the literature survey it is observed that different types of equalizations
are discussed for MIMO, OFDM and MC CDMA without channel estimation to
reduce ISI. Equalization can be effected in a better way if the channel characteristics
are estimated a priori through estimation techniques. Pilot based channel estimation
is proved to be effective with less estimation error and the performance is decided
basically on the number and location of pilots. In this work, MPCE technique is
proposed and incorporated, in which, the location and number of pilots in time and
frequency domain is dynamically varied with the help of channel estimation itself.
Utmost care is taken to ensure that the pilots are Nyquist rated all the time based on
the frequency selectiveness and fastness of the fading. Also, delay spread and
Doppler spread threshold(Th) is decided based on the traffic and target BER
required. This modification shows a clear improvement in performance and is
observed through simulation. A simple flow chart of proposed MPCE is depicted in
Fig.3.6 (a) and (b).
MappingData CP
Insertion
IFFT P/S
PN
Sequence
PilotS/P
FFT
PN
Sequence
Detector
CP
Removal
P/S Equalization
Estimated
Data
S/P
Channel
Estimation

45
The equalization is modified to adjust the amplitude and phase distortion
with the help of CSI obtained from MPCE technique. Fig.3.7 and 3.8 shows the
proposed model of MIMO MC CDMA transmitter and receiver respectively with the
modified pilot channel estimation and equalization. The transmitter of MIMO MC
CDMA consists of BPSK/QPSK modulator, PN sequence spreader and OFDM
modulator. After modulating, the data stream is multiplied by a spreading sequence
and the length of this spreading code is usually identical to the number of SC. The
pilot signals are first multiplexed to the data streams then OFDM modulation is
performed which is transmitted through multiple antennas.
Fig. 3.6 (a) Flow chart of MPCE (Delay spread). (b) MPCE (Doppler spread)
NoYes
Yes
Decrease
number of pilot
Delay
spread =Th
CSI
Increase
number of pilot
Delay
spread >Th
No
NoYes
Yes
Decrease
number of pilot
Doppler
spread =Th
CSI
Increase
number of pilot
Doppler
spread >Th
No

46
Fig.3.7 Block diagram of modified pilot insertion based transmitter
Fig.3.8 Block diagram of MPCE based MMSE equalization
PN
Sequence
FFTDetector
CP
Removal
P/S MMSE
Equalization
Estimated
Data
S/P
MPCE
MappingData
PN
Sequence
Pilot insertion
based on MPCE
S/P CP
Insertion
IFFT P/S
46

47
In the receiver the signals are received through multiple antennas and
demodulated using FFT. After OFDM demodulation the user data symbols and pilot
symbols are recovered by despreading with corresponding spreading codes. The
required transfer function for channel estimation and equalization is recovered from
pilot sequence. Finally the original data stream is recovered by dividing the
received signal by channel response. At the receiver end, the demodulator processes
the channel equalized waveform and reduces each waveform to a scalar (or) a vector
that represents an estimation of the transmitted data symbol. The detector, which
follows the demodulator, decides whether the transmitted bit is 0 or 1. The received
signal after FFT is given by
( ) ( ) ( ) ( ), 0,1,.., 1cY k X k H k W k k N    (3.1)
The pilot channel estimation procedure is shown in Fig.3.9, where the
received pilot signals YP(k) are extracted from Y(k), the channel transfer function
𝐻(k) can be obtained from the information carried by 𝐻P(k) with the knowledge of
the channel responses H(k), the transmitted data samples X(k) can be recovered by
simply dividing the received signal by sample response. Finally the signals are
equalized by MPCE based MMSE technique.
Fig.3.9 MPCE estimation procedure
Pilot signal
Extraction
Pilot SC
Estimation Channel
Interpolation
Received
Signal after
FFT
Estimated
Channel
Response… … … …
Known
pilots
XP(k)
Y(k) YP(k) 𝐻P(k) 𝐻(k)

48
3.6 SIMULATION RESULTS AND DISCUSSION
MIMO MC CDMA system with diversity technique is simulated using
MATLAB with the parameters given in Table 3.1. The result shows the BER
performance with respect to Eb/No of the system with/ without MPCE based MMSE
equalization under Rayleigh fading channel. The diversity technique uses 2, 3 and 4
antennas for both transmitter and receiver. The result indicates that with diversity,
performance of the system improves in terms of BER. Along with the diversity
technique the system was tested with different SCs (16, 64, and 128) and different
modulation techniques (BPSK, QPSK & 16QAM). From the result it is observed that
when the number of SCs is increased, the performance of the system gets increased
due to reduction of ISI. As the number of SCs increases, the frequency diversity also
increases which in turn reduces the error rate.
Table 3.1 Simulation parameters of the system with MPCE based MMSE
equalization
Spreading Codes Walsh Hadamard Code
Number of SCs 16/64 /128
Channel Rayleigh fading
Modulation BPSK/QPSK/16 QAM
Antennas 2x2/ 3x3/ 4x4/2x3/3x2
Estimation/Equalization technique MPCE/MMSE

49
3.6.1 Performance of the System with MPCE Based MMSE Equalization
(BPSK)
Fig.3.10 shows the BER performance of MIMO MC CDMA with/without
MPCE based MMSE equalization for BPSK modulation with different antenna
configuration. From the graphs it is evident that the system with MMSE equalization
and pilot based channel estimation or MPCE performs better due to the reduction of
ISI. The same system is tested for different SCs also, which is shown in Fig. 3.11
and 3.12 and it is observed that the performance of the system gradually increases
due to the increase in number of SCs i.e as frequency diversity increases the ISI
reduces by reducing the depth of the fading. The detailed report of the simulation
result is shown in the Table 3.2. For 16 SCs (4x4) the Eb/No requirement is 9.7 dB at
which is reduced to 9.2 dB for 64 SCs and 8.4 dB for 128
SCs.
Fig.3.10 Performance of the system with MPCE based MMSE (BPSK, 16 SCs)
0 5 10 15
10
-4
10
-3
10
-2
10
-1
10
0
Eb
/No
(dB)
BER
MIMO MC-CDMA (2 x 2) with MMSE & MPCE
MIMO MC-CDMA (4 x 4) with MMSE & MPCE

50
Fig.3.11 Performance of the system with MPCE based MMSE (BPSK, 64 SCs)
Fig. 3.12 Performance of the system with MPCE based MMSE (BPSK, 128 SCs)
0 5 10 15
10
-4
10
-3
10
-2
10
-1
10
0
Eb
/No
(dB)
BER
MIMO MC-CDMA(2 x2 )
MIMO MC-CDMA(4 x4 )
MIMO MC-CDMAwith MMSE & MPCE (2 x2 )
MIMO MC-CDMAwith MMSE & MPCE (4 x4 )
0 5 10 15
10
-4
10
-3
10
-2
10
-1
10
0
Eb
/No
(dB)
BER
MIMO MC-CDMA with MMSE & MPCE (2 x 2 )

51
Table 3.2 Impact of MPCE based MMSE equalization (BPSK) at target
BER of 10-3
Antenna
confi.
16 SCs 64 SCs 128 SCs
without
Equalization
Eb/No (dB)
MPCE
based
MMSE
Eb/No (dB)
without
Equalization
Eb/No (dB)
MPCE
based
MMSE
Eb/No (dB)
without
Equalization
Eb/No (dB)
MPCE
based
MMSE
Eb/No (dB)
2 x 2 14.8 13.2 14 12.2 13.8 11.8
3 x 3 14 12.5 13.5 11.2 12.8 10.6
4 x 4 11 9.7 10.3 9.2 9.8 8.4
It is also noticed that when the number of antennas are increased, the
performance of the system is increased due to the spatial diversity. For example in
Table 3.2 ,the Eb/No of 128 SCs (2x2) is 11.8 dB and it is reduced to 10.6 dB for
3x3 antenna configuration and 8.4 dB for 4x4 antenna configuration, moreover the
similar behavior is also observed in 64 SCs and 16 SCs. From the above discussion
it is clear that the performance of the system is improved by increasing the antenna
configuration and SCs. However the system performance of MIMO MC CDMA
system with MPCE based MMSE equalization is less compared with ICE discussed
in chapter 2, however the ICE is practically not possible.
3.6.2 Performance of the System with MPCE Based MMSE Equalization
(QPSK)
The performance of the proposed system is tested for QPSK modulation and
the simulation result is shown in Figs. 3.13, 3.14 and 3.15 Table 3.3. Comparing
Figs. 3.10, 3.11, 3.12 and Table 3.2 with Figs 3.13, 3.14, 3.15 and Table 3.3, it is
evident that the system with QPSK has lesser performance due to the higher order
modulation used. However in this case also performance improves due to increase in
SC and antenna configuration.

52
Fig.3.13 Performance of the system with MPCE based MMSE (QPSK, 16 SCs)
Fig.3.14 Performance of the system with MPCE based MMSE (QPSK, 64 SCs)
0 5 10 15
10
-4
10
-3
10
-2
10
-1
10
0
Eb
/No
(dB)
BER
0 5 10 15
10
-4
10
-3
10
-2
10
-1
10
0
Eb
/No
(dB)
BER
MIMO MC-CDMA w ith MMSE& MPCE(2 x 2 )
MIMO MC-CDMA w ith MMSE& MPCE(4 x 4 )

Performance improvement of mimo mc cdma system using equalization, beamforming and relay

Performance improvement of mimo mc cdma system using equalization, beamforming and relay

Recommandé

Recommandé

Contenu connexe

Tendances

Tendances (20)

Similaire à Performance improvement of mimo mc cdma system using equalization, beamforming and relay

Similaire à Performance improvement of mimo mc cdma system using equalization, beamforming and relay (20)

Plus de Tamilarasan N

Plus de Tamilarasan N (15)

Dernier

Dernier (20)

Performance improvement of mimo mc cdma system using equalization, beamforming and relay