461 470

ISSN: 2278 1323
International Journal of Advanced Research in Computer Engineering & Technology

Volume 1, Issue 4, June 2012

461
All Rights Reserved 2012 IJARCET

An Overview of PAPR Reduction
Techniques for an MC-CDMA System
Gagandeep Kaur, Rajbir Kaur

Abstract- MC-CDMA is the most promising technique for
high bit rate and high capacity transmission in wireless
communication. One of the challenging issues of MC-CDMA
system is very high PAPR due to large number of sub-carriers
which reduces the system efficiency. This paper describes the
various PAPR reduction techniques for MC-CDMA system.
Criterion for the selection of PAPR reduction technique and
also the comparison between the reduction techniques has been
discussed.
Index Terms- Complementary Cumulative Distribution
Function (CCDF), Multicarrier Code Division Multiple Access
(MC-CDMA), Peak-to-Average Power Ratio (PAPR),
I. INTRODUCTION
Multicarrier systems like CDMA and OFDM are now
days being implemented commonly. MC-CDMA (or
OFDM-CDMA) multiple access has become a most likely
technique for future generation broadband wireless
communication system such as 4G. This scheme is a
combination of both OFDM and CDMA that can provide
protection against frequency selective fading and time
dispersion. The CDMA part of this scheme provides
multiple access ability as well as spread each user signal
over the frequency domain to reduce the impact of
frequency selective fading. On the other hand OFDM
provides spreading across time domain of each spreading
codes chip which reduces the impact of inter-symbol
interference. This achieves in fulfillment of high data rate
transmission, so the number of subcarrier and spreading
codes must be carefully selected according to worst channel
conditions [1][2].
Although MC-CDMA is a powerful multiple access
technique but it is not problem free. OFDM signal has large
peak to average ratio (PAPR) which severely limits its
applications, and as long as basic operation of OFDM-
CDMA is identical to OFDM system, this undesirable

Manuscript received May 2012
Gagandeep Kaur, pursuing M.Tech(ECE), University College of
Engineering, Punjabi University Patiala, Patiala, India, 0171-2551636,
(email: gaganksodhi@gmail.com)
Rajbir Kaur, Assistant Professor (ECE), University College of
Engineering, Punjabi University Patiala, Patiala, India, +919779160093,
(email:rajbir277@yahoo.co.in)
property remains. High PAPR values causes a serious
problem to the power amplifier (PA) used at transmitter.
The power efficiency performance at such amplifiers
decreases as PAPR increases. Therefore signal suffers from
non-linear distortion at transmitter and degrades BER
performance at receiver. This forces the use of power
amplifier with large linear range which translates into
higher cost.
Therefore it is desirable to reduce PAPR by means of
PAPR reduction schemes. There are number of schemes
to deal with the issue of PAPR. such as, Signal
Distortion, Coding, and Symbol Scrambling techniques.
Signal distortion schemes reduce the amplitude by
linearly distorting the MC-CDMA signal at or around the
peaks. This includes techniques like clipping, peak
windowing, and peak cancellation. It is the simplest
technique but it causes in-band and out-band distortion.
Scrambling scheme is based on scrambling each MC-
CDMA signal with large PAPR. It includes techniques
such as Selected Mapping (SLM), Partial Transmit
Sequence (PTS). In case of PTS technique, MC-CDMA
sequences are partitioned into sub-blocks and each sub-
block is multiplied by phase weighting factor to produce
alternative sequences with low PAPR. However large
number of phase weighting factors increases the
hardware complexity and makes the whole system
vulnerable to the effect of phase noise. The SLM
technique pseudo-randomly modifies the phases of the
original information symbols in each OFDM block
several times and selects the phase modulated MC-
CDMA with best PAPR performance for transmission. In
Code allocation techniques, based on the orthogonal sets
of spreading sequences PAPR reduction schemes are
developed that excludes MC-CDMA signal with large
PAPR.
II. SYSTEM DESCRIPTION

A. MC-CDMA system model
The transmitter model of downlink MC-CDMA system is
shown in fig.1. The block of M complex valued symbol,
( ) ( ) ( ) ( )
1 2
[ , ,..., ]
k k k k
M
d d d d = of the k
th
user is firstly serial-
paralleled into a M-length vector as
( ) ( ) ( ) ( )
1 2
[ , ,..., ]
k k k k T
M
d d d d =
, k=1,2,,K. After this serial to parallel conversion, each
symbol is frequency domain spread by a user spreading
ISSN: 2278 1323

462

code
( ) ( ) ( ) ( )
1 2
[ , ,..., ]
k k k k
L
c c c c = , where L denotes length of the
spreading code. As spreading sequences, orthogonal sets of
sequences are preferred for reducing low multiuser
interference. Walsh-Hadamard (WH) sequences are used as
spreading sequences in this procedure. The input of K users
is summed up and is interleaved in frequency domain as
( )
0 1 1
1
, ,..., ] [
K
k T
N
k
X X X X X

=
= =
to obtain the diversity

effects by placing the chip symbol elements that
corresponds to same symbol on subcarrier that are distance
M apart. After frequency interleaving the symbol elements
are fed into IFFT block of size N M L = for OFDM
operation. The resultant baseband signal for one MC-
CDMA symbol is represented as
( ) ( ) 2 ( ( 1) ( 1)) /
1 1 1
1
( )
K M L
k k j M l m t NT
m l
k m l
d c e
N
s t
t +
= = =
=

(1)

Fig.1. MC-CDMA transmitter model
B. PAPR
Although MC-CDMA is a powerful multiple access
technique but it is not problem free. OFDM signal has large
peak to average ratio (PAPR) which severely limits its
applications, and as long as basic operation of OFDM-
CDMA is identical to OFDM system, this undesirable
property remains. When in time domain all the N
subcarriers are added up constructively, they produce a
peak power that is N times greater than the average power
of the signal. PAPR can be defined as the ratio between
instantaneous power and the average power of multicarrier
signal. The PAPR is calculated by the following equation:
10
2
max ( )
10log
2
( )
peak
average
s t
p
PAPR
p
E s t
= =
(
(

(
(
(2)

where peak p represents output peak power, average p means
output average power. E| | denotes the expected value.
Now to understand how and where the PAPR of a signal
becomes problem let us consider the power amplifier
characteristic curve shown in figure 2. It shows two
operating regions, the linear and the saturation region. For
better performance of the system the amplifier should work
in the linear region. But the problem appears when the
peaks of MC-CDMA signal are too large and do not fit in
the linear region. In this case they are treated by the
saturation region so they are non-linearly modified, causing
inter-modulation among subcarriers and out-of band
radiation.

Fig. 2. Non-linear amplifier characteristic curve
To hold these peaks the operational amplifier (OA) of the
transmitter needs a very large dynamic range which is very
expensive and it is inefficient in this case. Besides, the
larger the OAs dynamic range is, the more battery is
consumed, which is another reason why increasing the
dynamic range of OA is not a good solution. Therefore it is
desirable to reduce PAPR by means of PAPR reduction
schemes.
C. Cumulative Distribution Function
As PAPR is a random variable, an adequate statistic is
needed to characterize it. The cumulative distribution
function (CDF) is one of the most regularly used
parameters, which is used to measure the efficiency on any
PAPR technique. Normally, the Complementary CDF
(CCDF) is used instead of CDF, which helps us to measure
the probability that the PAPR of a certain data block
exceeds the given threshold [20].
The CDF of the amplitude of a sample signal is given by
( ) 1 exp( ) F z z =
(3)

The CCDF of the PAPR of the data block is desired is
our case to compare various reduction techniques. This is
given by [21]:
( ) 1 ( ) P PAPR z P PAPR z > = s
1 ( )
N
F z =
ISSN: 2278 1323

463

1 (1 exp( ))
N
z =
(4)

Fig. 3. CCDF of MC-CDMA signals with different number of
subcarriers
III. PAPR REDUCTION TECHNIQUES
Several PAPR reduction techniques have been proposed
in the literature. There are number of schemes to deal with
the issue of PAPR which are divided into two groups, i.e.,
Signal Distortion and Signal Scrambling techniques.

Fig. 4. Block diagram of PAPR reduction techniques
A. Clipping and Filtering
The simplest technique of PAPR reduction might be
amplitude clipping [1]. Amplitude clipping limits the peak
envelope of the input signal to predetermined value or
otherwise passes the input signal through unperturbed, i.e.,
transmitted signal is clipped at amplitude A as follows
( )
( )
( )
A if x A
y x if A x A
A if x A
<
= s s
>
(5)
where x denotes the signal before clipping and y denotes
the signal after clipping.

Fig. 5. MC-CDMA system using clipping
The distortion caused by amplitude clipping can be
viewed as another source of noise. The noise caused by
amplitude clipping falls both in-band and out-band. In-band
distortion cannot be reduced by filtering and results in error
performance degradation, while out-band distortion reduces
the spectral efficiency [5]. Filtering after clipping can
reduce the out of band radiation but, but may also cause
some peak re-growth so that signal after clipping and
filtering will exceed clipping level at some points [4]. To
reduce the overall peak re-growth, a repeated clipping and
filtering operations can be used. Generally, repeated
clipping and filtering takes much iteration to reach desired
amplitude level and thus it will increase the computational
complexity of the transmitter.
There are few techniques to mitigate the harmful effects
of the amplitude clipping. A method to iteratively
reconstruct the signal before clipping is proposed [6]. This
method is based on the fact that effect of clipping noise is
mitigated when decisions are made in frequency domain,
the signal is recovered somewhat from the harmful effects
of clipping. Another way to compensate for performance
degradation from clipping is to reconstruct the clipped
samples based on the samples in the oversampled signals.
Oversample signal reconstruction is used to compensate for
SNR degradation due to clipping for low values of clipping
threshold.
B. Companding
Companding is the process of compressing amplitude
signal at the transmitter and expanding them at the receiver.
In law companding technique the signals with lower
amplitude are amplified with greater gain and keeping the
peak signals unchanged and thus increasing the average
power of the transmitted signals resulting in increasing the
ISSN: 2278 1323

464

saturation region of HPA to make the system performance
worse [29]. The non-linear transform function is given as
[29]
2
2
( ) 6 1 exp
2
x
C x o
o
( | |
=
( |
\ .
(6)
This non-linear companding transform of (6) belongs to
exponential companding scheme. Based on this, there are
two types of non-linear companding techniques- which are
based on error function and exponential function [30]-[31].

Fig. 6. MC-CDMA system using companding
Error and exponential companding schemes have much
less impact on the original power spectrum of compared to
law companding scheme. It is a major reason that the
error and exponential companding schemes not only
enlarge the small amplitude signals but also enlarge the
large amplitude signals, while maintain the average power
unchanged by properly choosing parameters which can
increase the immunity of small amplitude signals from
noise.
Nonlinear companding transform is a type of nonlinear
process that may lead to significant distortion and
performance loss by companding noise. Companding can
be defined that noises are caused by peak re-growth after
DAC to generate in-band distortion and out-band noise, by
the excessive channel noises magnified after inverse non-
linear companding transform. For out-band noise it need to
be filtered and oversampled. For in-band distortion and
channel noises magnified, they need to iterative estimation.
Unlike AWGN, companding noise is generated by a
process known and that can be recreated at the receiver and
subsequently be removed.
In [32], simulation results has shown that companding
scheme could offer better system performance in terms of
PAPR reduction, power spectrum, BER for MC-CDMA
system than OFDM system.
C. Selected Mapping (SLM)
In the SLM technique, the transmitter generates a set of
sufficiently different candidate data blocks, all representing
the same information as the original data block, and selects
the most favorable for transmission. In this the input data
structure is multiplied by random series and resultant series
with the lowest PAPR is chosen for transmission. To allow
the receiver to recover the original data, the multiplying
sequence can be sent as side information. [7]- [8]
The block diagram of MC-CDMA system with SLM
technique is shown in fig.2. The input data sequences of
each user
( ) ( ) ( ) ( )
1 2
[ , ,..., ]
k k k k
M
d d d d = with length M are first
converted into M parallel data sequences
( ) ( ) ( ) ( )
1 2
[ , ,..., ]
k k k k
L
c c c c = and then each S/P converted output
is multiplied with the spreading code with length L.
Multiplexed symbol sequences
( )
0 1 1
1
, ,..., ] [
K
k T
N
k
X X X X X

=
= =
are multiplied by U-1

different phase sequences
0 1 1
[ , ,..., ]
u u u u
N
b b b b

= whose
length is equal to the number of carriers before IFFT
process resulting in U-1 modified data blocks
1
( ) ( ) ( )
0 0 1 1 1 1
0
[ , ,..., ]
U
u u u u T
n n N N
u
S X b X b X b X b

=
= =
. After the IFFT

process, the PAPR is calculated for U-1 phase rotated
symbols sequences
1
( ) ( ) ( )
0 0 1 1 1 1
0
( ) [ , ,..., ]
U
u u u u T
n n N N
u
s t x b x b x b x b

=
= =
and one original

sequence and then the symbol sequence with lowest PAPR
is selected for transmission and the corresponding selected
phase sequence
{ }
{ }
1
(0) (2) ( 1) ( ) ( )
0 1
(1), (2),..., ( 1) 0
, ,..., argmin max
U
U u v
n
n N
b b b U u
b b b b x
s s
=
| |
=
|
\ .

is
also transmitted to the receiver side for transmission.

Fig. 7. MC-CDMA system using SLM
In [9], a design criterion for phase sequence in SLM is
proposed. The proposed design criteria are that the
sequence having low average and large variance of PAPR
of the phase sequences themselves are good as phase
sequences in SLM, because the occurrence probability of
low PAPR becomes high because of small average and
large variance of PAPR of the sequences themselves. In
[10], Walsh hadamard transform is used as a phase
sequences and the simulation results shows that PAPR can
be reduced significantly proportional to an increase of the
number of different phase sequences. In [11], a new Pseudo
Random Interferometry code sequence is used as phase
sequence for SLM to reduce PAPR in MC-CDMA and the
simulation results shows that the proposed phase sequences
can effectively reduce the PAPR compared with Walsh
hadamard sequences and golay sequences.
D. Partial Transmit Sequence (PTS)
ISSN: 2278 1323

465

This method is based on phase shifting of sub-blocks of
data and multiplication of data structure by random vectors.
The main purpose behind this method is that the input data
frame is divided into non-overlapping sub-blocks and each
sub-block is phase shifted by constant factor to reduce
PAPR [9]. In the PTS technique, an input data block of N
symbols is partitioned into disjoint sub-bocks. The
subcarriers in each sub-block are weighted by phase factor
for that sub-block. The phase factors are selected such that
the PAPR of the combined signal is minimized. There are
three different kinds of sub-block partitioning schemes on
which the performance of PTS technique depends: adjacent,
interleaved, and pseudorandom. [12]- [13]
In the conventional PTS scheme in MC-CDMA, before
applied to IFFT operation, X, i.e., the sum of all active user
sequences after spreading and frequency interleaving is
input to PTS model as shown in fig.2. The data block, X
n
,
n=0,1,,N-1 is defined as vector, 0 1 1 [ , ,...., ]
T
N X X X X = .
Then partition X into V disjoint sets, represented by vectors
X
(v)
, v=1,2,,V such that
( )
1
V
v
v
X X
=
=
(7)

Fig. 8. Block diagram of conventional PTS scheme
Here, it is assumed that adjacent, interleaved and pseudo
random sub-block partition scheme are used and each sub-
block has equal size. The objective of PTS approach is to
form weighted combination of V clusters.
( ) ( )
1
V
v V
v
S b X
=
=
(8)
where b
(v)
, v=1,2,,V are weighting factors or phase
factors and are assumed to be pure rotations. After
transferring in time domain equation (6) becomes
( ) ( )
1
( )
V
v V
v
s t b x
=
=
(9)
The vector x
(v)
called partial transmit sequence is the
IFFT of X
(V)
. The weighting factors are chosen to minimize

the PAPR by searching for the appropriate combination of
each cluster and by corresponding weighting clusters.
{ }
{ }
(1) (2) ( ) ( ) ( )
0 1
(1), (2),..., ( ) 1
, ,..., argmin max
V
V v v
n
m N
b b b V v
b b b b x
s s
=
| |
=
|
\ .

(10)
The combination with weighting factors is called rotation
factor or combining sequence. Optimized transmit sequence
is
( ) ( )
1
( )
V
v v
m
v
s t b x
=
=
(11)
In [14], the PAPR reduction in synchronous downlink
MC-CDMA is investigated with different number of active
users and walsh hadamard and golay complementary
sequences are used as spreading sequences. The simulation
results shows that with increase in number of sub-blocks
and number of admitted angles there is a considerable
improvement in PAPR performance but it also results in
increase in computational complexity. Furthermore, to
compromise between PAPR reduction and system
complexity, suboptimal algorithms are used. In [15], a
novel grouping scheme is proposed in which the users are
partitioned into sub-groups and generate multiple
alternatives to achieve considerable PAPR reduction. The
simulation results showed that the proposed technique is
more effective than conventional PTS in MC-CDMA.
E. Spreading Code based Techniques
In Code allocation techniques, based on the orthogonal
sets of spreading sequences PAPR reduction schemes are
developed that excludes MC-CDMA signal with large
PAPR. There exists an additional space in the MC-CDMA,
i.e., user code space. Since the code spreading is performed
in frequency domain in MC-CDMA, the PAPR is
determined by the code set allocated to each user. So far,
most of efforts for peak power reduction are made mainly
focusing on searching a good code set achieving low PAPR.
There exists a high correlation between PAPR and the code
combination of simultaneous users for a given code set.
This implies that we may reduce PAPR by optimum
selection of code combination [16]. The PAPR reduction by
user code allocation does not require any signal processing
for the output signal in real-time and not require any side
information or redundancy which is normally required in
conventional PAPR reduction schemes.
In [16]-[19] various methods are used for the optimum
selection of code combinations to achieve minimum PAPR.
The code sets can be selected by various criteria, i.e. the
probability that PAPR exceed a certain threshold, the
maximum PAPR value, and the PAPR with a certain
probability. However, we can only get these criteria from
extensive simulations which require excessive amount of
time.
There are number of coding schemes that offers excellent
performance on PAPR reduction. In spreading code
shuffling technique [19], instead of using additional code
sequence, the same set of code is shuffled between different
users until a low PAPR multiplexed signal is produced at
output. In spreading sequence based technique [17], PAPR
ISSN: 2278 1323

466

property is studied based on two orthogonal sets of
sequences for spreading, i.e., Walsh Hadamard Sequences
and Golay Complementary Sequences. In this a theoretical
value of PAPR is considered, if the number of users are
few, then PAPR of WH sequence becomes larger than the
theoretical value, whereas the PAPR of CP sequence
coincide with the theoretical value. While for large number
of users PAPR of WH sequence becomes smaller than the
theoretical value. So the symbol with low PAPR value is
transmitted. In reduced search for optimum code
technique[18], in order to reduce the number of code
combinations to be searched, the incremental search
algorithm is used which maintains the previously obtained
code set for K-1 user and just find the single optimum code
set for the added kth user.
F. Chip Interleaving
This reduction technique can be used for uplink in M-
modification in MC-CDMA system. In M-modification the
total number of subcarriers N
c
is divided into m groups with
L sub-carriers in each. For one user (uplink) each group of
L subcarriers transmits one symbol spreaded with sequence
of length L. the user transmits M parallel data symbols on
all c N M L = sub-carriers. Walsh sequences are the
orthogonal sequences used for spreading in this case and
using it introduces some kind of redundancy to the system.
For example sequence 1-1 1 will continue with -1. OFDM
based systems (MC-CDMA included) are represented in
time domain by sum of sinusoids. By changing the position
of chip (chips) the peak value of sum of sinusoids
(representing chip sequences) can be reduced. For no chip
interleaving, the first chip is on first sub-carrier, the second
chip on the second sub-carrier, etc. this can be represented
by vector [1 2 3 4 5N
c
]. A permutation of this vector, for
example [3 6 9 13 ], represents the chip interleaving
pattern where the first chip is modulated on the third sub-
carrier, the second chip on the sixth sub-carrier and so on.
The principle of this approach is to find the interleaving
pattern that minimizes PAPR. The chip interleaving pattern
must be same for all users in the system to preserve
orthogonality among them the number of possible
interleaving pattern corresponds with the number of sub-
carriers in the system that makes direct search algorithm
improper for the system with more sub-carriers. So
optimum searching algorithms are applied and thus Genetic
Algorithm (GA) and Ant Colony Algorithm (ACA) are
used to solve this problem [22]-[23].

Fig. 9. M-modification of MC-CDMA and its optimization
G. Discrete Cosine Transform
The discrete cosine transform is a Fourier-like transform
is used to reduce the auto correlation function of the input
sequence to reduce peak-to-average power problem and it
requires no side information to be transmitted to the
receiver.
The 1D discrete cosine transform (1D DCT) A[k] of a
sequence a[n] of length n is defined as
1
0
(2 1)
[ ] [ ] [ ]cos
2
N
n
n k
A k a k a n
N
t
=
+ (
=
(

for k=0,1,,N-1 (12)
The inverse DCT is defined as
1
0
(2 1)
[ ] [ ] [ ]cos
2
N
n
n k
a n a k A k
N
t
=
+ (
=
(

n=0,1,,N-1 (13)
where a[k] is defined as
1
0
[ ]
2
1, 2,..., 1
for k
N
a k
for k N
N
(14)

The equation (11) is expressed in matrix below
A=C
N
a (15)
where A and a are both vector with N1 and CN is a DCT
transform matrix with N N . The row (or column) of the
DCT matrix CN are orthogonal matrix vectors. Then we can
use this property of the DCT matrix and reduce the peak
power of MC-CDMA signals. [24]- [25]

H. Discrete wavelet transform
Wavelet transforms can be derived from Fourier
transforms. The most interesting dissimilarity between
these two kinds of transforms is that the individual wavelet
functions are localized in space, while Fouriers sine and
cosine functions are not. The space localization feature,
ISSN: 2278 1323

467

along with wavelets localization of frequency, makes many
wavelet functions perform better in impulse interference
cancellation scenarios. This is also a part of important
reason why we prefer using wavelets bases to modulate
symbols in MCM systems. The wavelet packet functions
( )( ) n w t n Z+ e are defined by the following recursive
functions:
2 ( ) 2 ( ) (2 ) n n
k Z
w t h k w t k
+ e
=
(16)

2 1( ) 2 ( ) (2 ) n n
k Z
w t g k w t k
+
+
e
=
(17)

where h(k) and ( ) ( 1) ( 1 )
k
g k h L k = stand for a pair of
quadrature mirror filters (QMFs) of length L and Z
+
denotes
the non-negative integers.
The discrete wavelet transform (DWT) is a type of batch
processing, which analyzes a finite time domain signal by
breaking up the initial domain in two parts: the detail and
approximation information. The DWT property is only few
coefficients of DWT dominates the representation. [25]-
[26]

Fig. 10. MC-CDMA system using discrete transform
techniques
I. Discrete Hartley Transform
The DHT is a linear transform. In DHT N real numbers
x
0
,x
1
,,x
N-1
are transformed in to N real numbers
H
0
,H
1
,,H
N-1
. According to the N point DHT [27], N point
DHT can be defined as
1 1
0 0
2 2 2
cos sin
N N
k n n
n n
nk nk nk
H x x cas
N N N
t t t

= =
( | | | | | |
= + =
| | | (
\ . \ . \ .

(18)
where cos sin casu u u = + and k=0,1,,N-1
, 1
2
m
mn
a cas
N
t | |
=
|
\ .
(19)

The precoding matrix A can be constructed as
00 01 0( 1)
10 11 1( 1)
( 1) ( 1) ( 1)( 1)
L
L
L L L L
a a a
a a a
A
a a a

(
(
(
=
(
(

(20)

A is a precoding matrix of size L L , m and l are integers
from 0 to L-1. The DHT is also invertible transform which
allows us to recover the x
n
from H
k
and inverse can be
obtained by simply multiplying DHT of H
k
by 1/N. [25]-
[28]
IV. SPREADING CODES
Various spreading codes exist which can be distinguished
with respect to orthogonality, correlation properties,
implementation complexity and peak-to- average power
ratio (PAPR). The selection of spreading code depends on
the scenario. In synchronous downlink, orthogonal
spreading codes are of advantage, since they reduce the
multiple access interference compared to non-orthogonal
sequences. However, in the uplink, the orthogonality
between the spreading codes gets lost due to different
distortions of the individual codes. Thus, simple PN
sequence can be chosen for spreading in uplink. If the
transmission is asynchronous, gold codes have good cross-
correlation properties.
Here synchronous downlink MC-CDMA system is used.
The spreading codes used in MC-CDMA have their length
N. the individual element in the code is referred to as chips.
Each chip of a code belongs to the set {1, -1}. The property
of the codes that is desired is for the codes of different users
to be orthogonal, i.e.,
1
,
0
[ ] [ ]
N
l m
i
c i c i No
=
= (21)
A. PN Sequences
The property of a PN sequence is that he sequence
appears to be noise-like if the construction is not known at
the receiver. They are typically generated by using shift
registers. Often used PN sequences are maximum-length
shift register sequences, known as m-sequences. A
sequences has a length of
2 1
m
n = (22)
Bits and is generated by a shift register of length m with
linear feedback. The sequence has a period length of n and
each period contains 2
m-1
ones and 2
m-1
-1 zeros; i.e. t is
balanced equation.
B. Walsh Hadamard codes
Walsh-Hadamard code is generated from a hadamard
matrix whose rows form an orthogonal set of codes. The
code sequences in this code set are the individual rows of a
Hadamard matrix. Hadamard matrices are square matrices
whose entries are either +1 or -1 and whose rows and
columns are mutually orthogonal. If N is a non-negative
power of 2, the N N Hadamard matrix, is defined
recursively as follows:
ISSN: 2278 1323

468

2
N N
W
N
N
N
H H
H
H H
(
=
(

(23)
2
1 1
1 1
H
+ + (
=
(
+

(24)
Walsh-Hadamard codes are important because they form
the basis for orthogonal codes with different spreading
factors. This property becomes useful when we want
signals with different spreading factors to share the same
frequency channel.
C. Golay Complementary (CP) codes
In a similar manner, an orthogonal set of golay
complementary codes can be recursively obtained by
2
N
N C
N
N
N
H H
H
H H
(
= (
(

(25)
where N H is composed of H
N
of which right half columns
are reversed, e.g., if [ ] N N N H A B = where A
N
and B
N

are
2
N
N matrices, then N H [ ] AN BN .
It can readily be shown by induction that
C CT
N
N N
H H I = (26)
where I
N
denotes the NN identity matrix, hence the matrix
given above is orthogonal.
D. Orthogonal Binary User Codes
Orthogonal binary user codes are constructed through
performing search operations in the binary sample space
that consist of zero mean and linear phase codes. In contrast
to WH codes, the OBU codes do not necessitate the
condition that there can be only one code in the set for a
given number of zero crossings. Again, for WH codes,
decimal value of all n-bit codes are multiples of either
2
(n/2)
+1 or 2
(n/2)
-1. OBU do not impose this restriction.
Rather, efforts were made to reduce the number of codes
that fulfill this criterion in order to avoid the codes that are
common to WH codes. For a given code length, more than
one OBU code sets can be formulated. In this context one
may remember that for a given code length there exist only
one set of WH codes. Since shorter code length implies
smaller binary sample space, there exist a considerable
number of short OBU codes that are common with same
length WH codes. But this number reduces as code length
becomes higher.
E. Gold codes
PN sequences with better cross-correlation properties
than m-sequences are part of so-called Gold sequences. A
set of n Gold sequences is derived from preferred pair of m-
sequences with n cyclically shifted versions of the second
preferred m-sequences. By including the two preferred m-
sequences, a family of n+2 Gold codes is obtained. Gold
codes have a three-valued cross-correlation function with
values{ } 1, ( ), ( ) 2 t m t m , where
( 1)/ 2
( 1)/ 2
2 1
( )
2 1
m
m
for modd
t m
for meven
+
+
+
=

+
(27)
V. CRITERIA FOR THE SELECTION OF PAPR REDUCTION
TECHNIQUE

1. High capability of reducing PAPR:- It is the first and
the most important factor to be considered during the
selection of PAPR reduction technique considering the
few harmful side effects such as in-band and out-band
distortion.
2. Loss in data rate:- Some techniques i.e., signal
scrambling techniques with explicit side information
requires the data rate to be reduced. In the techniques
like SLM, PTS, interleaving the data rate is reduced
due to side information transmitted to the receiver that
gives the information about what has been done at the
transmitter side and this side information can be
received in error unless some kind of protection in the
form of channel coding is employed. Thus when
channel coding is used the loss in data rate due to side
information is further increased.
3. Increase in BER at the receiver: - The aim of PAPR
reduction is to increase the system performance as well
as the BER than that of the original MC-CDMA
system. Therefore all methods which have more
increase in BER at the receiver should be paid more
attention. Moreover, if the side information is received
in error at the receiver that may result in error in entire
data frame and thus this may also increase the BER at
the receiver.
4. Computational complexity:- Computational complexity
is an important factor to be considered during the
selection of PAPR reduction technique. The technique
such as PTS finds the best solution of PAPR reduction
of the signal by using much iteration. Also, the PAPR
reduction capability of interleaving technique is better
for large number of interleavers. Thus, more complex
techniques have better PAPR reduction capability in
most of the cases.
5. Without additional power increase: - The design of the
system should take into consideration the efficiency of
the power. If operation of the technique that reduces
the PAPR need more power, it degrades the BER
performance when the transmitted signals are
normalized back original power signal.
6. Other consideration: - More attention should be paid to
effects of non-linear devices such as digital-to-analog
ISSN: 2278 1323

469

converter, HPAs and mixers as PAPR reduction mainly
avoids non-linear distortion due to these memory-less
devices introducing into the communication channels.
Also the cost of these non-linear devices should also be
considered in designing the PAPR reduction scheme.

VI. COMPARATIVE STUDY OF PAPR REDUCTION
TECHNIQUES
TABLE I: Advantages and Disadvantages of Different PAPR
Reduction Techniques
S.
No.
Technique Advantages Disadvantages
1. Clipping - No loss in
data rate
- No side
information
required
- Distortion
(In-band and
Out-band)
- Increase
BER
2. Companding - Better power
spectral
density
- Low
implementati
on
complexity
- No constraint
in modulation
format and
sub-carrier
size
- Increases
average
received
power
- Out-band
distortion
3. Selected
Mapping
(SLM)
- No distortion
is introduced
- Does not
require any
complex
optimization
techniques
- More
effective than
PTS for same
amount of
side
information
- Side
information
needed
- Complexity
increases
with increase
in number of
phase factors
4. Partial
Transmit
Sequence
(PTS)
- Less complex
than SLM
- PAPR
reduction in
PTS performs
better than
SLM
- Side
information
is larger than
SLM
5. Chip
Interleaving
- Less complex
- No
performance
degradation
- No need of
side
information
- Multiple
iterations
needed
6. Spreading
Code based
Techniques
- Excellent
performance
in PAPR
degradation

- Cost of
complexity
is high
- Data rate
loss
7. Discrete
Cosine
Transform
(DCT)
- Does not
influence
BER
-
Computation
al
complexity
increases
8. Discrete
Wavelet
Transform
(DWT)
- High spectral
efficiency
- No loss of
data rate
- No distortion
- Complexity
increases
9. Discrete
Hartley
Transform
(DHT)
- Shows better
PAPR gain
when
compared
with SLM
- Complex

VII. CONCLUSION
MC-CDMA is an important multiple access candidate for
4G wireless communication system. The transmitted signal
of an MC-CDMA system exhibits a very high peak-to-
average power ratio (PAPR) when large number of sub-
carriers are used. Therefore amplifier should be operated in
large linear ranges to avoid non-linear distortion. This leads
to very inefficient amplification and very expensive
transmission. In this paper we have discussed various
PAPR reduction techniques for multicarrier transmission,
all which have the potential to reduce PAPR at the cost of
loss in data rate, increase in BER, transmit signal power and
computational complexity. The PAPR reduction technique
should be chosen according to the appropriate system
requirements. The comparative study of the PAPR
reduction techniques is also discussed in the paper with
their advantages and disadvantages.
ACKNOWLEDGMENT
We are grateful to the University College of Engineering,
Punjabi University, Patiala for providing necessary support
and infrastructure.
REFERENCES
[1] R.van Nee and R. Prasad, OFDM for Wireless Multimedia
Communication, Bartech house, 2000.
[2] S. Hara and R. Prasad, Overview of Multicarrier CDMA, IEEE
Communication Mag., vol. 35, no.12, pp.126-133,dec.1997.
[3] N. Yee, J. P. Linnartz and G. Fettweis, Multicarrier CDMA in
Indoor Wireless Radio Networks, Proc. IEEE PIMRC93, Pp. 109-
113, Sep. 1993
[4] X. Li And L. J. Cimini Jr., Effects of Clipping and Filtering on the
Performance of OFDM, IEEE Commn. Lett., vol. 2, no.5, Pp. 131-
133, May 1988.
[5] Masaki Ojima and Takeshi Hattori, PAPR Reduction Method using
Clipping and Peak Windowing in CI/OFDM System, Sophia
University Journal, 2007.
[6] Toshiyuki, Shigeru, Masaharu, And Junichiro, PAPR Reduction
Improvement in Iterative Clipping and Filtering with a Graded Band
Limiting Filter for OFDM Transmission, IEICE Trans
Fundamentals, Vol.E90-A, No.7, 2007
[7] K.W. Bauml, R.F.H. Fischer and J.B. Huber, Reducing the peak to-
average power ratio of multicarrier modulation by selected
mapping, Electron. Lett. vol. 32, no. 22, pp. 2056-2057, Oct. 1996.
ISSN: 2278 1323

470

[8] Yan Biao; Yang Juan; and Li Ming, Reduction effect of SLM on
PAPR in MC-CDMA systems, Journal of Systems in Engineering
and Electronics, 2005, pp 761-766
[9] Naoto Ohkubo and Tomoaki Ohtsuki, Design Criteria for Phase
Sequences in Selected Mapping, IEEE Vehicular Technology
Conference, 2003, pp 373-377
[10] S. Ruangsuthinarupap; K.M. Ahmed; W.A.C. Fernando and M.E.R.
Khan, PAPR Reduction by Combined Selected Mapping and
Selected Spreading Code in MC-CDMA systems, , Ninth
International Symposium on Computers and Communications, 2004.
Proceedings. ISCC 2004. 2004, pp-725-728
[11] Jizeng Wang; Jingyu Luo; Yanlong Zhang; A New Phase
Sequence for SLM in MC-CDMA System, WiCom Sept. 2007.
[12] S.H. Muller and J.B. Huber, OFDM with reduced peak to average
power ratio by optimum combination of partial transmit sequence,
Electronics Letters, Vol. 33, no. 5, pp. 368-369,1997.
[13] A.S. Madhukumar, A. B. Premkumar, Peng Xiaoming and Francois
Chin, A simplified approach for peak-to-average power reduction
for multicarrier code division multiple access system, IEEE Inter.
Synp.on PIMRC07, 2007
[14] N. Ruangsurat and R. M. A. P. Rajatheva, An Investigation of Peak
to Average Power Ratio in MC-CDMA Combined with Partial
Transmit Sequence, IEEE VTC 2001-Spring. Vol.1, pp. 761-765,
Rhodes, May 2001.
[15] Lilin Dan; Oingsong Wen; Yue Xiao and Shaoqian Li, A Novel
User Grouping Scheme for PAPR Reduction in MC-CDMA
System, Vehicular Technology Conference, pp 994-998, 2007
[16] Popovic, B.M., Spreading sequences for multi-carrier CDMA
systems, ZEE Colloquium on CDMA Techniques and Applications
for Third Generation Mobile Systems, pp. 811 -816 1997
[17] H. Ochiai and H. Imai, OFDM-CDMA with peak power reduction
based on the spreading sequences, IEEE ICC98, pp. 1299 1303,
Atlanta, June 1998.
[18] Kunseok Kang; Kweonhue Choi; Sooyoung Kim Shin, Reduced
search for optimum code sets to reduce PAPR in MC-CDMA
system, Wireless Personal Multimedia Communications, 2002. vol
1, pp.135 139
[19] E Alsusa & L Yang, MC-CDMA Specific PAPR Reduction
Technique Utilizing Spreading Code Redistribution, IEEE VTC/Fall
2006, Montreal, Canada, Sept. 2006, pp. 1 - 5.
[20] Wang Yi Gu linfeng An Investigation of Peak-to-Average
Power Reduction in MIMO-OFDM Systems, Blekinge
Institute of Technology, October 2009.
[21] Md. Abdullah Al Baki and Mohammad Zavid Parvez, Peak to
Average Power Ratio (PAPR) Reduction in OFDM Based Radio
Systems Blekinge Institute of Technology, May 2010.
[22] Zbynek Fedra and Vladimir Sebesta, The new PAPR reduction
approach in MC-CDMA, Signal Processing Symposium, 2006, pp
270-273, 2006
[23] Zbynek Fedra, Roman Marsalek and Vladimir Sebesta, Chip
Interleaving and its Optimization for PAPR Reduction in MC-
CDMA, radio engineering, vol. 16 , no.4, December 2007
[24] Zhongpeng Wang, Combined DCT and companding for PAPR
reduction in OFDM signals, Journal of Signal and Information
Processing 100-104
[25] B. Sarala, D.S. Venkateswarlu, MC-CDMA PAPR reduction
techniques using Discrete Transforms and Companding, IJDPS
nov.2011, vol.2, no.6, pp 253-270

[26] Haixia Zhang, Dongfeng Yuan and Cheng-Xiang Wang, A Study
on the PAPRs in multicarrier modulation systems with different
orthogonal bases, Wiley InterScience, Wireless Communication and
Mobile Computing, march 2007, pp 311-318
[27] R.N. Bracewell, Discrete Hatley Transform, Journal of the
Optivcal society of America, 73(12), pp 1832-1835
[28] Imran Baig and Varum Jeoti, A new discrete Hartley transform
precoding based interleaved-OFDMA uplink system with reduced
PAPR from 4G cellular networks, Journal of Engineering Science
and Technology, vol.6, no.6, pp 685-694, 2011.
[29] X.B. Wang, T.T. Tjhung and C.S. Ng, Reduction of peak-to-average
power ratio of OFDM system using A companding technique, IEEE
Trans. Broadcasting, vol. 45, no.3, pp 303-307, Sept 1999
[30] T. Jiang and G. X. Zhu, Nonlinear companding transform to reduce
peak-to-average power ratio of OFDM signal, IEEE Trans.
Broadcasting, vol. 50, no.3, pp 342-346, Sept 2004
[31] T. Jiang Y.Yang and Y. Song, Exponential companding transform
for PAPR reduction in OFDM system, IEEE Trans. Broadcasting,
vol. 51, no.2, pp 244-248, Jun. 2005
[32] R. Manjith, S.C. Ramesh and M. Mohamed Ismail Majeed, PAPR
reduction in OFDM and MC-CDMA system using nonlinear
companding techniques, IEEE region 8 SIBIRCON-2010, pp- 274-
278, July 2010

Gagandeep Kaur: Designation: Student,
Academic Qualifications: Currently pursuing M.Tech (ECE) from
UCOE, Punjabi University, Patiala. Done B.Tech (ECE) from RIMT-
IET, Mandi Gobindgarh. Areas of interest: Wireless Communication
and MATLAB. Published Work: International Conferences: 1

Rajbir Kaur: Designation: Assistant
Professor, Electronics and Communication Engineering. Academic
Qualifications: M.Tech., pursuing PhD. Area of Specialization: Analog
and Digital communication systems. Published Work: National
Conferences: 6, International Conferences: 1, International Journal: 1,
National Conference/Seminar Presentation: 1. M.Tech/M.Phil Students
Guided / Underguidance : 04

461 470

Uploaded by

Copyright:

Available Formats

461 470

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

461 470

Uploaded by

Copyright:

Available Formats

ISSN: 2278 1323

International Journal of Advanced Research in Computer Engineering & Technology

to obtain the diversity

are multiplied by U-1

. After the IFFT

and one original

You might also like