You are on page 1of 8

International Journal of Emerging Technology and Advanced Engineering

Website: www.ijetae.com (ISSN 2250-2459, Volume 2, Issue 2, February 2012)
43

Peak-To-Average Power Ratio Reduction by CB-ACE and
Adaptive Ace Algorithms
Madhuri P
1
, Dr Malleswari B L
2
1
Assistant Professor, GNITS, Hyderabad, A.P.
2
Professor, GNITS, Hyderabad, A.P.
1
madhuripocha@gmail.com

2
blmalleswari@gmail.com

Abstract - Orthogonal Frequency Division Multiplexing
(OFDM) is a method of Digital Modulation in which a signal is
split into several narrowband channels at different
frequencies. The OFDM technology was first conceived in the
1960s and 1970s during the research into minimizing
interference among the channels near each other in the
frequency [5]. The main idea behind OFDM is that since low-
rate modulations are less sensitive to multipath, the better
way is to send a number of low rate streams in parallel than
sending one high rate waveform [8].The OFDM signal has a
noise like amplitude with a very large dynamic range;
therefore it requires RF power amplifiers with a high Peak-to-
Average Power Ratio (PAPR) or Peak-to-Average Ratio
(PAR) [1].
The high PAPR increases the complexity of Analog-to-
Digital (A/D) and Digital-to-Analog (D/A) converters and also
lowers the efficiency of power amplifiers [1]. The Clipping-
Based Active Constellation Extension (CB-ACE) Algorithm
and Adaptive Active Constellation Extension (Adaptive ACE)
Algorithm can be efficiently used to reduce the high PAPR in
OFDM systems.

Keywords - Orthogonal Frequency Division Multiplexing
(OFDM), Peak-to-Average Power Ratio (PAPR), Peak-to-
Average Ratio (PAR), Clipping-Based Active Constellation
Extension (CB-ACE), Adaptive Active Constellation
Extension (Adaptive ACE)
I. INTRODUCTION
Orthogonal Frequency Division Multiplexing (OFDM)
is a Multi-Carrier Modulation (MCM) scheme, in which
many parallel data streams are transmitted at the same time
over a channel, with each transmitting only a small part of
the total data rate. With OFDM, a high-speed digital
message is divided into a large number of separate carrier
waves [7].
OFDM depends on Orthogonality principle.
Orthogonality is allowing the sub carriers, which are
orthogonal to each other. So that, the cross talk between co-
channels is eliminated and inter-carrier guard bands are not
required.
The OFDM signal has a noise like amplitude with a very
large dynamic range; therefore it requires RF power
amplifiers with a high Peak-to-Average Power Ratio
(PAPR). The high PAPR increases the complexity of
Analog-to-Digital (A/D) and Digital-to-Analog (D/A)
converters and also lowers the efficiency of power
amplifiers [1].
A. Block Diagram of OFDM Transmitter
The block diagram of Orthogonal Frequency Division
Multiplexing (OFDM) transmitter is shown in the figure
1.1. S[n], a serial stream of binary digits is the input to the
Orthogonal Frequency Division Multiplexing (OFDM)
transmitter. By inverse multiplexing, the serial stream of
binary digits is converted into N parallel streams, as any
programmable IC will take the input parallely. The N
parallel streams are converted into the state space
components by using modulation techniques like
Quadrature Amplitude Modulation (QAM) and Phase Shift
Keying (PSK).
Quadrature Amplitude Modulation (QAM) is a
combination of simple Amplitude Modulation (AM) and
simple Phase Modulation, which helps to transmit more
data over the same bandwidth as that of simple Amplitude
Modulation alone or Phase Modulation alone. Hence, QAM
increases the efficiency of transmission of the data for radio
communication systems.
Phase Shift Keying (PSK) is a digital modulation
scheme that conveys the data by changing or modulating
the phase of the reference signal or the carrier wave. PSK
uses a finite number of phases each assigned with a unique
pattern of binary digits. The various forms of the Phase
Shift Keying (PSK) are Binary Phase Shift Keying (BPSK),
Quadrature Phase Shift Keying (QPSK), 8-Phase Shift
Keying (8-PSK) and 16-Phase Shift Keying (16-PSK).
An Inverse Fast Fourier Transform (IFFT) is computed
on each set of symbols, giving a set of complex time-
domain samples. The time-domain samples are
then Quadrature mixed to pass band in the standard way.

International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, Volume 2, Issue 2, February 2012)
44

The real and imaginary components are first converted
to the analog domain by using Digital-to-Analog
Converters (DACs).
The analog signals are then used to modulate cosine and
sine waves at the carrier frequency, f
c
respectively. These
signals are then summed up to give the transmission
signal, S(t) [16].

Fig 1.1 : Block Diagram of OFDM Transmitter

B. Block Diagram of OFDM Receiver
The block diagram of Orthogonal Frequency Division
Multiplexing (OFDM) receiver is shown in the figure 1.2.
The received signal, r(t) is the input to the Orthogonal
Frequency Division Multiplexing (OFDM) receiver. The
received signal is Quadrature mixed down to baseband
signals by using cosine and sine waves at the carrier
frequency, f
c
. The frequency of the baseband signals is
equal to twice that of the carrier frequency i.e., 2f
c
,
so Low Pass Filters (LPFs) are used to attenuate the
frequency which is greater than the carrier frequency, f
c
.
The baseband signals are then sampled and digitized by
using Analog-to-Digital Converters (ADCs). A forward
Fast Fourier Transform (FFT) is used to convert back into
the frequency domain. The frequency domain signals are
then converted to N parallel streams, each of which is
converted to a binary stream using an appropriate
symbol detector. The binary streams are then recombined
into a serial stream, denoted by
ˆ
S[n]
which is an estimate
of the original binary stream at the transmitter [16].


1.2 : Block Diagram of OFDM Receiver

International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, Volume 2, Issue 2, February 2012)
45

II. PEAK-TO-AVERAGE POWER RATIO
The Peak-to-Average Power Ratio (PAPR) of an
Orthogonal Frequency Division Multiplexing (OFDM)
transmit signal is given as shown in equation 2.1 [4], [6].
2
T
2
0
max( x(t ) )
1
x(t ) dt
T
10
PAPR 10log dB
, ,
, ,
}
=
Eq. 2.1
Where,
PAPR – Peak-to-Average Power Ratio
x(t) – Baseband Representation of OFDM Symbol
T – Duration of the OFDM Symbol
The baseband representation of the Orthogonal
Frequency Division Multiplexing (OFDM) symbol is given
as in equation 2.2 [1].
j2 nt N 1
N
n
n 0
1
x(t) X e , t T
N
t ÷
=
= 0 s s
¿
Eq. 2.2
Where,
N – Number of Subcarriers
X
n
– Data Sequence
T – Duration of the OFDM Symbol
OFDM is currently being used in several new radio
broadcast systems including the proposal for High
Definition Digital Television (HDTV). The other major
applications of the OFDM systems are –
- Worldwide Interoperability for Microwave Access
(WiMAX)
- Digital Audio Broadcasting (DAB)
- Digital Video Broadcasting (DVB)
- Wireless Networking
A. Applications of CB-ACE And Adaptive ACE
The Clipping-Based Active Constellation Extension
Algorithm also called as the Repeated Clipping and
Filtering (RCF) process is used to reduce the Peak-to-
Average Power Ratio (PAPR) in OFDM Systems.
The Adaptive Active Constellation Extension Algorithm
is not only used to reduce the Peak-to-Average Power Ratio
(PAPR) but also to overcome the drawbacks of CB-ACE
Algorithm or RCF Process.
III. CLIPPING-BASED ACTIVE CONSTELLATION EXTENSION
(CB-ACE) ALGORITHM
The basic idea of the Clipping-Based Active
Constellation Extension (CB-ACE) Algorithm is to
generate the anti-peak signal for reducing the Peak-to-
Average Power Ratio (PAPR) by projecting the flipping in-
band noise into feasible extension area while removing the
out-of-band distortion with filtering [2].
The basic principle of Clipping-Based Active
Constellation Extension (CB-ACE) algorithm involves
switching between the time domain and the frequency
domain. Filtering and applying the ACE constraint in the
frequency domain, after clipping in the time domain, both
require iterative processing to suppress the subsequent re-
growth of the peak power [9].
The CB-ACE algorithm is first used to clip the peak
amplitude of the original Orthogonal Frequency Division
Multiplexing (OFDM) signal. The clipping sample
obtained after clipping the peak signals, denoted by
(i )
n
c
,
is given by the equation 3.1 [2].
n
j (i) (i) (i)
n n n
(i)
n
c ( x e , x
x
u
= | |÷A) | |> A
0, | |s A
Eq. 3.1
Where,
(i )
n
c
– Clipping Sample of the i
th
iteration
(i)
n
x

– Oversampled OFDM signal
A – Predetermined Clipping Level
θ
n
– arg(-
(i)
n
x )
The equation 3.1 says that the clipping sample is reduced
to a value equal to zero when the peak amplitude of the
original OFDM signal is less than or equal to the
predetermined clipping level, A. If the peak amplitude of
the original OFDM signal is greater than the predetermined
clipping level, then the clipping sample is given by
n
jθ (i)
n
( x A)e ÷ , where the predetermined clipping
level is subtracted from the oversampled OFDM signal an
is then multiplied by an exponential value [2].
The predetermined clipping level, denoted by A, is
related to the target clipping ratio, γ and is given by the
equation 3.2 [2].
2
2
n
A
γ=
E{ x }
Eq. 3.2
Where,
γ – Target Clipping Ratio
A – Predetermined Clipping Level
x
n
– Oversampled OFDM signal
The clipping of the peak signal results to distortion of
the original OFDM signal.
The distortion of the original signal can be assumed as
the noise, which results to an unreliable communication
between the transmitter and the receiver.


International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, Volume 2, Issue 2, February 2012)
46

The distortion caused by clipping the original OFDM
signal is categorized into two types, namely [2] –
- In-Band Distortion.
- Out-of-Band Distortion.
The in-band distortion results in the system performance
degradation and cannot be reduced, while, the out-of-band
distortion can be minimized by filtering the clipped signals.
The signal obtained after filtering the clipped signal is
given by the equation 3.3 [2].
(i+1) (i) (i)
x =x +μc Eq. 3.3
Where,
(i)
c
– Anti-Peak Signal at the i
th
iteration
µ – Positive Real Number (µ varies from 0.1 to 1)
The anti-peak signal at the i
th
iteration generated
for the PAPR reduction, denoted by
(i)
c , is given by the
equation 3.4 [2].
(i) (i) (i)
c =T c Eq. 3.4
Where,
(i)
c – Anti-Peak Signal at the i
th
iteration
T
(i)
– Transfer Matrix at the i
th
iteration
C
(i)
– Peak Signal above the Pre-Determined Level
The transfer matrix at the i
th
iteration, denoted by
T
(i)
, used for generating the anti-peak signal is given by the
equation 3.5 [2].
(i) *(i) (i)
ˆ ˆ
T =Q Q
Eq. 3.5
Where, T
(i)
– Transfer Matrix at the i
th
iteration

*(i)
ˆ
Q – Conjugate of Constellation Order

(i)
ˆ
Q – Constellation Order
Though, the process of filtering completely eliminates
the distortions caused by the clipping process, it introduces
peak re-growth at some of the peak signals of the OFDM
signal. The peak re-growth can be reduced by repeating the
filtering process, which may again introduce some
distortions. Therefore, the clipping and filtering processes
are to be repeated until the peak signals are completely
reduced. Hence, the Clipping-Based Active Constellation
Extension (CB-ACE) Algorithm is also named as the
Repeated Clipping and Filtering (RCF) process.
IV. ADAPTIVE ACTIVE CONSTELLATION EXTENSION
(ADAPTIVE-ACE) ALGORITHM
The main objective of the Adaptive Active Constellation
Extension (Adaptive ACE) Algorithm for reducing the
Peak-to-Average Power Ratio (PAPR) is to control both the
clipping level and the convergence factor at each step and
thereby minimize the peak power signal whichever is
greater than the initial target clipping level [2].
The Adaptive Active Constellation Extension (Adaptive
ACE) Algorithm can be initialized by selecting the
parameters namely the target clipping level, denoted by A
and the number of iterations, denoted by i. In the first step,
the iteration is taken as two i.e., i = 2 and the initial target
clipping level is to be taken as A [2].
The predetermined clipping level, denoted by A, is
related to the target clipping ratio, γ and given is by the
equation 4.1 [2].
2
2
n
A
γ =
E{ x }
Eq. 4.1
Where,
γ – Target Clipping Ratio
A – Predetermined Clipping Level
x
n
– Oversampled OFDM signal
The clipping of the peak signal results to
distortion of the original OFDM signal. The distortion of
the original signal can be assumed as the noise, which
results to an unreliable communication between the
transmitter and the receiver. The distortion caused by
clipping the original OFDM signal is categorized into two
types, namely [2] –
- In-Band Distortion.
- Out-of-Band Distortion.
The in-band distortion results in the system
performance degradation and cannot be reduced, while, the
out-of-band distortion can be minimized by filtering the
clipped signals. The signal obtained after filtering the
clipped signal is given by the equation 4.2 [2].
(i+1) (i) (i)
x = x + μc
Eq. 4.2
Where,
(i)
c
– Anti-Peak Signal at the i
th
iteration
µ – Convergence Factor
The Convergence Factor (CF), denoted by µ can
be estimated by using the equation 4.3.
(i) (i)
(i) (i)
[ c ,c ]
μ
c ,c
9
=
Eq. 4.3
Where,
µ – Convergence Factor
9 – Real Part
C
(i)
– Peak Signal above Pre- Determined
(i)
c
– Anti-Peak Signal at the i
th
iteration

International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, Volume 2, Issue 2, February 2012)
47

, – Complex Inner Part
The anti-peak signal at the i
th
iteration generated
for the PAPR reduction, denoted by
(i)
c , is given by the
equation 4.4 [2].
(i) (i) (i)
c = T c Eq. 4.4
Where,
(i)
c – Anti-Peak Signal at the i
th
iteration
T
(i)
– Transfer Matrix at the i
th
iteration
C
(i)
– Peak Signal above Pre-Determined
The transfer matrix at the i
th
iteration, denoted by
T
(i)
, used for generating the anti-peak signal is given by the
equation 4.5 [2].
(i) *(i) (i)
ˆ ˆ
T = Q Q Eq. 4.5
Where,
T
(i)
– Transfer Matrix at the i
th
iteration
*(i)
ˆ
Q – Conjugate of Constellation Order
(i)
ˆ
Q – Constellation Order
The original Orthogonal Frequency Division
Multiplexing (OFDM) signal, denoted by x
n
, is to be
clipped in order to reduce the peak signals. The clipping
signal is given by the equation 4.6.
n
j (i) (i) (i)
n n n
(i)
n
c = ( x e , x
x
u
| |÷A) | |> A
0, | |s A
Eq. 4.6
Where,
(i )
n
c
– Clipping Sample
A – Predetermined Clipping Level
θ
n
– arg(-
(i)
n
x )
The clipping level denoted by A, for the next
iteration is given by the equation 4.7 [2].
(i+1) (i)
A
A = A μ + V
Eq. 4.7
Where,
A
(i+1)
– Next Iteration Level
A
(i)
– Present Iteration Level
µ – Convergence Factor
A
V – Gradient with respect to A
The gradient with respect to the target clipping
ratio, denoted by
A
V , is given by the equation 4.8 [2].
(i) (i)
1 3
(i+1)
n
n I I
A
p
c
=
N
e
V
¿
Eq. 4.8
Where,
A
V – Gradient with respect to A
N
p
– No. of peak samples larger than A

The Peak-to-Average Power Ratio (PAPR) is to be
calculated to the signal obtained by the equation 4.2, which
reduces the PAPR than the PAPR calculated for the
original OFDM signal or PAPR obtained of the OFDM
signal obtained by using the Clipping-Based Active
Constellation Extension (CB-ACE) Algorithm.
V. RESULT AND DISCUSSION
The Peak-to-Average Power Ratio (PAPR) of the
original OFDM signal can be calculated by using equations
2.1 and 2.2. The simulated result obtained by using
MATLAB version 7.2.0 (R2006a) can be seen in Screen
Shot 5.1.
From the Screen Shot 5.1, the Peak-to-Average Power
Ratio (PAPR) of the original Orthogonal Frequency
Division Multiplexing signal is equal to 11.8 dB with a
CCDF of 10
-2
.
The Peak-to-Average Power Ratio (PAPR) of the
original Orthogonal Frequency Division Multiplexing
(OFDM) signal is very high, which is evident from the
Screen Shot 5.1. The high PAPR results to the increase in
the complexity of the Analog-to-Digital Convertors
(ADCs) and Digital-to-Analog Convertors (DACs), also
reduces the efficiency of the power amplifiers.
The high Peak-to-Average Power Ratio (PAPR) of the
original Orthogonal Frequency Division Multiplexing
(OFDM) signals can be reduced by using various reduction
techniques like Clipping-Based Active Constellation
Extension (CB-ACE) Algorithm, Exponential Companding
Transform, Adaptive Active Constellation Extension
(Adaptive ACE) Algorithm and New Companding
Transform.

International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, Volume 2, Issue 2, February 2012)
48


Screen Shot 5.1 – PAPR Vs CCDF of Original OFDM Signal
The Peak-to-Average Power Ratio (PAPR) of the
clipped and filtered OFDM signal obtained by Clipping-
Based Active Constellation Extension Algorithm can be
calculated by using equations 3.1 to 3.5. The simulated
result obtained by using MATLAB version 7.2.0 (R2006a)
can be seen in Screen Shot 5.2.
From the Screen Shot 5.2, the Peak-to-Average Power
Ratio (PAPR) of the Orthogonal Frequency Division
Multiplexing (OFDM) signal obtained by using the
Clipping-Based Active Constellation Extension (CB-ACE)
algorithm is equal to 10 dB, 8.5 dB and 8.0 dB for the
target clipping ratios of 0 dB, 2 dB and 4 dB respectively
with a Complimentary Cumulative Distribution Function
(CCDF) of 10
-2
or 0.01.The other problems faced by the
Clipping-Based Active Constellation Extension (CB-ACE)
algorithm are Out-of-Band Interference (OBI) and peak re-
growth.Here, the Out-of-Band Interference (OBI) is a form
of noise or an unwanted signal, which is caused when the
original Orthogonal Frequency Division Multiplexing
(OFDM) signal is clipped for reducing the peak signals
which are outside to the predetermined area and the peak
re-growth is obtained after filtering the clipped signal.
The peak re-growth results to, increase in the
computational time and computational complexity.
The Peak-to-Average Power Ratio (PAPR) of the
clipped and filtered OFDM signal obtained by Adaptive
Active Constellation Extension Algorithm can be
calculated by using equations 4.1 to 4.8. The simulated
result obtained by using MATLAB version 7.2.0 (R2006a)
can be seen in Screen Shot 5.3.
From the Screen Shot 5.3, the Peak-to-Average Power
Ratio (PAPR) or Peak-to-Average Ratio (PAR) or Crest
Factor of the Orthogonal Frequency Division Multiplexing
(OFDM) signal obtained by using the Adaptive Active
Constellation Extension (Adaptive-ACE) algorithm is equal
to 6.8 dB for the target clipping ratios of 0 dB, 2 dB and 4
dB with a Complimentary Cumulative Distribution
Function (CCDF) of 10
-2
or 0.01.
The table 4.1 gives a comparison of Peak-to-Average
Power Ratio (PAPR) of the original Orthogonal Frequency
Division Multiplexing (OFDM) signal and clipped &
filtered OFDM signal obtained by using CB-ACE
Algorithm and Adaptive ACE Algorithm.


International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, Volume 2, Issue 2, February 2012)
49


Screen Shot 5.2 – PAPR Vs CCDF by using CB-ACE Algorithm
(For Different Target Clipping Ratios)


Screen Shot 5.3 – PAPR Vs CCDF by using Adaptive-ACE Algorithm
(For Different Target Clipping Ratios)


International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, Volume 2, Issue 2, February 2012)
50

Table 5.1 – Comparison of PAPR for Different Algorithms

Original
OFDM Signal
CB-ACE
Algorithm
Adaptive ACE
Algorithm

PAPR in dB

11.8
10.0 (γ = 0 dB)
8.5 (γ = 2 dB)
8.0 (γ = 4 dB)
6.8 (γ = 0 dB)
6.8 (γ = 2 dB)
6.8 (γ = 4 dB)

VI. CONCLUSIONS
The Peak-to-Average Power Ratio (PAPR) of the
Orthogonal Frequency Division Multiplexing (OFDM)
system is equal to 11.8 dB (approximately 12 dB) without
using any algorithm. The high PAPR results to the increase
in the complexity of Analog-to-Digital Convertor (ADC)
and Digital-to-Analog Convertor (DAC), also lowers the
efficiency of power amplifiers.
The Peak-to-Average Power Ratios (PAPR) for the
target clipping ratios of 4 dB, 2 dB and 0 dB are 8.0 dB, 8.5
dB and 10.0 dB respectively by using Clipping Based
Active Constellation Extension (CB-ACE) Algorithm.
Though, the PAPR is reduced by using CB-ACE algorithm,
the minimum PAPR cannot be achieved for low target
clipping ratios.
The Peak-to-Average Power Ratios (PAPR) for the
target clipping ratios of 4 dB, 2 dB and 0 dB is 6.8 dB by
using Adaptive Active Constellation Extension (Adaptive
ACE) Algorithm. The drawbacks of CB-ACE Algorithm
are minimized by Adaptive ACE Algorithm.
References
[1] Abaurakhia S A, Badran E F and Darwish A E, “Linear Companding
Transform for the Reduction of Peak-to-Average Power Ratio of
OFDM Signals”, IEEE Transactions on Broadcasting, Vol. 55, No.
1, pp. 155 – 160, March 2009.
[2] Bae K, Andrews J G and Powers E J, “Adaptive Active
Constellation Extension Algorithm for Peak-to-Average Ratio
Reduction in OFDM”, IEEE Communications Letters, Vol. 14, No.
1, pp. 39 – 41, January 2010.
[3] Hou J, Ge J, Zhai D and Li J, “Peak-to-Average Power Ratio
Reduction of OFDM Signals with Nonlinear Companding Scheme”,
IEEE Transactions on Broadcasting, Vol. 56, No. 2, pp. 258 – 262,
June 2010.
[4] http://en.wikipedia.org/wiki/Crest_factor.
[5] http://searchnetworking.techtarget.com/definition/orthogonal-
frequency-divisi on-multiplexing.
[6] http://www.dliengineering.com/downloads/crest%20factor.pdf.
[7] http://www.linktionary.com/o/ofdm.html.
[8] http://www.mobileisgood.com/ofdm.php.
[9] Ouderaa E V, Schoukens J and Renneboog J, “Peak Factor
Minimization Using a Time-Frequency Domain Swapping
Algorithm”, IEEE Transactions on Instrumentation and
Measurement, Vol. 37, No. 1, pp. 145 – 147, March 1988.