International Journal of Electronics and Electrical Engineering Vol. 3, No.

1, February, 2015

Reduction of PAPR of OFDM Symbols Using

Modified Narahashi and Nojima Phasing
Sequences
Vallavaraj Athinarayanan
Caledonian College of Engineering, Muscat, Oman
Email: [email protected]

Brian G Stewart
Glasgow Caledonian University, Glasgow, United Kingdom
Email: [email protected]

AbstractThis paper discusses the limitation of the phasing cannot be used with applications involving digital
technique proposed by Narahashi and Nojima for reducing modulation schemes such as phase shift keying (PSK) or
the PAPR of OFDM symbols. It is shown that although the quadrature amplitude modulation (QAM). For example,
phasing sequence proposed by Narahashi and Nojima works
with quadrature phase shift keying (QPSK) modulation of
well for constant amplitude and constant phase angle signals,
the technique does not work for inputs with varying phase random input signals are mapped to produce an output
angles, for example, QPSK. To improve this situation a that could have one of four different phase angles. The
modified phasing sequence is proposed and numerical derivation of the phasing scheme in [6] does not consider
results show a very good reduction in PAPR for input this inherent phase shift associated with each of the
symbols with both constant and varying phase angles. QPSK symbols, and therefore, the phasing scheme does

not produce a reduced PAPR for OFDM systems that use

Index TermsOFDM, peak-to-average power ratio, phasing digital baseband modulation.
schemes This paper proposes a modification to the phasing
scheme proposed by NN [6]. The proposed modification
is derived from the instantaneous envelope power of an
I. INTRODUCTION
OFDM signal, and with an assumption that the baseband
Orthogonal Frequency Division Multiplexing (OFDM) modulation is performed using QPSK. However, the
is gaining popularity in wireless communications due to formulation obtained could be used for any PSK methods.
its high data rate capability and robustness to hostile The Newman phasing scheme [5] also suffers similar
channel interference [1]. One major drawback of OFDM limitations in that it cannot reduce the PAPR if the input
is its large peak-to-average power ratio (PAPR), which symbols are random. The proposed modification can also
results in reducing the efficiency of the RF power be applied to the Newman phasing scheme and the
amplifier used in the transmitter. Several techniques have reduction of PAPR is equally good for both constant
been proposed in the literature, and an overview of amplitude or varying amplitude input signals.
various PAPR reduction techniques is available in [2]. One of the drawbacks of the NN, the Newman and the
Phasing methods have been investigated by many proposed phasing schemes is that the receiver must have
researchers [3]-[4] as a tool for reducing PAPR, wherein a-priori knowledge of the phase shift being given to each
the initial phase of each tone is appropriately set to symbol, or the information must be transmitted as side
minimize the PAPR. While it has been shown in [3] that a information. This will increase the transmission overhead.
PAPR of 2.60dB could be achieved by using a Newman This paper is structured as follows. The proposed
phasing scheme [5], Narahashi and Nojima (NN) [6] have modification to the NN phasing scheme is derived in
proposed an improved phasing method. The PAPR values Section 2. Section 3 provides the results of numeric
achieved with their phasing scheme never exceeded calculations and simulation, where a comparison of
3.01dB and show improved PAPR reduction over PAPR obtained using the proposed method is compared
Newman phasing for tone numbers 3, 4, 6, 7, 9, 12, 14, with that obtained using NN phasing. Section 4 is a
17, 22, 24, and 29 [6]. However, the phasing scheme conclusions section and summarises the main results from
proposed in [6] gives the stated PAPR values only if the the letter.
complex tones have constant amplitude and zero or
constant phase shifts. As a result the phasing method II. DERIVATION OF PHASING SCHEME
The instantaneous amplitude of a baseband OFDM
Manuscript received December 13, 2012; revised April 17, 2014. signal can be written as

2015 Engineering and Technology Publishing 80

doi: 10.12720/ijeee.3.1.80-83
 International Journal of Electronics and Electrical Engineering Vol. 3, No. 1, February, 2015

N 1 equation reduces to zero. Therefore, for QPSK, the

~
x(t ) X n exp j(2nt / T ) (1) average power of the OFDM signal under consideration
n 0
is therefore given by
~
where X n X n exp j ( n ) is the complex QPSK symbol,
where n is the phase angle of the QPSK symbol, n
represents the number of the subcarrier and T is the

E x1 (t )
2
E X
N 1

n 0
n
2

N 1
E 1 N
n 0
(5)

symbol duration. The phasing scheme derived in [6]

Peak-to-Average Power Ratio (PAPR)
would reduce PAPR only if the phase angle, n , is a
constant for all symbols, however, QPSK mapping of PAPR is defined as the ratio of the maximum (peak) of
random input symbols results in varying values for n . the instantaneous envelope power to the average power.
Therefore, it is important to consider the phase angle of Thus,
QPSK symbols in the derivation of the proposed phasing

max x1 (t )
2

PAPR

scheme. (6)
2
We can rewrite (1), as E x1 (t )
N 1
x(t ) X n exp j( n 2nt / T ) (2) Substituting (4) and (5) in (6),
n 0
1 N 2 N 1
In an attempt to reduce the PAPR of OFDM signal, PAPR max N 2 cos
N n0 mn1
each QPSK symbol is given an initial phase shift. The
OFDM signal as a result of the phase-shifted QPSK
symbols can then be represented as follows: 2
max 1 P(t ) (7)
N 1 N
x1 (t ) X n exp j ( n n 2nt /T )
n 0
where,
N 2 N 1
where n [0 ,1,..., N 1 ] is the initial phase shift applied P(t ) cos (8)
to each QPSK symbol. The magnitude of QPSK symbols n 0 m n 1
is unity and, therefore, the above equation can be written
as Equation (8) can be written as
N 2

x1 (t )
N 1
exp j( n n 2nt / T ) (3)
P(t ) cos2t / T k 1 k k 1 k
k 0
n 0
N 3
The instantaneous envelope power is given by, cos2 2t / T k 2 k k 2 k
k 0
N 1 N 4
P1 (t ) x1 (t ) exp j ( n n 2nt / T ) cos2 3t / T k 3 k k 3 k
2
(9)
n 0 k 0

N 1
exp j ( n n 2nt / T ) cos2 ( N 1)t / T N 1 0 N 1 0
n 0
P(t) obtained in (9) represents the fluctuations from
N 2 N 1 average power of the OFDM transmission. The objective
N 2 cos (4) is to find the initial phase angles that minimize P(t). It is
n 0 m n 1
difficult to obtain solutions considering all terms in the
where right-hand side of (9). Based on the solution proposed in
[6], which considers only the first summation term of P(t),
( n m n m 2 (n m)t / T ) a solution to (9) can be formulated as shown below in
(10).
The average envelope power is given by,
k 1 k k k 1
Pavg (t ) E x1 (t ) 2
2 (10)
k 1 k k k 1
where E{.} is the expectation value. Using (1), we get an N 1
expression for average power as follows, Rearranging the terms in (10), gives
[( k 1 k 1 ) ( k k )

E x1 (t )
2
E X
N 1
n
2
2
N 2 N 1
X n X m cos 2
n 0 n0 mn1 ( k k ) ( k 1 k 1 )]
N 1
For orthogonal signals, as the peak of one symbol falls
at the null of other symbol, the second term in the above Replacing with , we get,

2015 Engineering and Technology Publishing 81

 International Journal of Electronics and Electrical Engineering Vol. 3, No. 1, February, 2015

2
( k 1 k ) ( k k 1 ) (11)
mapped using QPSK modulation. This ensures that the
N 1 input symbols do not have a constant initial phase shift
and helps to prove that the proposed modified phasing
Using the solution to k proposed in [6], we get scheme is better that the NN phasing scheme, especially
(assuming 1 2 0 ) , when any OFDM system deals with random input signals.

(k 1)( k 2) A. Modified vs. NN Phasing Schemes

k Table I gives the values of PAPR for the proposed
N 1
modified phasing scheme as a function of the number of
and therefore, subcarriers and compares this with the PAPR for NN
(k 1)( k 2) phasing and the maximum theoretical PAPR value when
k k (12) no phasing scheme is employed.
N 1
It can be seen that though NN phasing can produce a
Equation (12) gives an algebraic expression for the PAPR as low as 2.65dB [6] when the input symbols are
initial phase shift to be added to QPSK symbols. of constant amplitude and constant inherent phase shift, it
does not produce low PAPR when the input signal is
TABLE I. PAPR COMPARISON
random. However, the modified phasing scheme
PAPR (dB) suggested in this paper produces a PAPR as low as about
No. of Theoretical 2.61dB. As the number of subcarriers increases beyond
subcarrier Modified NN phasing maximum 200, the PAPR converges to around 2.61dB.
(N logN)
2.00 3.01 3.01 3.01
3.00 2.22 4.77 4.77
4.00 2.58 5.17 6.02
5.00 2.92 3.44 6.99
6.00 2.57 4.91 7.78
7.00 2.70 5.73 8.45
8.00 2.89 4.61 9.03
9.00 2.48 5.04 9.54
10.00 2.76 6.91 10.00
12.00 2.62 6.06 10.79
14.00 2.71 7.06 11.46
16.00 2.80 6.47 12.04
18.00 2.81 4.63 12.55
Figure 1. Simulated PAPR vs. number of subcarrier for modified and
20.00 2.75 5.48 13.01 NN phasing schemes.
22.00 2.68 4.75 13.42
24.00 2.68 6.05 13.80
26.00 2.74 6.69 14.15
28.00 2.74 8.18 14.47
30.00 2.76 6.21 14.77
32.00 2.75 5.55 15.05
64.00 2.69 7.27 18.06
100.00 2.65 8.07 20.00
128.00 2.64 7.34 21.07
200.00 2.61 9.89 23.01
256.00 2.62 7.75 24.08
300.00 2.61 7.35 24.77

III. NUMERICAL AND SIMULATION RESULTS

Figure 2. Theoretical and simulated PAPR vs. number of subcarriers
The PAPR due to the proposed phasing scheme is
numerically computed and compared with the PAPR due The numeric results obtained have been verified to
to the NN phasing method and also with the theoretical match with the simulation results. An OFDM system with
maximum PAPR when no phasing method is used. For various phasing schemes has been simulated using
computation of PAPR using NN phasing, the information Matlab, and the results are shown in Fig. 1. It can be seen
source used is a random binary sequence, which is that the modified phasing scheme produces a much

2015 Engineering and Technology Publishing 82

 International Journal of Electronics and Electrical Engineering Vol. 3, No. 1, February, 2015

reduced PAPR level when compared to the NN phasing REFERENCES

method. [1] J. A. C. Bingham, Multicarrier modulation for data transmission:
Fig. 2 compares the PAPR values due to the modified An idea whose time has come, IEEE Commun. Mag., pp. 5-14,
phasing schemes that are obtained theoretically using (8) 1990.
[2] S. H. Han and J. H. Lee, An overview of peak-to-average power
and through simulation. It is also noted that the PAPR ratio reduction techniques for multicarrier transmission, IEEE
from simulation closely matches with the theoretical Wireless Communications, pp. 56-65, Apr. 2005.
value, thus validating the simulation model. [3] S. Boyd, Multitone signals with low crest factor, IEEE Trans.
Circuits Syst., vol. 33, no. 10, pp. 1218-1022, Oct. 1986.
B. Peak Envelope Power [4] B. M. Popovic, Synthesis of power efficient multitone signals
with flat amplitude spectrum, IEEE Trans. Commun., vol. 39, no.
Fig. 3 provides a plot of peak envelope power of a 7, pp. 1031-1033, Jul. 1991.
simulated OFDM system containing 16 subcarriers with [5] D. J. Newman, An L1 extremal problem for polynomials, in
and without the modified phasing scheme employed. This Proc. Amer. Math. Soc., Dec. 1965, pp. 1287-1290.
plot clearly illustrates the ability of the phasing scheme in [6] S. Narahashi and T. Nojima, A new phasing scheme for
multitone signal systems to reduce peak-to-average power ratio,
containing the peak envelope power to within 3.01dB of Electronics and Communications in Japan (Part I:
the average power. Communications), vol. 80, no. 1, pp. 89-99, 1997.

Vallavaraj Athinarayanan was born in India

in 1965. He received his BE degree in
Electronics & Communication Engineering in
1987, ME degree in Microwave & Optical
Engineering in 1993, both from Madurai
Kamaraj University, India, and the PhD in
Communications Systems in 2008 from
Glasgow Caledonian University, UK.
In 1987 he became a lecturer at Mepco
Schlenk Engineering College, India. In 1999,
he joined the Caledonian College of Engineering, Oman. He has more
than two and a half decades of teaching and research experience.
Currently he is the Associate Dean Undergraduate Studies at
Caledonian College of Engineering, Oman. His research activity
includes peak power problems in multicarrier modulation systems,
wireless communication systems and signal processing. He has co-
Peak envelope power vs. time (16 subcarriers) authored a number of technical books published by McGraw Hill
International, Singapore and Tata McGraw Hill Publishing Company,
New Delhi. He has also published several research papers in the
IV. CONCLUSION international conferences and journals.
Professor Vallavaraj is a Senior Member of IEEE, a life member of
A modified phasing scheme is introduced that Indian Society for Technical Education and a member of Oman Society
overcomes the limitation of NN phasing to reduce the of Engineers.
PAPR of OFDM systems. The proposed modification is Brian G Stewart was born in the United
easy to implement and results in a significant reduction in Kingdom in 1959. He received his BSc (Hons)
PAPR for systems using either constant amplitude input degree in Physics and Astronomy in 1981 and
signals or random input signals. his PhD in Astronomical Instrumentation in
1985, both from the University of Glasgow,
The proposed phasing scheme has been validated Scotland, UK. He also received a BD (Hons)
through simulations and following results are obtained. in Theology in 1994 from the University of
The PAPR is contained within 3.01dB for OFDM Aberdeen, Scotland, UK.
systems with up to 300 subcarriers. The PAPR converges In 1985 he became a lecturer in the School of
Electronic and Electrical Engineering at the
to about 2.61dB as the number of subcarriers approaches now Robert Gordon University, Aberdeen, UK,. In 1998 he joined
and increases above 300. Glasgow Caledonian University, Scotland, UK and is currently a
A common drawback of such phasing schemes is that Professor in the School of Engineering and Computing. He has more
the receiver must have prior knowledge of the phase shift than 20 years of teaching and research experience. His research interests
lie in the fields of Communication Systems and High Voltage Partial
being given to each symbol, and hence the transmission Discharge. He has held numerous government and industrial research
overhead is increased. Investigations on reducing the grants and has published numerous research papers in international
overhead or enabling reception without a prior knowledge conferences and journals. He also holds a number of patents in the fields
of the phase shift require to be further investigated. of communication systems and high voltage instrumentation. Professor
Stewart is a Chartered Engineer, a Member of the IET and a Member of
the IEEE.

2015 Engineering and Technology Publishing 83