Enhanced Papr Reduction in OFDM Systems using Adaptive Clipping with Dynamic Thresholds

International Journal of Computer Networks & Communications (IJCNC) Vol.17, No.1, January 2025
DOI: 10.5121/ijcnc.2025.17104 45
ENHANCED PAPR REDUCTION IN OFDM
SYSTEMS USING ADAPTIVE CLIPPING WITH
DYNAMIC THRESHOLDS
Fatma Ben Salah 1
, Abdelhakim Khlifi 2
, Marwa Rjili 1
, Amina Darghouthi 1
and
Belgacem Chibani 1
1
MACS Laboratory, University of Gabes, Tunisia
2
Innov’COM laboratory, Sup’COM, University of Carthage, Tunisia.
ABSTRACT
Orthogonal Frequency Division Multiplexing (OFDM) is a highly efficient multicarrier modulation method
that is widely used in current high-speed wireless communication systems. It offers numerous benefits,
including high capacity and resilience to multipath fading channels, when compared to other techniques.
However, a significant drawback of OFDM is its high peak-to-average power ratio (PAPR), which can
result in in-band distortion and out-of-band radiation due to the non-linearity of high power amplifiers. To
address this issue, several techniques have been suggested, such as Selective Mapping (SLM), Partial
Transmit Sequence (PTS), Clipping, and Nonlinear Compounding, which will be discussed later in the
paper. The clipping technique, in particular, has been thoroughly analyzed as a simple and crucial method
for reducing PAPR. However, an arbitrary choice of clipping threshold can result in significant signal
distortion, degrading the transmission quality. Therefore, it is essential to find an optimum threshold that
minimizes PAPR while preserving signal quality, which is a challenging task. The classical clipping
scheme may not yield satisfactory results in this regard. This paper proposes a modified clipping scheme
that estimates the dynamic range of a noisy OFDM signal. The estimated parameters are then used to
determine the optimal threshold, which is more reliable than the previous technique that assumes an
arbitrary dynamic value. Simulation results indicate that the proposed modified clipping scheme has
achieved a PAPR reduction of 3.5 dB compared to the original OFDM.
KEYWORDS
OFDM; PAPR; Clipping; Threshold; Estimation; MUSIC; Eigen Decomposition
1. INTRODUCTION
The OFDM technology is a fundamental component of communication systems, offering
numerous advantages over other methods [1]. These benefits include the ability to achieve high
data rates, immunity to frequency selective fading and impulse interference, and tolerance of
multipath delay propagation. Additionally, OFDM technology provides high spectral power
efficiency, smooth equalization, and flexibility in hardware implementation [2,3,6]. Compared to
other multicarrier techniques, OFDM is known for efficient bandwidth utilization [1,2,11],
reduced susceptibility to echoes, and minimal non-linear distortion. These advantages have led to
the widespread adoption of OFDM in various wireless communication systems, such as BRAN,
WLAN, WiMAX, ETSI, LTE, DSL, and ADSL [2][4]. Furthermore, OFDM is useful in modern
applications, including F-OFDM, U-OFDM, GFDM, and COFDM. Despite these advantages,
OFDM faces challenges associated with the high peak-to-average power ratio (PAPR) [4,5]. This
can result in in-band (IB) and out-of-band (OOB) radiation, with the non-linear nature of the
High Power Amplifier (HPA) used in the transmitter being a significant contributor to the
problem [5,6]. The high PAPR also increases the complexity of using processing blocks such as

46
analog-to-digital converters (ADCs) and digital-to-analog converters (DACs) [11]. Therefore, it
is essential to address the PAPR issue. Instead of the classical solution of using an expensive
amplifier with a large linear range, it is advisable to find a solution to reduce the PAPR [5-7].
Many techniques have been proposed to reduce the PAPR, which can be divided into two
categories [5-8]. The first category consists of signal distortion techniques, including clipping
[15], clipping and filtering [14], peak windowing [10], active constellation extension (ACE) [14],
non-linear companding transforming [7,22], and trellis-assisted constellation subset selection
(TACSS) [8]. The second category includes signal scrambling techniques, such as selective
mapping (SLM) [16], partial transmit sequence (PTS) [19], block coding [14], tone reservation
(TR) [17], and tone injection [7].
This paper focuses on the clipping method, which is a distinguished PAPR reduction technique
that offers a simple and low-complexity solution. It reduces the PAPR to a predefined threshold
[9-11]. The clipping technique allows signal peaks to be avoided by using an arbitrarily chosen
threshold. However, an arbitrary threshold value can lead to signal distortion and compromise the
effectiveness of the method in terms of peak reduction. This is a common issue in classical
clipping approaches [14]. To address this issue, it is advisable to select a more appropriate
amplitude value before applying the clipping action. This can be achieved using various
estimation techniques, such as Pisarenko Harmonic Decom position (PHD), Multiple Signal
Classification (MUSIC) [13,15], Root-MUSIC [15], and Estimation of Signal Parameters via
Rotational Invariance Techniques (ESPRIT) [12].
In this paper, the MUSIC approach is used to analyze the dynamics of an OFDM signal and to
determine an optimal clipping threshold. This approach aims to improve the efficiency of the
PAPR clipping reduction method by selecting an appropriate threshold that improves
performance while minimizing signal distortion. The paper is organized as follows. Section 2
describes the related work. Section 3 analyzes the proposed modelling. Results and discussions
are debated in Section 4. Lastly, the paper’s conclusions are given in Section 5.
2. RELATED WORK
Reducing PAPR is a significant challenge in wireless communication systems that employ
OFDM modulation. Over time, several techniques have been proposed to address this challenge
while minimizing signal distortion and maintaining system efficiency. One commonly used
technique is the conventional clipping method, which involves limiting the amplitude of the
OFDM signal to a predefined threshold [9-11]. While this method effectively reduces PAPR, it
can also introduce adverse effects such as distortion and interference, which can negatively
impact overall communication quality [14]. Another technique is the encoding method, which
entails encoding the data bits before modulating them into OFDM symbols. This approach
reduces the probability of large peaks in the signal and subsequently reduces PAPR [20].
However, the encoding method can increase system complexity and introduce redundancy, which
can lead to a decrease in overall data throughput. The selective mapping method (SLM) is
another technique that generates multiple OFDM symbols from the same data block, each
utilizing a different phase sequence. The system then selects the symbol with the lowest PAPR
for transmission [16]. Although this method effectively reduces PAPR, it introduces some
complexity. Successful implementation of the SLM method requires transmitting additional
information to the receiver, which adds an extra cost to the system [16]. The Partial Transmission
Sequence (PTS) technique is an alternative. method that involves dividing the OFDM symbol
into multiple subblocks and applying different phase factors to each subblock. This strategic
manipulation aims to decrease the PAPR when combining the sub-blocks in the final signal [19].
However, this approach also introduces additional complexity to the system. To successfully

47
implement the PTS method, extra information needs to be transmitted to the receiver, resulting in
communication costs [19].
In recent years, various techniques have been proposed to enhance the performance of wireless
communication systems in reducing PAPR. For instance, a PAPR reduction method based on
deep learning was introduced in [21]. This method utilizes a deep neural network to predict and
decrease PAPR in OFDM signals.
3. SIGNAL MODELING
3.1. Basic OFDM Signal Modeling
The OFDM sequence is generated by summing all N modulated subcarriers when applying the
IFFT operation, the subcarriers may be orthogonal to each other [1-5]. To understand the concept
of OFDM, let X = {Xk, k = 0, 1…, N − 1}, be the complex representation of the input data block
symbols after the constellation mapping operation, where Xk represents the block data of the kth
subcarrier from 1 to N. Therefore, the complex baseband OFDM signal is defined as:
𝑥(𝑡) =
1
√𝑁
∑ 𝑋𝑘𝑒𝑗2𝜋𝑘∆𝑓𝑡
𝑁−1
𝑘=0
0 ≤ 𝑡 ≤ 𝑇 (1)
Where ej2πk∆f t
is the twist factor of the kth
subcarrier, T is the total time of the symbol, ∆f is the
frequency space between the subcarriers. The bandwidth of the symbol is B = N.∆f, and ∆f is set
to 1/T to ensure the orthogonality between the subcarriers of the symbol [1,2]. The baseband
OFDM signal is sampled using the Nyquist rate (t = T /N). Therefore, the discrete OFDM signal
can be expressed in the time domain as:
𝑥𝑛 =
1
√𝑁
∑ 𝑋𝑘𝑒
𝑗𝜋𝑘𝑛
𝑁
𝑁−1
𝑘=0
0 ≤ 𝑛 ≤ 𝑁 – 1 (2)
where n is the discrete sampling index, and the discrete OFDM signal vector is written as:
𝑥𝑛 = [𝑥0, 𝑥1, … , 𝑥𝑁−1] (3)
The principle of the OFDM system is shown in Figure 1. At the transmitter end, the input data is
mapped by one of the modulation (Mod) schemes [4]. The signal is then converted from serial to
parallel (S/P). An IFF is then applied. A cyclic prefix is added to the signal. The cyclic prefix
adds a copy of the end of the OFDM symbol to its beginning, then the received signal is
converted from parallel to serial (P/S) before being transmitted to the channel [5]. At the receiver
side, the received signal is converted from S/P and demodulated by FFT. Then the signal is
converted from P/S and demodulated to obtain data output as shown below:

48
Figure 1.OFDMTransmitterandReceiverBlock Diagram
3.1. PAPR in OFDM System
The presence of peaks in an OFDM signal corresponds to a high PAPR [5]. PAPR is defined as
the ratio of the maximum peak power to the average power of the signal, as expressed in equation
4:
𝑃𝐴𝑃𝑅{𝑋(𝑛)} = 10𝑙𝑜𝑔10
𝑚𝑎𝑥|𝑥(𝑛)|2
𝐸{|𝑥(𝑛)|2}
(4)
E{.}is the average power of the signal. The continuous-time representation of the OFDM signal
is obtained by applying the oversampling operation. This operation ensures that peaks not
captured in the PAPR calculation are taken into account [6]. To improve the accuracy of the
PAPR calculation, the discrete baseband signal is sampled with L ≥ 4 by inserting (L−1) N
zeropadding between samples, where L is the oversampling factor. As a result, the oversampled
OFDM signal is written as [3]:
𝑥𝑛 =
1
√𝑁𝐿
∑ 𝑋𝑘𝑒
𝑗𝜋𝑘𝑛
𝑁𝐿
𝑁𝐿−1
𝑘=0
0 ≤ 𝑛 ≤ 𝑁𝐿 − 1 (5)
The evaluation of the performance of PAPR is commonly done using the Complementary
Cumulative Distribution Function (CCDF) [6]. The CCDF represents the probability of the PAPR
value exceeding a specific threshold. According to the central limit theorem [5], when there is a
large number of subcarriers N, the real and imaginary parts of the OFDM signal in the time
domain follow a Gaussian distribution with a mean of zero and a variance of 0.5. To calculate the
CCDF of the time-domain signal at the Nyquist sampling rate, the appropriate formula is used.
[14]:
𝑃𝑟(𝑃𝐴𝑃𝑅 > 𝑃𝐴𝑃𝑅0) = 1 − (1 − 𝑒𝑃𝐴𝑃𝑅0)𝑁 (6)
Where PAPR0 is the threshold value. Furthermore, when oversampling L is conducted, the
CCDF of the OFDM signal can be written as [3-6]:
𝑃𝑟(𝑃𝐴𝑃𝑅 > 𝑃𝐴𝑃𝑅0) = 1 − (1 − 𝑒𝑃𝐴𝑃𝑅0)𝑁𝐿 (7)
The CCDF curves of the PAPR performance can be used to compare the effectiveness of
different PAPR reduction techniques [4]. A lower CCDF curve indicates a lower probability of
having a high PAPR value, which means a better PAPR reduction [3-6].

49
3.2. Classical Clipping
In this section, we explore classical clipping technique used to reduce the PAPR value[14]. As it
was formerly indicated, the classical clipping retains a predistortion technique used to reduce the
PAPR in OFDM systems [9,10]. The basic idea is to trim the peak of the OFDM signal above a
given named CR (Clipping Ratio) threshold. In this paper such clipping can be expressed as [11-
14]:
𝑥𝑐(𝑛) = {
𝑥(𝑛), |𝑥(𝑛)| ≤ 𝐶𝑅
𝐶𝑅
𝑥(𝑛)
|𝑥(𝑛)|
, |𝑥(𝑛)| > 𝐶𝑅
(8)
Where:
xc(n) represents the clipped output signal
x(n) is the input OFDM signal at time n
CR is a threshold. It defines the limit value beyond which the samples of the signal x(n) will be
clipped.
3.3. Contribution
To effectively apply clipping to a signal x, it is essential to have a deep understanding of the
signal's dynamics. However, in many cases, the clipping threshold is arbitrarily set without
proper justification, which may not accurately reflect the true characteristics of the signal, as
highlighted in [14]. To address this, an estimation approach can be employed to make a more
informed selection of the threshold. By utilizing the MUSIC algorithm, a more accurate clipping
can be applied to the signal, leading to a better assessment of its effects [13,15].
The MUSIC algorithm, renowned for its spectral analysis capabilities, is employed to achieve a
higher resolution of the signal's dynamics. It is noteworthy that the signal subspace technique
heavily relies on the covariance matrix, as discussed in [18]. This matrix is utilized to extract
information about the signal subspace from the OFDM signal, facilitating the separation of signal
components from different subcarriers and the reduction of noise. To estimate the signal, an
Eigen Decomposition of the noisy observed signal is performed, dividing it into two subspaces
[18]. In comparison, while the ESPRIT algorithm is more computationally efficient due to its
avoidance of spectral searches, MUSIC offers superior accuracy in identifying closely spaced
frequencies, which is crucial for resolving the fine spectral details in OFDM signals.
Additionally, MUSIC provides higher resolution and robustness to noise, making it well-suited
for applications where precision in frequency estimation and the separation of subcarriers are
critical for accurate signal reconstruction.
Figure 2 bellow summarizes our contribution:

50
Figure 2. Proposed Method
The flowchart of the proposed method, based on the MUSIC algorithm, is illustrated in Figure 3:
Figure 3. Proposed Method Flowchart using MUSIC Algorithm
Step1. Observed Signal
We have the mixture of signal OFDM, x(n), and additive noise, w(n), with variance σ2
at the
observation, y(n)
𝑦(𝑛) = 𝑥(𝑛) + 𝑤(𝑛) (8)
Where x(n) contains N subcarriers.
𝑥(𝑛) = ∑ 𝐴𝑘𝑒𝑘(𝑛)
𝑁
𝑘=1
(9)
𝑒𝑘 = [1, 𝑒𝑖2𝜋𝑓𝑘
, 𝑒𝑖4𝜋𝑓𝑘
, … , 𝑒𝑖2𝜋(𝑁−1)𝑓𝑘]
𝑇
(10)
where Ak, fk are amplitude and frequency respectively of the kth
subcarriers.

51
Step2.Correlation Matrix Calculation
Based on the observed signal defined by (8), the related autocovariance matrix can be obtained
by:
𝑹𝒚 = [
𝑅𝑦(0) … 𝑅𝑦(𝑁 − 1)
… … …
𝑅𝑦(𝑁 − 1) … 𝑅𝑦(0)
] (11)
With the Toplitz matrices, the auto covariance matrix can also be written as:
𝑹𝒚 = 𝑦𝑦𝐻 (12)
𝑹𝒚 = 𝑅𝑥 + 𝑅𝑤 (13)
Where Rx, I and H are autocovariance matrix for x, identity matrix and the Hermitian transpose
respectively and Rw = σ2
I. So, Rx can be written as:
𝑅𝑥 = ∑ 𝑃𝑘𝑒𝑘𝑒𝑘
𝐻
𝑁
𝑘=1
(14)
So we have:
𝑹𝒚 = ∑ 𝑃𝑘𝑒𝑘𝑒𝑘
𝐻
𝑁
𝑘=1
+ σ2
𝐼 (15)
Where PK represents the powers set of various OFDM subcarriers. It is possible to extract all the
parameters relating to x from the eigenvalues and eigenvectors by the Eigen Decomposition of
Ry described in the next step.
Step3. Eigen Decomposition
The MUSIC method relies on the Eigen Decom position where the eigenvalues of Ry are ordered
as λ1> λ2 > λ3... >λM. In the noise free case, the eigenvalues λ1,λ2,λ3 ...λN are non-zero and the
eigenvalues λN+1,λN+2,...,λM are zero. The Eigen Decomposition of autocovariance matrix of Ry
can be written as:
R = UΛUH (16)
Where U is the matrix of eigenvectors, Λ is the diagonal matrix containing the eigenvalues.
Λ = [
λ1 0 … 0
… λ2 … …
0 0 … λ𝑁
] (17)
U = [u1, u2, u3, … , u𝑀] (18)
Step4. OFDM amplitudes estimation
The amplitudes of the OFDM signal are estimated by focusing on the eigenvalues associated with
the signal subspace. The larger eigenvalues, corresponding to the first N values (N < M), indicate
the presence of the signal. Conversely, the remaining (M − N) smaller eigenvalues represent the
noise. The orthogonal nature of the noise eigenvalue, with each column of the matrix
representing the signal subspace, is exploited. Since each row of this matrix corresponds to a
source signal, the orthogonality simplifies the identification of the subcarriers associated with the

52
source signal. Consequently, the first rows of the matrix are dedicated to the source signal, while
the subsequent rows represent the noise. By equation (15) and (17), we have:
𝑅𝑥 = ∑ λ𝑘𝑢𝑘𝑢𝑘
𝐻
𝑁
𝑘=1
(19)
𝑅𝑤 = ∑ σ2
𝑢𝑘𝑢𝑘
𝐻
𝑁
𝑘=1
(20)
𝑅𝑌 = ∑(λ𝑘 + σ2
)𝑢𝑘𝑢𝑘
𝐻
𝑁
𝑘=1
+ ∑ σ2
𝑢𝑘𝑢𝑘
𝐻
𝑀
𝑘=𝑁+1
(21)
𝑹𝒚 = 𝑢𝑥𝑢𝑥
𝐻
+ 𝑢𝑤𝑢𝑤
𝐻 (22)
The signal subspace is: Ux=span {u1, u2,…,uN}. The noise subspace, is orthogonal to signal
subspace: Uw= span{uN+1,uN+2,...,uM}.
So, we can deduce the MUSIC Spectrum as:
𝑃𝑚𝑢𝑠𝑖𝑐(𝐾) =
1
𝑢𝑘𝑢𝑘
𝐻
𝑒𝑘𝑒𝑘
𝐻 (23)
As a result, the estimated amplitudes can also be deduced from the signal powers Pmusic(k) as
follows:
𝐴𝑘
̃ = √𝑃𝑚𝑢𝑠𝑖𝑐(𝐾) (24)
Step5.Clipping threshold selection
To select an optimal threshold 𝐶𝑅
̃ we have used the amplitudes estimates 𝐴𝑘
̃as:
𝐶𝑅
̃ =
𝑚𝑒𝑎𝑛(𝐴𝑘
̃)
max(𝐴𝑘
̃)
(25)
Where:
 𝐴𝑘
̃: These are the estimated amplitudes of the OFDM subcarriers, derived from the
MUSIC spectrum (Equation 24).
 𝑚𝑒𝑎𝑛(𝐴𝑘
̃) : This represents the average value of all the estimated amplitudes. The mean
gives a general idea of the overall strength of the signal across all subcarriers.
 max(𝐴𝑘
̃): This is the maximum value among the estimated amplitudes, identifying the
strongest subcarrier in terms of amplitude.
𝐶𝑅
̃ helps define a clipping threshold that balances between the average signal level and the
strongest signal component. The idea is to avoid excessive clipping, which can distort the signal,
while ensuring that weaker signals are still adequately captured.
Subsequently, we used this optimal threshold value to perform the new clipping procedure, as
indicated in the following equation:
𝑥
̃𝑐(𝑛) = {
𝑥
̃(𝑛), |𝑥
̃(𝑛)| ≤ 𝐶𝑅
̃
𝐶𝑅
̃
𝑥
̃(𝑛)
|𝑥
̃(𝑛)|
, |𝑥
̃(𝑛)| > 𝐶𝑅
̃
(26)
The proposed approach is also described in Algorithm1 below:

53
Algorithm 1 Proposed PAPR Reduction Approach
Input: Observed Signal: y=x+w, Number of Subcarriers N
Output :Estimates 𝐴𝑘
̃, 𝐶𝑅
̃ , PAPR of modified clipping
Procedure
 Noised OFDM signal y
 Applying MUSIC algorithm to signal y
1. Finding covariancematrix of y by equation (13)
2. Finding Eigen Decomposition of Ry as equation (22)
3. Extracting Eigen vectors Ui and Eigen values λi
4. Finding MUSIC Spectrum by 𝑃𝑚𝑢𝑠𝑖𝑐(𝐾) =
1
𝑢𝑘𝑢𝑘
𝐻𝑒𝑘𝑒𝑘
𝐻
5. Obtaining OFDM amplitudes by 𝐴𝑘
̃ = √𝑃𝑚𝑢𝑠𝑖𝑐(𝐾)
6. Select optimal threshold clipping as 𝐶𝑅
̃ =
𝑚𝑒𝑎𝑛(𝐴𝑘
̃)
max(𝐴𝑘
̃)
7. Applying Modified Clipping Action
End procedure
4. RESULTS AND DISCUSSIONS
This section presents the simulation results used to evaluate the performance of the proposed
PAPR reduction technique. The simulations were conducted using 64 subcarriers, with
Differential Phase Shift Keying (DPSK) symbols input into the OFDM system. A Monte Carlo
simulation was performed to assess the effectiveness of the proposed method.
The simulation results show that the proposed PAPR reduction technique outperforms the traditional
approach. Figure 4 compares the PAPR between the proposed method and conventional clipping. At a
clipping threshold of 0.9, the proposed technique achieves an improvement of approximately 2.5 dB over
the conventional approach, as indicated by the PAPR CCDF at 10−2
.Additionally, compared to the original
OFDM system, the proposed method successfully reduces the PAPR by approximately 3.5dB.
Figure 4. PAPR performance based on proposed scheme
Figure 5 bellow illustrates the Complementary Cumulative Distribution Function (CCDF) of the
PAPR for SNRs ranging from 3 dB to 10 dB. The results clearly show that as the SNR increases,
the PAPR decreases, suggesting that systems operating at higher SNRs experience less
pronounced signal peaks relative to the average power.

54
Figure 5. PAPR performance of OFDM with different SNR
Table 1 quantifies the relationship between SNR and PAPR. The measured PAPR values for
SNRs. These results further confirm the trend observed in the CCDF curves: as SNR increases,
the PAPR reduction becomes more effective.
Table 1. PAPR values with various SNR
SNR(dB) PAPR(dB)
3 9.2
5 8.1
7 7.1
10 6.2
To enhance the effectiveness of our proposed method, we conducted additional simulations by
varying the modulation type and the number of subcarriers. Figures 6 and 7 present the results.
Figure 6 displays the CCDF curves of the PAPR for BPSK, DPSK, QAM, and PAM modulation
schemes under identical conditions, using 64 subcarriers and a SNR of 10 dB. At a CCDF level
of 10−2
, the PAPR values are: DPSK (3.4 dB), PAM (3.8 dB), BPSK (4.2 dB), and QAM (4.8
dB).
DPSK exhibits the lowest PAPR due to its differential encoding, which smooths transitions
between symbols. PAM shows a slightly higher PAPR due to occasional increases in signal peaks
from a lack of phase continuity. BPSK has a marginally worse PAPR due to abrupt phase
changes. QAM has the highest PAPR due to simultaneous modulation of phase and amplitude,
causing significant signal fluctuations.
So, DPSK offers the best PAPR performance, followed by PAM. BPSK shows a higher PAPR,
and QAM has the highest PAPR, requiring careful management in power-constrained systems
.

55
Figure 6. PAPR performance of OFDM with differents modulation type
In Figure 7 we examined the PAPR in an OFDM system with varying numbers of subcarriers.
The results show that PAPR increases as the number of subcarriers grows. This behaviur is to be
observed, as the probability of large peaks in the transmitted signal increases in direct proportion
to the number of subcarriers. As the number of subcarriers is increased, the OFDM signal
becomes a sum of a greater number of independent modulated signals, which increases the
likelihood of constructive interference and thus results in higher PAPR values. This trend serves
to illustrate the difficulties inherent in PAPR reduction for systems comprising a considerable
number of subcarriers, a configuration that is becoming increasingly prevalent in modern
communication standards, including 5G and beyond.
Figure 7. PAPR performance of OFDM with differents Number of subcarriers
Table 2 summarizes how our approach reduces PAPR by an average of 2.75 dB to 3.34 dB,
depending on the number of subcarriers. It also highlights that the clipping threshold increases
with more subcarriers, indicating higher clipping levels for larger OFDM signals.

56
Table 2. PAPR values with various subcarriers
Number of Sub-
carriers
Estimated
Threashold
PAPR of Noisy
OFDM(indB)
(mean)
PAPR of
proposed method
(indB) (mean)
64 0.6 8.97 6.5
128 0.7 9.04 7
256 0.75 9.34 7.4
512 0.8 10 7.6
5. CONCLUSIONS
This research introduces a novel and effective approach to reducing the peak-to-average power
ratio (PAPR) in OFDM systems, achieving a substantial reduction of 3.5 dB compared to the
original OFDM and 2.5 dB compared to conventional clipping methods. A significant
contribution of this research is the integration of the Multiple Signal Classification (MUSIC)
algorithm for the estimation of the dynamic range of the OFDM signal. By accurately assessing
the signal dynamics, the MUSIC algorithm enables the precise determination of the optimal
clipping threshold, thereby achieving a balance between PAPR reduction and signal distortion
minimisation. The deployment of the MUSIC algorithm guarantees that the clipping threshold is
dynamically calibrated in accordance with the estimated characteristics of the OFDM signal,
thereby facilitating more efficient PAPR reduction without the introduction of excessive
distortion. This approach not only enhances the performance of power amplifiers by reducing
power consumption but also preserves the integrity of the transmitted signal, thereby improving
both efficiency and reliability in wireless communication systems.The results demonstrate the
potential of the proposed method to significantly enhance the performance of OFDM systems,
rendering it highly applicable in modern wireless communication networks where power
efficiency and signal quality are paramount.Moreover, the techniques developed in this study can
be extended to a range of wireless communication technologies, particularly as they evolve
towards more complex and bandwidth-intensive standards, such as 5G and beyond. The
combination of MUSIC-based dynamic range estimation with adaptive clipping offers significant
advantages in PAPR reduction across a range of platforms.
CONFLICTS OF INTEREST
The authors declare no conflict of interest.
REFERENCES
[1] Javier Lorca Hernando, Ana García Armada “Frequency-Modulated OFDM: A New Waveform for
High-Mobility Wireless Communications” in IEEE transactions on communications, VOL. 71, NO.
1,2023;
[2] F. Conceição, M. Gomes, V. Silva, R. Dinis, A. Silva, and D. Castanheira, “A Survey of Candidate
Waveforms for beyond 5G Systems” in Electronics, vol. 10, no. 1, p. 21, 2021.
[3] N. A. Galande and A. M. Shah, “Implementation of OFDM by using wavelet for optimization of
wireless communication system,” in IEEE International Conference on Recent Trends in
Electronics, Information & Communication Technology (RTEICT), 2016.
[4] A. Darghouthi, A. Khlifi, H. Shaiek, F. B. Salah, and B. Chibani, "A Hybrid Framework OTFS-
OFDM Based on Mobile Speed Estimation” in International Journal of Computer Networks and
Communications(ijcnc) 16(3):49-67, 2024.
[5] F. B. Salah, A. Khlifi, M. Rjili, and B. Chibani, "PAPR Reduction Based on FrFT in OFDM
Systems," SSD 2024, vol. 745, pp. 745-749, 2024.
[6] U. S.M. and M. H.B., "A new approach to improve the performance of OFDM signal for 6G
communication," International Journal of Computer Networks & Communications (IJCNC), vol. 14,

57
no. 3, pp. 121-134, Jun. 2022. doi: 10.5121/ijcnc.2022.14308.
[7] B. Adebisi, K. Anoh, and K. M. Rabie, “Enhanced nonlinear companding scheme for reducing
PAPR of OFDM systems,” in IEEE Syst. J., vol. 13, no. 1, pp. 65–75, Mar. 2019.
[8] K. Liu, L. Wang, and Y. Liu, “A new nonlinear companding algorithm based on tangent
linearization processing for PAPR reduction in OFDM systems,” in China Commun, vol. 17, no. 8,
pp. 133–146, Aug. 2020.
[9] Z. Xing, K. Liu, A. S. Rajasekaran, H. Yanikomeroglu, and Y. Liu, “A hybrid companding and
clipping scheme for PAPR reduction in OFDM systems,” in IEEE Access, vol. 9, pp. 61565–61576,
2021.
[10] Liu & al “An Enhanced Iterative Clipping and Filtering Method Using Time-Domain Kernel
Matrix” in IEEE Access, 2019.
[11] F. B. Salah, A. Khlifi, M. Rjili, and B. Chibani, "PAPR Reduction in OTFS Systems Using Multi
Weighted Fractional Fourier Transform Analysis," HSI 2024, pp. 1-6, 2024.
[12] Liu & al “An Improved ESPRIT Algorithm for DOA Estimation of Coherent Signals” in IEEE
Access,2019.
[13] C Liu, X Cheng, J Su, Y Yuan “Research and Simulation of Super resolution Root-MUSIC
Algorithm” in 4th International Symposium on Computer Engineering and Intelligent
Communications (ISCEIC), 2023.
[14] S. S. Haque and M. M. Mowla, "Performance improvement for PAPR reduction in LTE downlink
system with elliptic filtering," Int. J. Comput. Netw. Commun(IJCNC)., vol. 7, no. 1, pp. 1-10, Jan.
2015.
[15] Jin& al “Improved MUSIC algorithm for frequency estimation” in Signal, Image and Video
Processing, 2017.
[16] M. A. Taher & al “Sliding the SLM-technique to reduce the non-linear distortion in OFDM
systems”, in Elektrotechn, 2014.
[17] H. Liang & al “PAPR of orthogonal frequency division multiplexing systems using modified tone
reservation techniques”,in IJCS,2016
[18] Aidan O. T. Hogg; Vincent W. Neo; Stephan Weiss; Christine Evers “A Polynomial Eigenvalue
Decomposition Music Approach for Broadband Sound Source Localization” in IEEE Workshop on
Applications of Signal Processing to Audio and Acoustics (WASPAA),2021.
[19] R. P. S. M. Pallavi, G. Rajalakshmi, and S. R. Mahadeva Prasanna, "Efficient PAPR Reduction in
OFDM Systems Using Modified PTS Scheme," IEEE Access, vol. 9, pp. 30497-30504, 2021
[20] J. Wang, Z. Ma, and D. Qu, "Joint Source-Channel Coding and PAPR Reduction Scheme for OFDM
Systems," IEEE Transactions on Communications, vol. 71, no. 2, pp. 887-898,2023
[21] S. Chen, X. Zhang, and C. Zhao, "PAPR Reduction in OFDM Systems Using Deep Learning," IEEE
Access, vol. 8, pp. 57580-57589, 2020.
[22] M. M. El-Nabawy, M. A. Aboul-Dahab, and K. El-Barbary, "PAPR Reduction of OFDM Signal by
Using Combined Hadamard and Modified Meu-Law Companding Techniques," International
Journal of Computer Networks & Communications (IJCNC), vol. 6, no. 5, pp. 71-84, Sep. 2014
AUTHORS
Fatma BEN SALAH, born in Gafsa, Tunisia, in 1989, earned her Bachelor's degree in
Engineering in 2014 from the National School of Engineers of Gabes(ENIG), with a
specialization in Communication and Networking. Presently, she is pursuing her doctoral
studies in Telecommunication Networks. Fatma is currently a contracted professor at
ENIG, where she imparts her knowledge and expertise to students.
Abdelhakim KHLIFI serves as an assistant professor at the National Engineering School
of Gabes, Tunisia. He earned his Engineer degree in 2007, followed by a Master's degree
from the National Engineering School of Tunis in 2010, and a Ph.D. degree in 2015. His
teaching specializes in signal processing and digital communications. His primary research
interests center on the performance analysis of waveform optimization in 5G/6G systems.

58
Marwa RJILI born in Medenine, Tunisia, in 1991, obtained her Bachelor's degree in
Engineering in 2014 specializing in Communication and Networking. She is currently a
doctoral student in Electrical Engineering at the same institution. Additionally, Marwa
serves as a contractual lecturer at the Higher Institute of Computer Science and
Multimedia of Gabes.
Amina Darghouthiborn in Tozeur, Tunisia, in 1993, is a doctoral student researcher in
Electrical Engineering at the National Engineering School of Gabes (Tunisia). Additionally,
Amina serves as a contractual lecturer at the National School of Engineers of Gabes, sharing
her knowledge and expertise with students.
RHAIMI Belgacem Chibani is an Associate Professor in Computer Sciences & Information
Engineering (CSIE). He has been employed at the National Engineering High School at Gabes
(ENIG) since September 1991. After earning his Doctorate Thesis at the National Engineering
High School at Tunis (ENIT), he received his Ph.D. degree from ENIG, University of Gabes,
Tunisia, in 1992. He is a member of the Research Laboratory MACS at ENIG, where he
serves as an activities supervisor in the field of Signal Processing and Communications
Research. Currently, his research areas encompass Signal Processing and Mobile Communications. He is
affiliated with the University of Gabes, and his research interests include Information and Signal
Processing, as well as Communications Engineering. He has published numerous papers in international
conferences and journals, such as CESA, IFAC, autumn, spring, A2I, and Summer Schools. Additionally,
he is a member of the Communications Engineering staff at ENIG and has served as a Program Committee
Member for various top national schools and activities related to Communications.

Enhanced Papr Reduction in OFDM Systems using Adaptive Clipping with Dynamic Thresholds

More Related Content

Similar to Enhanced Papr Reduction in OFDM Systems using Adaptive Clipping with Dynamic Thresholds (20)

More from IJCNCJournal (20)

Recently uploaded (20)

Enhanced Papr Reduction in OFDM Systems using Adaptive Clipping with Dynamic Thresholds