Review

Volume xx(Issue), Year, Page

https://doi.org/10.53623/gisa.v1ix

Future OFDM-based Communication Systems Towards

6G and Beyond: Machine Learning Approaches
Filbert H. Juwono1*, Regina Reine2
1 Curtin University, Malaysia; filbert@ieee.org
2 Twigx Research, UK; reginareine@twigx.com
* Correspondence: filbert@ieee.org;

SUBMITTED: date; REVISED: date; ACCEPTED: date

ABSTRACT: The vision towards 6G and beyond communication systems demands higher
rate transmission, massive amount of data processing, and low latency communication.
Orthogonal Frequency Division Modulation (OFDM) has been adopted in the current 5G
networks and has become one of the potential candidates for the future communication
systems. Although OFDM offers many benefits including high spectrum efficiency and high
robustness against the multipath fading channels, it has major challenges such as frequency
offset and high Peak to Power Ratio (PAPR). In 5G communication network, there is a
significant increase in the number of sensors and other low-power devices where users or
devices may create large amount of connection and dynamic data processing. In order to deal
with the increasingly complex communication network, Machine Learning (ML) has been
increasingly utilised to create intelligent and more efficient communication network. This
paper discusses challenges and the impacts of embedding ML in OFDM-based communication
systems.

KEYWORDS: OFDM; 6G; machine learning; cyber-physial social system

1. Introduction

The fifth generation (5G) communication has revolutionized the relationship between
technology, individuals, and society which leads to the new philosophy of Society 5.0 [1]. At
the same time, the cyber-physical social system (CPSS) concept emerges where everything is
connected, and data-driven [2]. In CPSS, there are three type of communications: humans to
humans, humans and machines and machines to machines (M2M) [3]. In M2M network,
sensors are strategically positioned throughout the environment, resulting in enormous data
communication where processing data in high data rates become a challenging task. Intelligent
and efficient methods are required to satisfy the standard of the future communication network.
The following characteristics are desired in the 6G and beyond networks [4]:
intelligent, high energy efficiency, and always-on global coverage. The data rate and end-to-
end latency can be up to 1 Tb/s and 1 ms, respectively. As 6G and beyond communication
systems often considered as green communication systems due to the utilisation of low-power
devices [5]. OFDM has been proposed as the modulation waveform for 6G and beyond
networks due to its flexibility in merging with other technologies [6], [7].
This paper briefly discusses the benefit of artificial intelligence (AI) in OFDM-based
communication systems through the perspective of ML. The rest of this paper will be structured

1
 Green Intelligent Systems and Applications Volume(Issue), Year, Page

as follows. Section 2 will briefly discuss the OFDM-based communication systems and its
challenges. Section 3 presents the impacts of utilising ML in OFDM-based communications.
Section 4 presents the challenges in implementing ML in OFDM-based communications.
Finally, conclusion remarks are presented in Section 5 .

2. OFDM-based Communication Systems and Its Challenges

The fundamental principle of OFDM is to divide a stream of high-rate data into many of parallel
low-rate data streams which are mapped to a number of orthogonal subcarriers. The spectrums
of the OFDM subcarriers overlap with each other without introducing interference.

Consider the transmitter part of an OFDM system as illustrated in Fig. 1 where the total
bandwidth, W, is divided into N subcarriers. The modulated symbol for the kth subcarrier is
represented as bk, with symbol interval Ts. The spectrums of the OFDM subcarriers overlap
with each other without introducing interference. Digital OFDM system can be implemented
by employing inverse Discrete Fourier Transform (IDFT) at transmitter and DFT at receiver.

Fig. 1. OFDM system.

Bit stream, bk , is passed to the modulation block (M-QAM or M-PSK, where M is the
modulation order) and subsequently passed to IDFT block, i.e. xk  IDFT  X k  to form
OFDM signal. The analog baseband of the OFDM transmitted signal 𝑥𝑘 can be written as

N 1
xk   bk e j 2 kt / T , 0  t  T (1)
k 0

Guard interval (GI) may be appended in OFDM signal to eliminate intersymbol interference
(ISI), resulting in a longer sequence x k . GI can be in a form of either cyclic prefix or zero
padding. The OFDM signal is then transmitted over the communication channel and received
by the receiver. The received signal is expressed by

yk  x k  hk  nk (2)

where hk is the channel impulse response (CIR) coefficients and nk is the additive noise. After
the GI is removed, the received OFDM signal is passed to the DFT and demodulation blocks,
respectively.

2
 Green Intelligent Systems and Applications Volume(Issue), Year, Page

OFDM-based communication systems have two major challenges: frequency offset and high
peak to average power ratio in the transmitted OFDM signal.

A. Frequency Offset

Due to the orthogonality between the OFDM subcarriers, the receiver can demodulate the
transmitted data without introducing inter carrier interference (ICI). However, frequency
mismatch (frequency offset) between the transmitter and receiver, causing the carriers to
become no longer orthogonal [10]. Frequency offset results in system performance degradation
in terms of signal-to-noise ratio (SNR) and hence bit error rate (BER).

Support Vector Machine (SVM) has been employed to estimate frequency offset in
demodulation process as discussed in [11]. It was shown that the system with SVM can achive
relatively the same BER as the maximum likelihood (ML) demodulator. Symbol timing offset,
in addition to carrier offset, can produce phase distortion and intersymbol interference (ISI).
SVM was proposed by the authors of [12] to tackle symbol timing offset. The proposed method
was demonstrated to be capable of estimating the timing offset, albeit the performance is
dependent on the number of pilots.

The strict orthogonality and synchronization in OFDM system will restrict the capability of
M2M communication where sporadic traffic is generated by Internet of Things (IoT) or other
devices in the M2M network. This type of devices should not be forced to be integrated into
the strict synchronization procedure. Ideally, these devices awake occasionally, and then
should transmit their messages in an instance without strictly synchronized. Future
communications such as 6G communication systems are envisioned to connect devices over
the air, underwater, and space. Hence, more efficient techniques in channel estimation to
minimize frequency offset are required.

B. High Peak to Average Power Ratio

Another major challenge in OFDM-based communication is the large PAPR in the transmitted
signal. PAPR is an essential performance parameter of wireless communication system as it
reflects the cost and energy efficiency of the hardware requirement. Power efficiency is always
an important factor in mobile communication networks to keep the total operational cost low.
High PAPR of the transmitted OFDM signal has been one of the major drawbacks in OFDM
system design and implementation. The large peaks resulted from high PAPR requires
additional back-off to prevent nonlinearity. Nonlinearity will create intermodulation between
carriers and introduce signal distortion, thereby deteriorating the BER. The additional back-off
means larger dynamic range. This type of HPA is costly and it consumes more power for
portable systems, such as IoT devices. This is a major impediment in implementing OFDM for
portable appliances. With the significant increase of utilising IoT devices in various systems,
reducing the PAPR of the OFDM transmitted signal has been a popular topic in wireless
communication. Various conventional methods have been proposed to reduce the PAPR in
OFDM communication systems, such as clipping [13], tone reservation [14], [15], and pulse
shaping [16]. However, these techniques have various limitation in minimising the PAPR.

3
 Green Intelligent Systems and Applications Volume(Issue), Year, Page

3. Machine Learning in OFDM based Communications

Data-intensive applications over low power devices will be the future of the fully connected
world. In CPSS, data and humans become the focus of the technology where data will be
transmitted and collected from machines/devices and humans. The technology will create a
connecting world where the collected data are used for solving the society problems. ML-
OFDM has been proposed as an intelligent and efficient solution to solve various issues in the
communication systems.

A. ML-OFDM for Channel Estimation

The use of deep learning (DL) has been proposed to be used at the OFDM receiver [8]. Training
datasets are created offline under various environment condition where the received signals are
recorded and used as training data for the system to learn. Furthermore, ML can be used to
estimate channel impulse response (CIR) coefficients for equalization purpose [9]. In the future
communication, the OFDM signal will be transmitted over the air, space or underwater
channels to create fully connected world. Underwater channel characteristics are different from
radio frequency (RF) as underwater channel have different scenario compared to air or space.
Underwater acoustic (UWA) communications are used for surveillance system, sensors or
unmanned vehicle underwater [1]. The underwater environment is complex yet reliability and
high data rate are required for reliable communications. OFDM-ML can improve the BER
performance in UWA compared to the existing optimization method [21].

B. ML-OFDM for Power Efficiency

An autoencoder of deep learning, called PAPR reducing network (PRNet), was used as
modulator and demodulator to both reduce PAPR and achieve undeteriotating BER in [13]. In
particular, the loss function of PRNet was designed to meet the minimum PAPR as well as
undeteriotating BER. Simulation results showed that PRNet outperforms traditional clipping
and partial transmit scheme (PTS) methods. In [14], a low complexity extreme machine
learning (ELM)-based tone reservation was proposed to efficiently reduce the PAPR.
Compared to PRNet, the PAPR of ELM-based tone reservation is higher due to its low
complexity. However, it can achieve better symbol error rate (SER). Another DL-based tone
reservation technique was also proposed in [15]. Furthermore, ML techniques have also been
used to lower out-of-band (OOB) emission. In [16], autoencoder together with orthogonal
precoding were used to achieve both low PAPR and OOB emission. In [22], ML was used to
improve the performance of clipping and filtering to reduce the PAPR of the OFDM signals.

C. ML-OFDM for M2M Networks

Future communication systems should have flexibility, not strict orthogonality and
synchronisation, where different applications can be addressed by a single solution with
different parameter settings instead of using multiple solutions, able to control the interference
among multiple users and maintain the relationship with other technologies [23], [24], [25].
4
 Green Intelligent Systems and Applications Volume(Issue), Year, Page

Implementing a large amount of IoT devices in M2M network requires efficient and effective
infrastructure to detect failure. Phenomenon called “sleeping cells” can reduce the performance
of the cell without triggering any warning and can be undetected by the system. In [26], the
authors proposed to use ML framework to detect this type of failure efficiently.

D. ML-OFDM for Biomedical

6G technology has been envisioned to support Internet of BioNano-Things (IoBNT) to provide

reliable health monitoring and diagnostic system with nano-bio sensors, machine learning, and
molecular communications (MC) concepts. In [27], the authors discussed the visibility of
diagnostic system based on biomarker detection using molecular communication principles
with ML algorithm.

4. Challenges in Machine Learning in OFDM based Communications

ML in OFDM-based communication shows potential to improve various performances and

effective in reducing inefficiencies in order to satisfy the standard of future communication
systems. However, implementing ML is complex and many of the current approaches rely on
dynamic training models. Producing high quality training models require high computational
cost and different parameters in the training models will produce different results. Training
models require a large number of datasets and sensitive data may be uploaded to the cloud
server. The sensitive data may be stored in a central location and may be accessed
inappropriately. This may lead to privacy and security issues. In time, the datasets may rapidly
increase in size and this will affect the speed of the processing which will cause an issue for
M2M network where low latency is required. Most current ML algorithms and training models
are mainly using centralized servers where super powerful servers are a necessity. Servers with
super high computational speed can be costly where it is one of the major restrictions for
researchers and small-medium businesses.

5. Conclusion

For future communication systems, such as 6G and beyond networks, embedding ML in

OFDM-based communication will improve the performance and efficiency. However, there
are challenges such as computational complexity, computational cost, distribution, security,
privacy, and hardware cost in implementing ML in the communication system that need to be
tackled.

References
[1] A. Deguchi, C. Hirai, H. Matsuoka, T. Nakano, K. Oshima, M. Tai, and S. Tani, “What
is society 5.0?” in Society 5.0, Hitachi-UTokyo Laboratory, Ed. Singapore: Springer,
2020.
[2] R. Reine, F. H. Juwono, Z. A. Sim, and W. K. Wong, “Cyber-physical-social systems:
An overview,” in Smart Connected World, S. Jain and S. Murugesan, Eds. Cham:
Springer, 2021.
[3] Y. Zhou, L. Liu, L. Wang, N. Hui, X. Cui, J. Wu, Y. Peng, Y. Qi, and C. Xing, “Service-
aware 6G: An intelligent and open network based on the convergence of communication,
computing, and caching,” vol. 6, no. 3, pp. 253 – 260, 2020.
5
 Green Intelligent Systems and Applications Volume(Issue), Year, Page

[4] K. B. Letaief, W. Chen, Y. Shi, J. Zhang, and Y. A. Zhang, “The roadmap to 6G: AI
empowered wireless networks,” IEEE Comm. Mag. Vol. 57, no. 8, pp. 84-90, 2019.
[5] S. Verma, S. Kaur, M. A. Khan, and P. S. Sehdev, “Toward green communication in 6G-
enabled massive internet of things,” IEEE Internet Things J., vol. 8, no. 7, pp. 5408 –
5415, 2021.
[6] Y. L. Lee, D. Qin, L-C. Wang, and G. H. Sim, “6G massive radio access networks: Key
applications, requirements and challenges,” IEEE Open J. Vehi. Technol., vol. 2, pp. 54-
66, 2021.
[7] W. Lu, P. Si, X. Liu, B. Li, Z. Liu, N. Zhao, and Y. Wu, “OFDM based bidirectional
multi-relay SWIPT strategy for 6G IoT networks,” China Commun., vol. 17, no. 12, pp.
80-91, 2020.
[8] H. Ye, G. Y. Li, and B.-H. Juang, “Power of deep learning for channel estimation and
signal detection in OFDM systems,” IEEE Wireless Comm. Lett., vol. 7, no. 1, pp. 114-
117, 2018.
[9] B. A. Jebur, S. H. Alkassar, M. A. M. Abdullah, and C. C. Tsimenidis, “Efficient machine
learning-enhanced channel estimation for OFDM systems,” IEEE Access, vol. 9, pp.
100839-100850, 2021.
[10] J. Lee, H.-L. Lou, D. Toumpakaris, and J. M. Cioffi, “Effect of carrier frequency offset
on OFDM systems for multipath fading channels,” IEEE Global Telecomm. Conf., 2004,
pp. 3721-3725, Dallas, TX, USA.
[11] P. Siyari, H. Rahbari, and M. Krunz, “Lightweight machine learning for efficient
frequency-offset-aware demodulation,” IEEE J. Selected Areas Commun., vol. 37, no.
11, pp. 2544-2558, 2019.
[12] S. Somarouthu, S. Manam, and A. Thakre, “Symbol detection in presence of symbol
timing offset using machine learning technique,” IEEE Int. Conf. Recent Advances
Innovations Eng., 2020, pp. 1-6, Jaipur, India.
[13] B. Tang, K. Qin, H. Mei and C. Chen, "Iterative Clipping-Noise Compression Scheme
for PAPR Reduction in OFDM Systems," in IEEE Access, vol. 7, pp. 134348-134359,
2019.
[14] T. Jiang, C. Ni, C. Ye, Y. Wu and K. Luo, "A Novel Multi-Block Tone Reservation
Scheme for PAPR Reduction in OQAM-OFDM Systems," in IEEE Trans. Broadcast.,
vol. 61, no. 4, pp. 717-722, Dec. 2015.
[15] C. Ni, Y. Ma and T. Jiang, "A Novel Adaptive Tone Reservation Scheme for PAPR
Reduction in Large-Scale Multi-User MIMO-OFDM Systems," in IEEE Wireless
Commun. Lett., vol. 5, no. 5, pp. 480-483, Oct. 2016.
[16] Z. A. Sim, R. Reine, Z. Zang, F. H. Juwono and L. Gopal, "Reducing the PAPR of GFDM
Systems with Quadratic Programming Filter Design," 2019 IEEE 89th Vehi. Technol.
Conf. (VTC2019-Spring), 2019, pp. 1-5.
[17] M. Kim, W. Lee, and D.-H Cho, “A novel PAPR reduction scheme for OFDM system
based on deep learning,” IEEE Commun. Lett., vol. 22, no. 3, pp. 510-513, 2018.
[18] Z. Li, N. Jin, X. Wang, and J. Wei, “Extreme learning machine-based tone reservation
scheme for OFDM systems,” IEEE Wireless Commun. Lett., vol. 10, no. 1, pp. 30-33,
2021.
[19] L. Li, C. Tellambura, and X. Tang, “Improved tone reservation method based on deep
learning for PAPR reduction in OFDM system,” Int. Conf. Wireless Commun. Signal
Process., 2019, pp. 1-6, 2019.
[20] R. Kuwahara and M. Ohta, “PAPR and OOBE suppression of OFDM signal using deep
learning,” IEEE Global Conf. Consumer Electron., 2020, pp. 905-906, Kobe, Japan.
[21] Y. Kim, H. Lee, J. Ahn, J Chung, “Selection of CDMA and OFDM using machine
learning in underwater wireless networks”, ICT Express,Vol. 5, Issue 4, 2019, pp. 215-
218.

6
 Green Intelligent Systems and Applications Volume(Issue), Year, Page

[22] I. Sohn and S. C. Kim, "Neural Network Based Simplified Clipping and Filtering
Technique for PAPR Reduction of OFDM Signals," in IEEE Commun. Lett, vol. 19, no.
8, pp. 1438-1441, Aug. 2015.
[23] J. Lin, W. Yu, N. Zhang, X. Yang, H. Zhang and W. Zhao, "A survey on Internet of
Things: Architecture enabling technologies security and privacy and applications", IEEE
Internet Things J., vol. 4, no. 5, pp. 1125-1142, Oct. 2017.
[24] F. Malandra, L. O. Chiquette, L.-P. Lafontaine-Bédard and B. Sansò, "Traffic
characterization and LTE performance analysis for M2M communications in smart
cities", Pervas. Mobile Comput., vol. 48, pp. 59-68, Aug. 2018.
[25] M. Z. Shafiq, L. Ji, A. X. Liu, J. Pang and J. Wang, "A first look at cellular machine-to-
machine traffic: Large scale measurement and characterization", Proc. 12th ACM
SIGMETRICS/PERFORMANCE Joint Int. Conf. Meas. Model. Comput. Syst.
(SIGMETRICS), vol. 40, pp. 65-76, 2012.
[26] O. G. Manzanilla-Salazar, F. Malandra, H. Mellah, C. Wetté and B. Sansò, "A Machine
Learning Framework for Sleeping Cell Detection in a Smart-City IoT
Telecommunications Infrastructure," in IEEE Access, vol. 8, pp. 61213-61225, 2020.
[27] F. H. Juwono, R. Reine, W. K. Wong, Z. A. Sim and L. Gopal, "Envisioning 6G
Molecular Communication for IoBNT Diagnostic Systems," 2021 Int. Conf. Green
Energy, Comput., Sustainable Technol. (GECOST), 2021, pp. 1-5.

© 2021 by the authors. This article is an open access article distributed under the
terms and conditions of the Creative Commons Attribution (CC BY) license
(http://creativecommons.org/licenses/by/4.0/).