Content uploaded by Cesar A. Azurdia-Meza
Author content
All content in this area was uploaded by Cesar A. Azurdia-Meza on Sep 09, 2019
Content may be subject to copyright.
BER Performance of OFDM-Based Visible Light
Communication Systems
Pablo Palacios J´
ativa∗, Cesar A. Azurdia-Meza∗, Milton Rom´
an Ca˜
nizares‡,
David Zabala-Blanco§
∗Deparment of Electrical Engineering, Universidad de Chile, Chile
pablo.palacios@ug.uchile.cl, cazurdia@ing.uchile.cl
‡Departamento de Redes y Telecomunicaciones, Universidad de Las Am´
ericas, Ecuador
milton.roman@udla.edu.ec
§Department of Computing and Industries, Universidad Cat´
olica del Maule, Talca, Chile
davidzabalablanco@hotmail.com
Abstract—In this paper, OFDM techniques designed for In-
tensity Modulation/Direct Detection (IM/DD) modulation applied
to Visible Light Communications (VLC) systems are evaluated.
These are Asymmetrically Clipped Optical OFDM (ACO-OFDM),
DC Biased Optical OFDM (DCO-OFDM) and Asymmetrically
Clipped DC Biased Optical OFDM (ADO-OFDM). These tech-
niques are evaluated and compared for different types of M-QAM
modulation, in terms of Bit Error Rate (BER). It is shown that
ACO-OFDM, in general terms, presents a better performance
compared to the conventional DCO-OFDM and ADO-OFDM
schemes, based on the BER parameter established.
Keywords—ACO-OFDM, ADO-OFDM, DCO-OFDM, IM/DD,
OFDM-based VLC.
I. INTRODUCTION
Visible Light Communication (VLC) is a complementary
technique to radiofrequency (RF) communications. In compar-
ison with RF, it presents several advantages, such as resistance
to electromagnetic interference, because the signal emitted is
light, spectrum not licensed, low energy consumption due to
the characteristics of its transmitters and receivers, among
others [1]. Consequently, these characteristics lead it to be
part of the group of enabling technologies for 5G, such as
Cognitive Radio (CR) [2], Milimetric Waves (mm-waves) [3],
Non-Ortoghonal Multiple Access (NOMA) [4], etc. In the
VLC scheme, the Light-Emitting Diode (LED) is used for data
transmission as well as maintaining its main lighting function.
For the reception of the signal, the device most used in these
systems is the photodiode (PD).
Orthogonal Frequency Division Mltiplexing (OFDM) is
commonly used in wired broadband communication systems,
such as fiber optic and wireless communications, specifically in
Long Term Evolution (LTE), due to its remarkable advantages,
among them the Inter-Symbol Interference (ISI) resistance and
abetter performance against fading, which produces a high
transmission rate [5].
Currently, OFDM is being implemented as a modulation
technique to improve the performance of optical systems, and
achieve high transmission rates, as in [6], where transmission
is presented over 3 Gbps, through a single LED. In the OFDM
used for RF, the signals transmitted are bipolar (positive and
negative) and complex. Given the conditions of VLC systems
and their wireless optical environment in which an intensity
modulation / direct detection (IM/DD) technique is applied,
bipolar signals cannot be transmitted directly because the
intensity of the light cannot be negative. Therefore, OFDM
signals designed for IM/DD systems must be real and not
negative.
In order to convert the OFDM symbols into unipolar and
positive signals, several optical OFDM techniques have been
developed, among the best known, asymmetrically clipped
optical OFDM (ACO-OFDM) [7], DC Biased Optical OFDM
(DCO-OFDM) [8], and other forms based on ACO-OFDM
and DCO-OFDM, as the so-called Asymmetrically Clipped DC
Biased Optical OFDM (ADO-OFDM) [9].
To make an introductory comparison between these OFDM
schemes, it can be found that theoretically, ACO-OFDM only
modulates the odd subcarriers and trims the original bipolar
OFDM signal to zero, to build the positive signal [7], while
DCO-OFDM adds a DC bias to the original bipolar OFDM
signal and all the subcarriers carry data symbols [8]. Therefore,
because ACO-OFDM does not require any DC bias, it is
estimated to be better in terms of spectral efficiency and
average optical power [10]. In ADO-OFDM, the advantages
of the two techniques described above are combined, ACO-
OFDM is used in the odd subcarriers and DCO-OFDM is used
in the even subcarriers.
In this document, the performance of OFDM-based VLC
communication systems; such as ACO-OFDM, DCO-OFDM,
and ADO-OFDM are evaluated and compared, for different
types of M-QAM modulations, as it is an important factor in
OFDM schemes to achieve higher data rates. This analysis is
done in terms of the Bit Error Rate (BER), a key evaluation
metric of digital communication systems.
The organization of this paper is presented as follows: a
brief description of the state of the art on OFDM schemes
is reviewed in section II, The VLC system model and the
mathematical analysis of the OFDM schemes applied to the
VLC system are explained in section III. Then, in Section IV,
the performance evaluation and the numerical results of the
VLC system simulations are presented. The conclusions and
future directions are given in Section V.
II. STATE O F TH E ART
In order to know current works on OFDM techniques
applied to VLC systems, in [11] a hybrid and asymmetric
trimmed OFDM scheme, called EHACO-OFDM, is presented.
The scheme uses ACO-OFDM in the odd subcarriers, modu-
lation discrete multitone modulated in pulse amplitude (PAM-
DMT) in the imaginary parts of even subcarriers and DCO
-OFDM in the real parts of the pairs. Since all available
bandwidth is used for data modulation, this scheme can achieve
greater efficiency compared to traditional schemes. The OFDM
scheme proposed in [12], called µ-OFDM, is formulated
as a combination of the µ-law companding technique and
a transformation process of bipolar signals to the unipolar
ones that are required in VLC systems. Due to the inherent
characteristics of the companding operation, low PAPR signals
are obtained and the clipping noise is significantly eliminated.
To conclude, the most recent work in [13], the authors analyze
experimentally an improved adaptive scheme with paired cod-
ing (PWC) for a VLC OFDM system. Among its features are
the non-overload of coding and low complexity, through the
sub-carriers of low frequency (SC) modulated in 16 quadrature
amplitude modulation (16-QAM), while high frequency SC are
modulated by 4-QAM, and in between them, 32-QAM is used.
III. SYS TE M MOD EL A ND MATHEMATICAL ANALYSIS O F
OFDM SCHEMES
In this section, an indoor multi-cell VLC system mathemat-
ical model and the mathematical description of ACO-OFDM,
DCO-OFDM and ADO-OFDM is presented.
A. VLC System Model
This model is based on a basic principle of VLC tech-
nology, where the LED light (Transmitter) is modulated with
data and transmitted to the free space for lighting and wireless
communication in typical indoor environments [14]. According
to the literature, there are two main types of light components
that can be received by a PD (receiver): one is a component of
Line of Sight (LOS) and another that is the diffuse component
due to reflections of the signal in the surfaces that may exist
inside the room, that is, No Line of Sight (NLOS). Some works
has shown that the stronger diffuse component is much lower
in electrical energy than the LOS component in typical indoor
environments [15], so it will not be considered in this work,
and may be considered for future work.
The LED LOS irradiation follow a generalized Lambertian
pattern [16]. To generalize the system, we define Las the
number of LEDs (transmitters) and Kas the number of PDs
(receivers). The LOS optical channel gain between the lth LED
with l∈ {1,2...L}and the kth PD with k∈ {1,2...K}is
defined by [17]
h[k]
l=
(m+1)ApTs(θ[k]
l)g(θ[k]
l)
2π(d[k]
l)2cosm(φ[k]
l)cos(θ[k]
l)θ[k]
l≤Θ[k]
0θ[k]
l>Θ[k],
(1)
where d[k]
lis the distance from a LED lto an PD k,Apis the
physical area of PD, since it is assumed to be the same for all
users, θ[k]
lis the angle at which the light impacts on the PD of
the kth user, φ[k]
lis the irradiance angle, Ts(θ[k]
l)is the gain
of the LED optical filter, Θ[k]is the field of view (FOV) of
the PD in the kth user.
The mcomponent is the Lambertian index defined by [17]
m=1
log2(cos(θ1/2)) ,(2)
this expression indicates that it will be a function of half the
incidence angle of light in the PD. By last, g(θ[k]
l)is the
concentrator gain defined by [17]
g(θ[k]
l) = (η2
sin2(Θ[k])0≤θ[k]
l≤Θ[k]
0θ[k]
l>Θ[k],(3)
where ηis the reflective index of the PD. Therefore, the signal
received by the kth PD, is given as [17]
y[k]=h[k]x+n[k],(4)
where xis the signal transmitted by the lth LED and n[k]is
the additive noise of the kth PD. For our later analysis of the
OFDM schemes, we will focus on the xsignal emitted by the
LED.
B. Mathematical Analysis of OFDM VLC Schemes
For a more robust mathematical analysis, we will represent
the signal xtransmitted by some LED as a vector of complex
data signals and previously converted from serie to parallel
and modulated in M-QAM, defined by
X= [X0X1X2, ... , XN−1],(5)
where Xkis usually a QAM symbol. The vector Xis entered
into the fast inverse Fourier transform (IFFT). Previous to
this, to comply with the IM/DD characteristic of the VLC
technology, Xis limited to having Hermitian symmetry, as
shown
Xn=X∗
N−mfor 0< n < N/2,(6)
where the components X0=XN/2= 0. Through the
Hermitian symmetry process of the input, the output signal
of the IFFT is not complex, although, this reduces the spectral
efficiency by 50%.
After the process of Hermitian symmetry, the signal can
enter the IFFT, for which it is necessary to represent it in
the discrete time domain. Therefore, the sample of the time
domain kth of x,xkis called
xk=1
N
N−1
X
n=0
Xnexp(j2πnk
N),(7)
where Nis the number of points on the IFFT and Xnis the nth
subcarrier of signal X. Defined the output signal of the IFFT,
we will proceed to the description of the OFDM schemes.
1) DCO-OFDM Scheme: Due to the Hermitian symmetry
presented in (6), the number of unique data transporting
subcarriers present in DCO-OFDM is (N/2)-1. In Fig. 1, the
block diagram of this OFDM scheme is presented. The IFFT
output, x, is converted from parallel to serial (P/S), then a
cyclic prefix (CP) is added, its objective is to eliminate the
Intersymbol Interference (ISI) and its length depends on the
optical environment. The resulting signal is converted from
digital to analog (DAC), with an accuracy between 12 and
16 bits, and its speed is determined by the bandwidth used.
In the same block, after the DAC, a low pass filter (LPF) is
added, this is to reduce the noise at adjacent frequencies. In
this work, for ease, we will consider an ideal LPF. The output
of this block will be called x(t). For long Nvalues, the signal
x(t)can be modeled as a Gaussian random variable, with zero
mean and variance σ2=E{x2(t)}[18].
Fig. 1: DCO-OFDM System.
The main focus of the OFDM schemes for VLC is to
have a real and positive signal. The DCO-OFDM scheme has
its approach on converting the signal to positive, since the
conversion to real is made by the Hermitian symmetry. Thus,
after adding CP and making the DAC, a suitable DC bias is
added to x(t)and any remaining negative peaks are clipped
resulting in signal called as
xDCO (t) = x(t) + IDC .(8)
Given the characteristics of OFDM signals, having a high
peak-to-average power ratio (PAPR), a large IDC is necessary
in order to eliminate negative peaks [18]. The disadvantage of
a high IDC , is that the optical energy-per-bit to single sided
noise power spectral density (Eb(opt)/No), becomes large,
returning to the inefficient scheme in terms of optical power.
Therefore, instead of a high IDC , a moderate bias is used,
and the remaining negative peaks are cut out, resulting in a
clipping noise. In DCO-OFDM systems, all subcarriers carry
data symbols, so clipping noise affects all subcarriers.
According to, the IDC can be represented as the standard
deviation of x(t), defining the expression [10]
IDC =µpE{x2(t)},(9)
where µis a proportionality constant. Any negative peak that
remains after the addition of the IDC , is clipped to zero, as
seen in the example of Fig. 2, so this can generate distortion
and cause an increase in the Bit Error Rate (BER).
Fig. 2: DCO-OFDM signal clipping to zero.
Following with the detail of the block diagram of Fig. 1,
The signal xDCO (t), will be the input for an optical modulator,
which is responsible for converting the electrical signal into
optical signal to be transmitted by the LED. For ease, we will
assume an ideal optical modulator, therefore the electrical input
signal will be directly proportional to the optical output signal.
Finally, the optical signal is transmitted through the optical
channel defined at the beginning of this section.
In the receiver of the DCO-OFDM scheme, first the re-
ceived signal must be converted from an optical signal to an
electrical signal by the PD. In addition in this conversion a
shot noise is added, which affects the optical systems. The
post processing to this point is the same as a conventional
OFDM receiver [10].
2) ACO-OFDM: Unlike the DCO-OFDM technique, in
which all subcarriers carry data symbols, in ACO-OFDM only
odd subcarriers carry data symbols, while even subcarriers
form a bias signal that ensures that the transmitted OFDM
signal comply with the requirement to be positive [18]. In
compliance with this, the input signal to the IFFT, X, is defined
as
X= [0X10X3, ... , XN−1].(10)
In addition, the elements of vector X, previously, are
limited to having Hermitian symmetry as defined in 6. Thus,
the resulting signal in the time domain, xk, also to being real,
has the property of being antisymmetric, as shown
xk=−xk+N/2for 0< k < N/2,(11)
The block diagram in Fig. 3, represents the transmitter and
receiver of the ACO-OFDM scheme. As we can see, the ACO-
OFDM transmitter is similar to the DCO-OFDM transmitter,
where xpasses through the parallel to serial converter and
then adds the CP. It then passes through the DAC and is sent
through an ideal LPF that results in the signal x(t). Unable to
transmit the negative samples, the signal x(t)must be clipped
to zero, resulting finally the ACO-OFDM signal xACO(t)[19].
Fig. 3: ACO-OFDM System.
Due to the antisymmetry of the temporal signal x, the
clipping to zero does not produce any loss of information,
apart from that the distortion only falls on the even carriers,
which could produce a new noise that would interfere in the
odd carriers. Finally, the xACO(t)signal is input to an ideal
optical modulator and transmitted through the optical channel
described in (1). The processing in the receiver is similar to
a DCO-OFDM receiver, except that in ACO-OFDM only the
odd subcarriers are demodulated, since they only carry the data
symbols [19].
3) ADO-OFDM: In ADO-OFDM, a combined transmis-
sion of the ACO-OFDM and DCO-OFDM schemes is per-
formed. ACO-OFDM is transmitted on the odd subcarriers
and DCO-OFDM is transmitted on the peer subcarriers [20].
At reception, the odd subcarriers are demodulated with a
traditional ACO-OFDM receiver, but the even subcarriers are
demodulated after an interference cancellation process.
Fig. 4: ADO-OFDM Transmitter.
In Fig. 4 and Fig. 5, the block diagram of the ADO-OFDM
transmitter and receiver respectively is shown. The upper path
in the transmitter generates the ACO-OFDM signals while
the lower path generates the DCO-OFDM signals. As we can
see, the signal processing in ACO-OFDM is the same as in
Fig. 5: ADO-OFDM Receiver.
a traditional ACO-OFDM transmission [20]. The input data
vector, X, which is limited to have Hermitian symmetry, is
divided into odd component, Xodd and even component Xeven,
both with Hermitian symmetry, where
Xodd = [0X10X30, ... , 0XN−1],(12)
Xeven = [X00X20, ... , XN−20].(13)
These signals are the inputs of two separate IFFT blocks,
which produce two signals, xodd and xeven respectively. For
ACO-OFDM, the important signal is xodd, which is real but
still bipolar. Finally, by cutting the negative peaks of xodd to
obtain the unipolar signal, the ACO-OFDM signal is generated,
defined as
xACO =1
2xodd +nACO ,(14)
where nACO is the ACO-OFDM clipping noise.
In the DCO-OFDM part, the signal is produced by adding
a DC bias to the xeven signal, and clipping the remaining
negative peaks. Since xeven is generated from even subcarriers,
they will have even symmetry. Therefore, the resulting DCO-
OFDM signal is as follows
xDCO =xev en +nDCO +ID C ,(15)
where nACO is the DCO-OFDM clipping noise hat will affect
only the even index subcarrier and IDC is the DC component.
It is important to mention that the performance of the ACO-
OFDM is not affected by the clipping noise DCO-OFDM.
Following with the block diagram of Fig. 4, The xACO and
xDCO signals are serialized, converted to analog and filtered
through an ideal LPF, resulting in continuous signals in the
time domain xACO (t)and xDC O(t). To generate the ADO-
OFDM signal, we add xACO (t)and xDC O(t), obtaining
xADO (t) = xACO (t) + xDC O(t).(16)
Finally, a CP is added to the signal xADO(t)and then sent
to an ideal optical modulator.
For the ADO-OFDM receiver which is shown in Fig. 5,
there are two routes, one to demodulate the ACO-OFDM
symbols and another to demodulate the DCO-OFDM symbols.
The optical channel developed in (1), and a perfect equalizer
are assumed. For simplicity, we consider that the channel
noise consists of separate components due to the odd and
even received subcarriers, called nodd,h and nev en,h [20]. In
addition, since the optical channel is flat, the elements of
these two vectors are independent and identically distributed
Gaussian variables. The resulting signal is
y=xACO +xDC O +nodd,h +neven,h ,(17)
substituting (14) and (15) in (18)
y=1
2xodd +nACO +xeven +nD CO +IDC +nodd,h +nev en,h.
(18)
Since only xodd/2 and nodd,h contribute to the odd fre-
quency subcarriers, the data transmitted on the odd subcarriers
using ACO-OFDM can be easily retrieved at the receiver, so
it would be
Yodd =1
2Xodd +Nodd,h.(19)
Then, passing (18) through the FFT we obtain
Y=1
2Xodd+NAC O +Xeven +NDCO+ID C +Nodd,h+Neven,h .
(20)
For the DCO-OFDM signals demodulation, we previously
need to subtract the interfering ACO-OFDM signal. By means
of a local estimation of the ACO-OFDM signal and subtracting
it from the received signal. As shown in the diagram of Fig. 5,
Yodd obtained previously, will be input of an IFFT, obtaining
yodd, then it is multiplied by two to be able to estimate xodd
and clipping to zero, obtaining
yACO =xAC O +nACOodd,h +nACOeven,h.(21)
Since the ACO clipping results in equal total power on
even and odd subcarriers, nACOodd,h and nACOeven,h have
equal average noise power. This estimated ACO-OFDM signal,
yACO , is subtracted from 18, to obtain an estimated yeven
signal, which is passed by an FFT. The result will be even
subcarriers that are used to estimate the DCO-OFDM signal,
as follows
Yeven =Xeven +NDC O +IDC +Xeven +NAC Oeven,h +Neven,h .
(22)
As we can see, there are three sources of noise, the original
DCO clipping noise, the noise added in the channel to the
subcarrier pairs and an additional noise component caused
by the noisy estimation of the ACO-OFDM signal in the
interference cancellation process.
IV. SIMULATION RESULTS
In this section the BER performance simulations of a VLC
system based on the ACO-OFDM, DCO-OFDM and ADO-
OFDM schemes are presented and compared R as a function
of electrical energy per bit to single-sided noise power spectral
density (Eb/No)for different QAM constellations, in the
presence of an optical LOS channel. When using this model,
we are considering that the main source of noise is the electric
front end of the receiver, so the noise is bipolar (not unipolar).
For all OFDM schemes, the number of subcarriers is set to
512 and the length of CP is set to 128. For practical purposes,
the IDC of the DCO-OFDM scheme is 0.5 V [3] .
Fig. 6, shows the comparison of the OFDM schemes for the
modulations 4-QAM, 8-QAM and 16-QAM, this is for better
visualization and a fairer comparison between modulations and
OFDM schemes.
0 5 10 15 20 25 30
Eb/No, dB
10-4
10-3
10-2
10-1
100
BER
DCO-4QAM
ACO-4QAM
ADO-4QAM
DCO-8QAM
ACO-8QAM
ADO-8QAM
DCO-16QAM
ACO-16QAM
ADO-16QAM
Fig. 6: BER performance of ACO-OFDM, DCO-OFDM and
ADO-OFDM for 4-QAM, 8-QAM, 16-QAM
From Fig. 6, it is clear that the ACO-OFDM scheme
performance is better in general terms than the DCO-OFDM
and ACO-ODFM schemes for low M-QAM modulations.This
is due to two reasons: 1) Because ACO-OFDM transmits data
symbols on only half of its subcarriers, the bit rate of ACO-
OFDM is half that of DCO-OFDM for a given constellation.
Despite this contrast, only when DCO-OFDM uses a 4-QAM
modulation, its BER exceeds the ACO-OFDM with 16-QAM
modulation by approximately 2 dB. So, for low modulations
the BER perfomance in general is better for ACO-OFDM.
2) The BER performance of ADO-OFDM will be negatively
affected by the DC bias that is added in the DCO part of ADO-
OFDM, in addition to the ACO path estimation noise disturb-
ing the detection. Therefore, the traditional ACO-OFDM and
DCO-OFDM schemes greatly outperform the ADO-OFDM for
low M-QAM modulations.
Fig. 7, shows the comparison of the OFDM schemes for
the modulations 32-QAM and 64-QAM, which allows us to
have a better visual comparison of the schemes for higher
QAM modulations. It should be noted that the performance
of the BER, for the higher QAM modulations, is more similar
between the ACO-OFDM and DCO-OFDM schemes, because
the clipping noise becomes the dominant distortion factor in
both schemes. However, in general terms of the performance
of the BER, the ACO-OFDM scheme continues to perform
better than the DCO-OFDM and ADO-OFDM schemes.
0 5 10 15 20 25 30
Eb/No, dB
10-4
10-3
10-2
10-1
100
BER
DCO-32QAM
ACO-32QAM
ADO-32QAM
DCO-64QAM
ACO-64QAM
ADO-64QAM
Fig. 7: BER performance of the ACO-OFDM for 32-QAM,64-
QAM
V. C ONCLUSIONS
In this work, we evaluated a VLC system on different
OFDM schemes; specifically ACO-OFDM, DCO-OFDM, and
ADO OFDM techniques, which were compared in terms of
BER, for M-QAM modulations ranging between 4-QAM and
64- QAM. It is shown that, the ACO-OFDM scheme, besides
being a simple solution for signal dispersion, has a greater
optical capacity. The performance of DCO-OFDM depends
on the bias current that produces a clipping noise that limits
performance. Finally, in the ADO-OFDM scheme, the bias
current of the DCO-OFDM section affects the ACO-OFDM
section and the effect of estimating this signal produces a low
performance of the ADO-OFDM technique. This study was
conducted specifically under the BER parameter, for future
research, we can consider making an analysis in terms of
spectral efficiency and power utilization.
ACKNOWLEDGMENT
This work was partially funded by Project STIC AMSUD
19-STIC-08, Project 11160517, CONICYT PFCHA/Beca de
Doctorado Nacional/2019 21190489, and SENESCYT ”Con-
vocatoria abierta 2014-primera fase Acta CIBAE-023-2014”.
REFERENCES
[1] C. Wang, F. Haider, X. Gao, X. You, Y. Yang, D. Yuan, H. M.
Aggoune, H. Haas, S. Fletcher, and E. Hepsaydir, “Cellular architecture
and key technologies for 5G wireless communication networks,” IEEE
Communications Magazine, vol. 52, no. 2, pp. 122–130, February 2014.
[2] P. Palacios, A. Castro, C. Azurdia-Meza, and C. Estevez, “SVD de-
tection analysis in cognitive mobile radio networks,” in 2017 Ninth
International Conference on Ubiquitous and Future Networks (ICUFN),
July 2017, pp. 222–224.
[3] P. Palacios, J. J. Freire, and M. Rom´
an-Ca˜
niz´
ares, “Millimeter-Wave
Channel Estimation Using Coalitional Game,” in Conference on Infor-
mation Technologies and Communication of Ecuador. Springer, 2018,
pp. 3–17.
[4] P. P. J´
ativa, M. R. Ca˜
nizares, C. S. Arancibia, and J. J. Freire, “Non-
Orthogonal Multiple Access for Cognitive Mobile Radio Networks
in 5G Communications,” in 2019 22nd Conference on Innovation in
Clouds, Internet and Networks and Workshops (ICIN), Feb 2019, pp.
344–350.
[5] B. Farhang-Boroujeny and H. Moradi, “OFDM Inspired Waveforms for
5G,” IEEE Communications Surveys Tutorials, vol. 18, no. 4, pp. 2474–
2492, Fourthquarter 2016.
[6] D. Tsonev, H. Chun, S. Rajbhandari, J. J. D. McKendry, S. Videv, E. Gu,
M. Haji, S. Watson, A. E. Kelly, G. Faulkner, M. D. Dawson, H. Haas,
and D. O’Brien, “A 3-Gb/s Single-LED OFDM-Based Wireless VLC
Link Using a Gallium Nitride µLED,” IEEE Photonics Technology
Letters, vol. 26, no. 7, pp. 637–640, April 2014.
[7] Y. Sun, F. Yang, and L. Cheng, “An Overview of OFDM-Based
Visible Light Communication Systems From the Perspective of Energy
Efficiency Versus Spectral Efficiency,” IEEE Access, vol. 6, pp. 60 824–
60 833, 2018.
[8] M. Zhang and Z. Zhang, “An Optimum DC-Biasing for DCO-OFDM
System,” IEEE Communications Letters, vol. 18, no. 8, pp. 1351–1354,
Aug 2014.
[9] Y. Sun, F. Yang, and J. Gao, “Comparison of Hybrid Optical Modulation
Schemes for Visible Light Communication,” IEEE Photonics Journal,
vol. 9, no. 3, pp. 1–13, June 2017.
[10] L. Wu, Z. Zhang, J. Dang, and H. Liu, “Adaptive Modulation Schemes
for Visible Light Communications,” Journal of Lightwave Technology,
vol. 33, no. 1, pp. 117–125, Jan 2015.
[11] R. Guan and N. Huang and H. Wang and J. Wang and M. Chen, “Novel
spectral efficient OFDM for optical wireless communication,” in 2016
IEEE International Conference on Communications (ICC), May 2016,
pp. 1–6.
[12] Y. Yang and Z. Zeng and S. Feng and C. Guo, “A Simple OFDM
Scheme for VLC Systems Based onµ-Law Mapping,” IEEE Photonics
Technology Letters, vol. 28, no. 6, pp. 641–644, March 2016.
[13] J. He, J. He, and J. Shi, “An Enhanced Adaptive Scheme With Pairwise
Coding for OFDM-VLC System,” IEEE Photonics Technology Letters,
vol. 30, no. 13, pp. 1254–1257, July 2018.
[14] Z. Ghassemlooy, W. Popoola, and S. Rajbhandari, Optical wireless
communications: system and channel modelling with Matlab R
. CRC
press, 2017.
[15] C. Chen, D. Basnayaka, and H. Haas, “Non-line-of-sight channel
impulse response characterisation in visible light communications,” in
2016 IEEE International Conference on Communications (ICC), May
2016, pp. 1–6.
[16] T. Komine and M. Nakagawa, “Fundamental analysis for visible-
light communication system using LED lights,” IEEE Transactions on
Consumer Electronics, vol. 50, no. 1, pp. 100–107, Feb 2004.
[17] F. Miramirkhani and M. Uysal, “Channel Modeling and Characteri-
zation for Visible Light Communications,” IEEE Photonics Journal,
vol. 7, no. 6, pp. 1–16, Dec 2015.
[18] J. Armstrong and B. J. C. Schmidt, “Comparison of Asymmetrically
Clipped Optical OFDM and DC-Biased Optical OFDM in AWGN,”
IEEE Communications Letters, vol. 12, no. 5, pp. 343–345, May 2008.
[19] K. Asadzadeh, A. Dabbo, and S. Hranilovic, “Receiver design for
asymmetrically clipped optical OFDM,” in 2011 IEEE GLOBECOM
Workshops (GC Wkshps), Dec 2011, pp. 777–781.
[20] S. D. Dissanayake, K. Panta, and J. Armstrong, “A novel technique
to simultaneously transmit ACO-OFDM and DCO-OFDM in IM/DD
systems,” in 2011 IEEE GLOBECOM Workshops (GC Wkshps), Dec
2011, pp. 782–786.
