0% found this document useful (0 votes)

81 views35 pages

Visible Light Communication Survey

This article provides a survey of visible light communication (VLC) technology. It discusses the advantages of LED lighting that enables VLC, including high energy efficiency and fast switching speeds. The survey covers VLC system components, physical channel characteristics, modulation methods, medium access techniques, system designs, and applications for indoor localization, gesture recognition, and more. It also outlines challenges that must be addressed to design high-speed mobile VLC networks.

Uploaded by

Tiến Nguyễn Phúc

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

81 views35 pages

Visible Light Communication Survey

Uploaded by

Tiến Nguyễn Phúc

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 35

This article has been accepted for publication in a future issue of this journal, but has not been

fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/COMST.2015.2476474, IEEE Communications Surveys & Tutorials
1

Visible Light Communication, Networking and

Sensing: A Survey, Potential and Challenges
Parth H. Pathak∗ , Xiaotao Feng,† , Pengfei Hu,∗ , Prasant Mohapatra∗
∗ Computer Science Department, † Electrical and Computer Engineering Department,
University of California, Davis, CA, USA.
Email: {phpathak, xtfeng, pfhu, pmohapatra}@ucdavis.edu

Abstract—The solid-state lighting is revolutionizing the indoor mens/watt in 2015 [1], and is projected to be around
illumination. Current incandescent and fluorescent lamps are 200 lumens/watt by the year 2020. This is a many fold
being replaced by the LEDs at a rapid pace. Apart from ex- increase compared to current incandescent and fluores-
tremely high energy efficiency, the LEDs have other advantages cent bulbs which provide an average luminous efficacy
such as longer lifespan, lower heat generation and improved of 15 and 60 lumens/watt [1] respectively. Similarly, the
color rendering without using harmful chemicals. One additional lifespan of LEDs ranges from 25,000 to 50,000 hours
benefit of LEDs is that they are capable of switching to different - significantly higher than compact fluorescent (10,000
light intensity at a very fast rate. This functionality has given rise hours). Apart from the energy savings and lifespan
to a novel communication technology (known as Visible Light advantages, the LEDs also have other benefits like com-
Communication - VLC) where LED luminaires can be used for
pact form factor, reduced usage of harmful materials in
high speed data transfer. This survey provides a technology
design and lower heat generation even after long period
overview and review of existing literature of visible light com-
of continuous usage. Due to these benefits, LED adoption
munication and sensing.
is on a consistent rise and it is expected that nearly 75%
This paper provides a detailed survey of (1) visible light
communication system and characteristics of its various com-
of all illumination will be provided by LEDs by the year
ponents such as transmitter and receiver, (2) physical layer 2030 [1].
properties of visible light communication channel, modulation The rapid increase in the usage of LEDs has provided
methods and MIMO techniques, (3) medium access techniques, a unique opportunity. Different from the older illumi-
(4) system design and programmable platforms and (5) visible nation technologies, the LEDs are capable of switching
light sensing and application such as indoor localization, ges- to different light intensity levels at a very fast rate. The
ture recognition, screen-camera communication and vehicular switching rate is fast enough to be imperceptible by a
networking. We also outline important challenges that need to human eye. This functionality can be used for commu-
be addressed in order to design high-speed mobile networks nication where the data is encoded in the emitting light
using visible light communication. in various ways. A photodetector (also referred as a
light sensor or a photodiode) or an image sensor (matrix
of photodiodes) can receive the modulated signals and
1 I NTRODUCTION decode the data. This means that the LEDs can serve
The indoor lighting is going through a revolution. The dual purpose of providing illumination as well as com-
incandescent bulb that has been widely used to lit our munication. In last couple of years, VLC research has
surroundings since its invention over a century ago is shown that it is capable of achieving very high data rates
slowly being phased out due to its extremely low energy (nearly 100 Mbps in IEEE 802.15.7 standard and upto
efficiency. Even in the most modern incandescent bulbs, multiple Gbps in research). The communication through
no more than 10% of the electrical power is converted visible light holds special importance when compared
to useful emitted light. The compact fluorescent bulbs to existing forms of wireless communications. First, with
introduced in 1990s have gained increasing popularity in the exponential increase of mobile data traffic in last two
the last decade as they provide a better energy efficiency decades has identified the limitations of RF-only mo-
(more lumens per watt). However, recent advancements bile communications. Even with efficient frequency and
in solid-state lighting through Light Emitting Diodes spatial reuse, the current RF spectrum is proving to be
(LEDs) have enabled unprecedented energy efficiency scarce to meet the ever-increasing traffic demand. Com-
and luminaire lifespan. Average luminous efficacy (how pared to this, the visible light spectrum which includes
much electricity is used to provide the intended illu- hundreds of terahertz of license free bandwidth (see
mination) of best-in-class LEDs is as high as 113 lu- Fig. 1) is completely untapped for communication. The

1553-877X (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/COMST.2015.2476474, IEEE Communications Surveys & Tutorials
2

3 KHz 300 MHz 300 GHz 430 THz 790 THz 30 PHz 30 EHz Frequency

Radio Microwave Infrared Visible Ultraviolet X-ray Gamma

105 m 1m 1 mm 750 nm 380 nm 10 nm 0.01 nm Wavelength

Red Orange Yellow Green Blue Violet

Fig. 1: Human eye can perceive the electromagnetic signals between the frequency range of 430 THz and 790 THz
which is referred as the visible light spectrum.

Visible Light Communication (VLC) can complement the layer design specifications. In last couple of years, the
RF-based mobile communication systems in designing achievable VLC link capacity has surpassed 1 Gbps, and
high-capacity mobile data networks. Second, due to its increasing research efforts are being directed towards
high frequency, visible light cannot penetrate through realizing the full potential of VLC.
most objects and walls. This characteristic allows one In this survey, we provide a systematic view of VLC
to create small cells of LED transmitters with no inter- research and identify important challenges. Specifically,
cell interference issues beyond the walls and partitions. we provide technology overview and literature review
It can also increase the capacity of available wireless of
channel dramatically. The inability of signals to penetrate 1) Visible light communication system components
through the walls also provides an inherent wireless and, details of transmitter and receiver characteris-
communication security. Third, VLC facilitates the reuse tics,
of existing lighting infrastructure for the purpose of 2) Physical layer characteristics such as channel
communication. This means that such systems can be de- model and propagation, modulation and cod-
ployed with relatively lesser efforts and at a lower cost. ing schemes, and Multiple-Input Multiple-Output
This untapped potential of visible light communication (MIMO) techniques,
has motivated us to compile this survey. 3) Link layer, multiple user access techniques and
The pioneering efforts of utilizing LEDs for illumi- issues,
nation as well as communication date back to year 4) System design and various programmable VLC
2000 when researchers [2] in Keio University in Japan platforms,
proposed the use of white LED in homes for building 5) Visible light sensing and applications such as vis-
an access network. This was further fueled by rapid re- ible light indoor localization, human computer in-
search, especially in Japan, to build high-speed commu- teraction, device-to-device communication and ve-
nication through visible light with development of VLC hicular communication applications.
support for hand-held devices and transport vehicles. Based on the review, we then outline a list of challenges
This led to formation of Visible Light Communications that need to be addressed in future research to realize
Consortium (VLCC) [3] in Japan in November of 2003. full potential of VLC.
VLCC proposed two standards - Visible Light Commu- The growing interest in VLC has resulted in a few
nication System Standard and Visible Light ID System surveys in past couple of years. This article differs from
Standard - by 2007. These standards were later accepted these surveys in many ways. In [9], authors discussed
by Japan Electronics and Information Technology In- LED-based VLC where the primary focus of discussion
dustries Association (JEITA) [4] as JEITA CP-1221 and was on design of physical layer techniques (modulation,
CP-1222 respectively. The VLCC also incorporated and circuit design etc.) that can enhance the performance
adapted the infrared communication physical layer pro- of VLC. Compared to [9], this article focuses on a
posed by international Infrared Data Association (IrDA) broader discussion about VLC, covering other aspects
[5] in 2009. In parallel, hOME Gigabit Access project of networking such as medium access as well as sensing
(OMEGA) [6], sponsored by European Union, also de- using visible light. Medium access protocols for VLC
veloped optical communication as a way to augment have been surveyed in [10], however, no comprehen-
the RF communication networks. In 2014, VLCA (Visible sive overview and comparison of networking techniques
Light Communications Associations) [7] is established as have been provided. Also, in this paper, we show that
a successor of VLCC in Japan for further standardization the usage of smartphone camera and light sensor for
of VLC. The first IEEE standard for visible light com- receiving visible light signals extend the VLC to other
munication was proposed in 2011 in the form of IEEE related fields of mobile computing and sensing. Multiple
802.15.7 [8] which included the link layer and physical research topics in this area such as indoor localization

and smartphone screen-camera communication are not TABLE 1: Acronyms and their full names
surveyed in any earlier work before this paper. In this Acronym Full form
paper, we provide a comprehensive survey of these ACO-OFDM Asymmetrically-Clipped Orthogonal Frequency Di-
vision Multiplexing
topics with additional focus on visible light sensing. Com- ADC Analog to Digital Converter
pared to [11] and [12] where authors surveyed free- AoA Angle of Arrival
space communication along with other forms of opti- BER Bit Error Rate
BIBD Balanced Incomplete Block Designs
cal wireless communications, the primary focus of this CAP Contention Access Period
survey is narrower and more detailed towards visible CCA Clear Channel Assessment
light communication. In another related survey, authors CC Convolutional Coding
CCM Code Cycle Modulation
provided a detailed overview of how optical wireless CRC Cyclic Redundancy Check
communication can be used for cellular network design CSK Color Shift Keying
in [13], with different aspects of outdoor environment CSMA-CA Carrier Sense Multiple Access - Collision Avoidance
DAC Digital to Analog Converter
and its impact on the communication performance. Com- DCO-OFDM Direct Current biased Optical Orthogonal Fre-
pared to this, our primary focus in this paper is on quency Division Multiplexing
visible light communication primarily in indoor settings. DMT Discrete MultiTone
DOPPM Differential Overlapping Pulse Position Modulation
Authors provided a brief survey of VLC applications in DPPM Differential Pulse Position Modulation
[14] with some discussion on vehicular networks and DSRC Dedicated Short-Range Communication
indoor broadcasting. However, in this paper, we survey EPPM Expurgated Pulse Position Modulation
FEC Forward Error Correction
a growing body of literature since the publication of [14] FET Field Effect Transistor
focusing on novel applications of VLC such as indoor FOV Field Of View
localization, screen-camera communication etc. We also FPS frames per second
Gbps Gigabits per second
detail various practical aspects of communication system GPS Global Positioning System
design by reviewing currently available programmable GTS Guaranteed Time Slot
platforms and LED transmitters/receivers. This will en- HCI Human Computer Interaction
HetNets Heterogeneous Networks
able researchers with RF communication background IM/DD Intensity Modulation/Direct Detection
to easily extend their expertise in visible light wireless JT Joint Transmission
access networks. Kbps Kilobits per second
LCD Liquid-crystal-display
The rest of the survey is organized as follows. We start LED Light Emitting Diode
by providing an overview of various components of a LOS Line Of Sight
visible light communication system with introduction LTE Long Term Evolution
MAC Medium Access Control
to LED luminaires and different types of receivers in Mbps Megabits per second
Section 2. In Section 3, we survey the physical layer MCS Modulation and Coding Scheme
properties of VLC with details on channel and propaga- MEPPM Multi-level Expurgated Pulse Position Modulation
MIMO Multiple Input Multiple Output
tion, modulation methods and MIMO techniques. It also MISO Multiple Input Single Output
includes an overview of VLC standard IEEE 802.15.7 [8]. MPPM Multipulse Pulse Position Modulation
This is followed by Section 4 where various link layer MU-MIMO Multiple User - Multiple Input Multiple Output
NFC Near Field Communication
and medium access protocols are discussed. Section 5 NRZ Non Return to Zero
describes various aspects of VLC system design and OCDMA Optical Code Division Multiple Access
surveys available programmable platforms that can be OFDMA Orthogonal Frequency Division Multiple Access
OFDM Orthogonal Frequency Division Multiplexing
used for research. Section 6 reviews a wide variety of OMPPM Overlapping Multipulse Pulse Position Modulation
topics in visible light sensing and applications which OOC Optical Orthogonal Codes
includes indoor localization, screen-camera communica- OOK On Off Keying
OPPM Overlapping Pulse Position Modulation
tion, vehicular communication and human-computer in- PAPR Peak to Average Power Ratio
teraction. Based on the review, Section 7 outlines various PDP Power Delay Profile
challenges that need further research in order to build PPM Pulse Position Modulation
PWM Pulse Width Modulation
high-capacity, mobile VLC networks. We have compiled QAM Quadrature Amplitude Modulation
the acronyms used throughout the paper and presented RC Repetition Coding
them with their full forms in Table 1. RF Radio Frequency
RGB Red Green Blue
RLL Run Length Limited
2 VLC S YSTEM OVERVIEW RS Reed-Solomon coding
RSS Received Signal Strength
In this section, we provide an overview of visible light RTS/CTS Request To Send/Clear To Send
communication system and its transmitter and receiver SFO Sampling Frequency Offset
SISO Single Input Single Output
components. We then discuss various modes of VLC. SLM Spatial Light Modulator
SMP Spatial Multiplexing
SM Spatial Modulation
2.1 VLC Transmitter SNR Signal to Noise Ratio
The transmitter in a visible light communication system VLC Visible Light Communication
VPPM Variable Pulse Position Modulation
is an LED luminaire. An LED luminaire is a complete WDM Wavelength Division Multiplexing

lighting unit which consists of an LED lamp, ballast, mercial photodetectors can easily sample the received
housing and other components. The LED lamp (also visible light at rates of tens of MHz.
referred as an LED bulb in simpler terms) can include An imaging sensor or a camera sensor can also be used
one or more LEDs. The lamp also includes a driver to receive the transmitted visible light signals. Because
circuit which controls the current flowing through the such camera sensors are available on most of today’s
LEDs to control its brightness. When an LED luminaire mobile devices like smartphones to capture videos and
is used for communication, the driver circuit is modified images, it has the potential to convert the mobile devices
(further details in Section 5) in order to modulate the in readily available VLC receivers. An imaging sensor
data through the use of emitted light. For example, in a consists of many photodetectors arranged in a matrix
simple On-Off Keying modulation, the data bit “0” and on an integrated circuit. However, the limitation of an
“1” can be transmitted by choosing two separate levels imaging sensor is that in order to enable high-resolution
of light intensity. photography, the number of photodetectors can be very
A crucial design requirement for VLC system is that high. This significantly reduces the number of frames
illumination, which is the primary purpose of the LED per second (fps) that can be captured by the camera
luminaries, should not be affected because of the com- sensor. For example, the fps of commonly used camera
munication use. Hence, performance of the VLC system sensors in smartphones is no more than 40. This means
is also affected depending on how the LED luminaires that direct use of camera sensor to receive visible light
are designed. White light is by far the most commonly communication can provide very low data rate.
used form of illumination in both indoor as well as The “rolling shutter” property of camera sensor can be
outdoor applications. This is because colors of objects used to receive the data at a faster rate. Due to a large
(also known as color rendering) as seen under the white number of available photodetectors in a camera sensor, it
light closely resemble the colors of the same objects is not possible to read the output of each pixel in parallel.
under the natural light. In solid-state lighting, the white Instead modern camera sensors employ row scanning
light is produced in following two ways - where photodetectors of one row of the matrix is read at
1) Blue LED with Phosphor: In this method, the a time. This procedure of reading photodetector output
white light is generated by using a blue LED that row by row (or column-by-column) is referred as rolling
has yellow phosphor coating. When the blue light shutter. Fig. 2a shows how the rolling shutter process can
traverses through the yellow coating, the combina- be leveraged to increase the data rate. For illustration
tion produces a white light. Different variations of purposes, we assume that the transmitter uses ON-
the white light (color temperatures) are produced OFF modulation. The transmitter can change its state
by modifying the thickness of the phosphor layer. (transmit the next symbol) in a time shorter than the time
2) RGB Combination: White light can also produced required to scan a row of pixels. As shown in Fig. 2a,
by proper mixing of red, green and blue light. In the transmitter is in ON state first which results in higher
this method, three separate LEDs are used which intensity output for pixels of the first column. At the next
increases the cost of LED luminaire compared to time instance, it changes its state by switching to OFF
using the Blue LED with Phosphor. state. This can be recorded as low intensity output for
pixels of the second column. Once all the columns are
Due to ease of implementation and lower cost, the
scanned, all the columns of the resultant image can be
first method with blue LED and phosphor is more
converted to binary data. It was shown in [16] that multi-
commonly used for designing white LED. However, in
kbps of throughput can be achieved using the rolling
terms of communication, the phosphor coating limits the
shutter process of camera sensor.
speed at which LED can switched to a few MHz. As
Note that image sensor can allow any mobile device
we will discuss in Section 3.2, various solutions have
with camera to receive visible light communication.
been proposed to alleviate this limitation. On the other
However, in its current form, it can only provide very
hand, RGB combination is preferable for communication
limited throughput (few kbps) due to its low sampling
as it also creates an opportunity of using Color Shift
rate. On the other hand, stand-alone photodetectors have
Keying to modulate the data using three different color
shown to achieve significantly higher throughput (hun-
wavelength LEDs.
dreds of mbps). In this survey, we assume the receiver
to be the photodetector unless otherwise mentioned
2.2 VLC Receiver specifically.
Two types of VLC receivers can be used to receive the
signal transmitted by an LED luminaire 2.3 VLC Modes of Communication
1) photodetector - also referred as photodiode or non- Visible light communication can be classified into two
imaging receiver, modes: (1) Infrastructure-to-device communication and
2) imaging sensor - also called a camera sensor. (2) Device-to-device communication. An indoor scenario
The photodetector is a semiconductor device that con- where LED luminaires are used to illuminate the room
verts the received light into current. The current com- is shown in Fig. 2b. In this case, the luminaires can

LED lights
Safety alarm

Thermostat

Transmitter
ON OFF ON OFF Television
LED state

Image
sensor
readout

(a) The rolling shutter effect observed when receiving data using an image sensor. (b) An example scenario showing that LEDs can
communicate to various devices including user’s
mobile devices and other smart devices;
reproduced using [15]

Fig. 2: The rolling shutter effect and typical usage scenario of an indoor VLC network

transmit data to various devices inside the room. The is possible to choose an LED with appropriate specifica-
LEDs can also coordinate between themselves to reduce tions and estimate its communication link performance.
the interference and even enable coordinated multi-point Note that the notations symbols used throughout this
transmission to receiving devices. The uplink transmis- section are listed in Table 2 with their meaning.
sion from the devices are difficult to achieve because
using LEDs on end-user devices can cause noticeable
3.1.1 Transmitted Power of an LED - Luminous Flux
disturbance to users. In such case, RF or infrared com-
munication can be used for the uplink transmissions. An LED transmitter serves dual purpose of illumination
Similar to the indoor case, the LEDs used in street lamps and communication. Therefore, it is necessary to first
as well as traffic lights can be used to provide internet establish an understanding of relevant photometric and
access to users in cars and pedestrians. We will discuss radiometric parameters. Using these parameters, we will
such vehicular application in Section 6.3. be able to calculate the Luminous Flux which is the
Due to omni-present camera sensor for mobile de- transmitted power of an LED transmitter. First, we will
vices, the visible light communication can also be used calculate the transmitted power, path loss and received
for near-field device-to-device communication. Here, the power of a Line-Of-Sight (LOS) link and then analyze
LED pixels on the display of one smartphone can be the multipath impact of reflected paths.
used to transmit data to the camera sensor of another Photometric parameters quantify the characteristics of
smartphone. With recent advances in design of efficient light (such as brightness, color etc.) as perceived by the
codes, such screen-to-camera streaming has been shown human eye. They are useful in understanding the illumi-
to achieve very high throughput. We discuss these tech- nation aspects of LEDs. Radiometric parameters measure
niques in Section 6.2. In another form of device-to-device the characteristics of radiant electromagnetic energy of
communication, cars and other vehicles on the road can light. They are useful in determining communication
communicate with each other to form an ad-hoc network related properties of LEDs. There are two ways of cal-
using VLC. culating the Luminous Flux - using spectral integral or
Although we discussed the vehicular networking and using spatial integral. Depending on which parameters
screen-camera communication, our primary focus in are available for a given LED transmitter, one of the two
this survey is towards design and analysis of indoor methods can be chosen for calculation of luminous flux.
infrastructure-to-device networking using visible light. Spectral Integral: The spectral integral method uses
luminosity function of human eye and spectral power
distribution of an LED to derive the luminous flux.
3 P HYSICAL L AYER Luminosity Function V (λ): The photopic vision of human
We start with a comprehensive overview of VLC physi- eye allows humans to distinguish different colors, mak-
cal layer by discussing (1) channel model and character- ing it a crucial factor in designing lighting technology
istics, (2) modulation methods and (3) MIMO techniques [17]. It was shown in [18] that human’s photopic vision
for VLC. exhibits different levels of sensitivity to different wave-
lengths of visible light spectrum. This aspect is shown
in Fig. 3 using the luminosity function V (λ). The function
3.1 Channel Model and Propagation Characteristics shows that human eye can see the colors within the
In this section, we describe the channel model for prop- range of 380 nm to 750 nm with the maximum sensitivity
agation of visible light. Based on the channel model, it at wavelength of 555 nm (the yellow-green region).

TABLE 2: Symbols and their meaning

Visible spectrum
Symbol Meaning 1

Human eye sensitivity V(λ)

λ Wavelength Maximum sensitivity
V (λ) Luminosity functiuon at 555 nm
ST (λ) Transmitter spectral power distribution function 0.1
FT Transmitter luminous flux
FR Receiver luminous flux
gt (θ) Luminous intensity distribution 0.01
I0 Axial intensity

Orange
θmax Half beam angle

Yellow
Green
Violet

Cyan
Ωmax Full beam angle 0.001

Blue

Red
LL Luminous path loss
LP Optical power path loss
D Distance between transmitter and receiver 0.0001
r Radius of the receiver aperture 300 400 500 600 700 800 900
α Incident angle Wavelength (λ) (nm)
β Irradiation angle
Ar Receiver aperture area Fig. 3: Luminosity function representing human eye’s
Ωr Receiver solid angle from transmitter sensitivity to different wavelengths in the visible
m Order of Lambertial emission
φ1/2 Semi-angle at half illuminance spectrum.
Rf (λ) Spectral responsivity function
P Ro Received optical power
SR (λ) Receiver spectral power distribution function 100 Warm white

Normalized Radiant Power (%)

λrL Lower wavelength cut-off for optical filter 90
λrH Higher wavelength cut-off for optical filter
80
PR (i) Received optical power from LOS link of ith LED
PR (total) Total received optical power 70
ρ(λ) Spectral reflectrance 60 Natural white
N Number of LED transmitters 50
k Number of bounces of light
h(t) Power delay profile 40
δ Dirac delta function 30 Cool white
c Speed of light 20
F OV Acceptance angle of receiver 10
(k)
Γn Power of reflected ray after kth bounce 0
σshot Standard deviation of shot noise 400 450 500 550 600 650 700 750
σthermal Standard deviation of thermal noise
x Number of photons collected in unit time Wavelength (λ) (nm)
κ Boltzmann’s constant
IB Photocurrent due to background noise Fig. 4: Power spectral distribution for LED of three
Gol Open-loop voltage gain color types - warm white, natural white and cool
Tk Absolute temperature white. Warm white and natural white have more
Cpd Capacitance of the photodetector per unit area
η FET channel noise factor radiated power for green-yellow-orange wavelengths
gm FET transconductance compared to cool white which provides a more bluish
I2 , I 3 Noise-bandwidth factors illumination; Figure reproduced from [19].

Spectral Power Distribution ST (λ): The ST (λ) of an LED the “perceived” power emitted by the LED. It weighs the
is the function representing the power of the LED at ST (λ) function with V (λ) (the sensitivity of human eye
all wavelengths in the visible light spectrum. The LED to different wavelengths) because we know from Fig. 3
vendors typically publish the distribution to explain that human eye does not respond to all wavelengths
how different colors will be rendered in the presence equally. The luminous flux of the transmitter LED (FT )
of the LED. It is a radiometric parameter measured is measured in lumens and it can be calculated as
in Watts/nm. The spectral power distribution of three 750
Z nm
different colored LEDs are shown in Fig. 4. It can be
FT = 683 (lumens/watt) ST (λ)V (λ)dλ (1)
observed that all three LEDs have high radiant power
at two wavelengths - blue and yellow. As described 380 nm

in Section 2.1, most current LEDs produce white light The constant 683 lumens/watt is the maximum luminous
by combining blue light emitted by a blue LED with efficiency. The luminous efficiency is the ratio of luminous
yellow phosphor coating. Depending on the desired flux to the radiant flux, which measures how well the
type of white color (warm, natural or cool), blue and radiated electromagnetic energy and required electricity
yellow light emissions are controlled using the phosphor of an LED was transformed to provide visible light
coating. For example, more yellow light is allowed in illumination. We know from Fig. 3 that human eye
warm and natural white compared to the cool white is most sensitive to detect the wavelength of 555 nm
LED. (green). The electrical power necessary to produce one
Luminous Flux: The luminous flux combines luminos- lumen of light at the wavelength of 555 nm is derived to
ity function and spectral power distribution to calculate be 1/683rd of a watt [20]. This means that for any other

color source, the power necessary to produce one lumen solid angle Ωmax . Different from Equ. (1) which was a
of light is always higher than 1/683rd of a watt. Hence, spectral integral, here the flux is calculated using spatial
the maximum luminous efficiency is 683 lumens/watt integral as below
which occurs at 555 nm wavelength. ΩZmax
Spatial Integral: Another way of calculating the lumi-
FT = I0 gt (θ)dΩ (3)
nous flux is to utilize LED’s spatial emission properties.
For this, we will use luminous intensity and axial inten- 0

sity as described next. where gt (θ) is the normalized spatial luminous intensity
distribution. Combining Equs. (2) and (3), we get
90
θZ
max
Relative Luminous Intensity (%)

=47o
max
FT = I0 2πgt (θ)sinθdθ (4)

Luminous intensity (candela)

LED 1
100 LED 2 300
0
80 494
45
600 3.1.2 Path Loss and Received Power
60
Based on the luminous flux calculated above, we will
40 900 now derive the value of path loss. It was proven in
I0= 987

20
candela [24] that the path loss in photometric domain (referred
-100 -80 -60 -40 -20 0 20 40 60 80 100 1200
as luminous path loss LL ) is the same the path loss
0
Angle Angle
in radiometric domain (referred as optical power path
(a) (b) loss LP ). This is due to the fact that in line-of-sight
Fig. 5: (a) Luminous intensity distribution for two LED free space propagation, the path loss can be assumed
- (1) Cree XLamp XP-E High-Efficiency White [19] (2) to be independent of the wavelength. Therefore, we can
Cree XLAMP XR-E [21] (b) Luminous intensity calculate LL using the luminous flux derived in the
distribution of Cree LMH6 in polar coordinates [22] previous section. Specifically, LL is the ratio of luminous
and its half-beam angle; Figures reproduced from [19], flux of the receiver (FR ) and the transmitter (FT ). FT can
[21], [22]. be calculated as Equ. (4).

Luminous Intensity gt (θ): While luminous flux mea- Transmitter

sures the total amount of light emitted by an LED, the

luminous intensity measures how bright the LED is in
Receiver
a specific direction. It is measured in Candela which is
luminous flux per unit solid angle (1 steradian). This max Ar

Transmitter
Distance (D)
allows us to understand where the LED directs its light.
r
Fig. 5 shows the luminous intensity distribution of three
different LEDs. In Fig. 5a, both the LEDs emit light
at wider angles allowing better illumination in many radius (r)
directions, while in Fig. 5b, it can be observed that LED
emits light in a narrower beam (much like spotlighting).
Most LED sources have Lambertial beam distribution Receive
r
[23] which means that the intensity drops as the cosine
of the incident angle. Fig. 6: Relative position of transmitter and receiver in
There are two important parameters to be derived LOS settings; reproduced from [24].
from the intensity distribution
In order to calculate the FR , it is necessary to specify
Axial Intensity (I0 ) is defined as the luminous intensity
the relative positions of the transmitter and the receiver.
in candelas at 0o solid angle. For LED in Fig. 5b, the axial
This relative positioning is shown in Fig. 6. Here, the
intensity is 987 candela. Typically, the luminous intensity
distance between the receiver and the transmitter is
distribution provided by the vendors are normalized
D, and radius of the receiver aperture is r. The angle
with the axial intensity as shown in Fig. 5a.
between the receiver normal and transmitter-receiver
Half Beam Angle (θmax ) is the angle at which the light
line is α (also referred as incident angle). The transmitter
intensity decreases to half of the axial intensity. For the
viewing angle is β (also referred as irradiation angle). Let
LED in Fig. 5b, the half beam angle is 47o . For the
the receiver solid angle as observed from the transmitter
Lambertian sources like LEDs, the half beam angle is
be Ωr and receiver’s area Ar as shown in Fig. 6, then
calculated from the entire beam angle (Ωmax ) as follows
Ar cos(α) = D2 Ωr (5)
Ωmax = 2π(1 − cos θmax ) (2)
From Fig. 6, the receiver flux FR can be calculated as
The luminous flux can now be calculated by integrat-
ing the luminous intensity function over the entire beam FR = I0 gt (β)Ωr (6)

The optical path loss LL can be calculated using Equa- angle (β). These three factors are independent of trans-
tions (4), (5) and (6) as mitter and receiver hardware, and depend on receiver’s
FR gt (β)Ar cosα movement and orientation. As an example, if the receiver
LL = = θmax
(7) is a smartphone equipped with a photodiode, the three
FT
factors will change based on user’s movement and de-
R
D2 2πgt (θ)sinθdθ
0 vice orientation. It is crucial to understand the impact of
Most LED sources have Lambertial beam distribution these factors on received power in order to evaluate the
which means that the spatial luminous intensity distri- achievable capacity. Authors in [26] studied the impact
bution is a cosine function using a smartphone photodiode as the receiver. Fig. 8
show how the normalized received power (measured
gt (θ) = cosm (θ) (8) as light intensity on smartphone photodiode) varies
where m is the order of Lambertial emission. The value with changes in D, α and β. Fig. 8a shows how the
of m depends on the semi-angle at half illuminance Φ1/2 received power attenuates with D as inverse square low
of the LED (Equ. (10)). The incident angle measures the changes
ln(2) in smartphone’s orientation (0o means photodiode is
m= (9)
ln(cosΦ1/2 ) directly facing the LED). As the incident angle (α) in-
creases, the energy at which the photons strike the pho-
Substituting Equ. (8) and θmax in Equ. (7), we get the
todiode decreases, which in turn results in decrease of
path loss value for a Lambertian LED source as follows
received power. Similarly, the received power decreases
(m + 1)Ar with increase in the irradiation angle (β) confirming
cos α cosm (β)
LL = (10)
2πD2 the lambertian emission pattern of the LED. The impact
If the LED emission can not be modeled using the of these three factors have important implications on
Lambertian cosine function, it is necessary to measure guaranteeing high SNR in VLC access networks and
gt (θ) for the given LED, and use it to calculate LL from managing inter-cell interference as we will discuss in
Equ. (7). Section 4.2.
The received optical power can be now calculated
using the path loss. It is typical that the receiving pho- 3.1.3 Multipath Propagation with Reflected Paths
todetector is equipped with an optical filter. Let Rf (λ) As we saw in Section 2, typically there are more than one
denote the spectral response of the optical filter. Fig. 7 LED in a luminaire. The receiving photodetector can si-
shows Rf (λ) of a typical photodetector. Using Rf (λ), the multaneously receive (intensity modulated) signals from
multiple LEDs as shown in Fig. 2b. The received optical
0.6 power of the receiver can be calculated by summing the
received power of each LOS link within receiver’s field-
0.5
Responsivity (A/W)

of-view (FOV) can be expressed as

0.4
N
X
0.3 PR (total) = PR (i) (12)
0.2
i=0

0.1
where N is the total number of LEDs and PR (i) is
300 400 500 600 700 800 900 1000 1100 1200 the received optical power from LOS link of ith LED
Wavelength (nm)
calculated from Equ. (11).
Fig. 7: Spectral response of a typical photodetector Since the majority of the indoor surfaces are more or
receiver; responsivity (measured in A/W) is the ratio of less reflective of visible light, it is necessary to under-
output photocurrent in amperes to incident radiant stand the impact of reflected paths on the performance
energy in watts; reproduced from [25]. of communication. Spectral reflectance (ρ(λ)) represents
reflectivity of a surface (such as wall, ceiling etc.) as
received optical power PRO for the direct line-of-sight a function of wavelength. It was noted in [27] that
optical link can be calculated as reflectivity of Infrared signal is higher compared to the
λ
ZrH visible band. The spectral reflectance of commonly used
PRO = SR (λ)Rf (λ)dλ (11) building materials like plaster wall, ceiling etc. was
measured in [27] using a spectrophotometer. Fig. 9 shows
λrL
the results of measured reflectivity. It can be observed
where SR (λ) = LP ST (λ) = LL ST (λ) and λrL and λrH that plastic wall has the least reflectivity while the plaster
are lower and upper wavelength cut-off values for the wall has the highest reflectivity.
optical filter respectively. Because of the reflections, the receiver receives signal
Considering Equations (10) and (11), the received from many different paths. Such multipath propagation
power is dependent on three factors - the transmitter- can be characterized using Power Delay Profile (PDP).
receiver distance (D), incident angle (α) and irradiation The PDP gives the distribution of received power as a

Normalized received power

1
Incident angle = 0 Distance = 4 m Distance = 4 m
0.96 Irradiance angle = 0 Irradiance angle = 0 Incident angle = 0
0.8
0.76 1
0.57 0.8
0.6
0.6
0.38
0.4 0.4
0.19 0.2
0 0 0.2
1 2 3 4 5 6 90 75 60 45 30 15 0 0 15 30 45 60 75
Distance - D (m) Incident angle Irradiance angle
(a) (b) (c)

Fig. 8: Impact of (a) transmitter-receiver distance, (b) incident angle (α) and (c) irradiation angle (β) on the
received power; reproduced from [26]

1 N LEDs at time instance t as

Plastic wall N X ∞
Ceiling X
0.8 Floor h(t) = h(k) (t; Sn ) (13)
Spectral reflectance

Plaster wall
n=1 k=0
0.6
where Sn is the spectral power distribution of nth LED
0.4 and k is the number of bounces. When k = 0, the
resultant PDP [27] is that of an LOS path as
0.2
α0 D0
0
h(0) (t; Sn ) = L0 Pn rect δ t− (14)
350 400 450 500 550 600 650 700 750
F OV c
Wavelength (nm) where L0 = LL is the path loss for the LOS case (derived
Fig. 9: Different indoor surfaces exhibit different levels in Equ. (10)), δ is a dirac delta function, D0 is the distance
of spectral reflectance depending on the wavelength; between the LED and the receiver and c is the speed
reproduced from [27]. of light. Because the photodiode can only detect the
light whose angle of incidence is smaller than its FOV, a
rectangular function [27] is used where
(
function of propagation delay. A non-LOS signal can 1 for |x| ≤ 1
rect(x) =
be bounced from many surfaces before it reaches the 0 for |x| > 1
receiver photodetector as shown in Fig. 10. Authors in This means that when if a ray does not reach within the
FOV of the receiver after k bounces, its effect on the total
Transmitter received power is considered 0.
When k ≥ 1, the PDP after k bounces (refer Fig. 10)
for the nth LED can be calculated [27] as
D1
Z
1 k (k) (k) α0
k-th h (t; Sn ) = L1 L2 · · · Lk+1 Γn rect × (15)
0 k+1
bounce F OV
First 1 s∈S

bounce 2 D0 D1 + D2 + · · · + Dk+1
δ t− dAs (16)
D k+1 c
D2 0
where
2
k+1
FOV As (m + 1) cos α1 cosm β1
L1 = (17)
2πD1 2
For the path loss of the first bounce L1 , the ray origi-
Receiver
nated from the LED which we have previously modeled
Fig. 10: A non-LOS signal can bounce off the surfaces as a Lambertian emitter (Equ. (8)). For the remaining
many times before reaching the receiver; β and α bounces, we can calculate the path loss of each path as
denote the angle of irradiation and incident
As cos β2 cos α2
respectively; reproduced from [27]. L2 = (18)
πD2 2
AR cos βk+1 cos αk+1
[27] modeled the PDP of multiple bounces for a total of Lk+1 = (19)
πDk+1 2

The integration in Equ. (16) for each surface s of all pass filter at the receiver. Most of the previous studies
reflectors S where As is the area of the surface. For Lk+1 , assume that this ambient noise floor remains stationary
(k)
AR is the area of the photodiode receiver. Γn is power over space and time, however, no systematic evaluation
th is present in the literature. For example, the indoor solar
of the reflected ray after k bounce. It is calculated [27]
as Z radiation changes at different places depending on win-
(k)
Γn = Sn (λ)ρ1 (λ)ρ2 (λ) . . . ρk (λ)dλ (20) dows and doors. The radiation also changes depending
on the time of the day (and year) and orientation of
λ
the windows/doors. Radiation from other illumination
where ρk (λ) is the spectral reflectance of the surface of sources will also remain an unavoidable source of noise
k th bounce. until we completely transition to LED technology. It
is required that exhaustive indoor measurements are
10-5 carried out to accurately account for such noise.
4
Plaster wall Once the noise due to solar radiation and artificial
Received power (W)

3.5
Plastic wall illumination sources is filtered, the SNR at the receiver
3 can be calculated based on the shot noise and the thermal
2.5 noise of the photodetector circuitry as
2
1.5 PR E 2
SN R = (21)
(σshot )2 + (σthermal )2
1
0.5 where σshot and σthermal are the standard deviation of
0 shot noise and thermal noise respectively. The shot noise
0 5 10 15 20 25 30 is due to inherent statistical fluctuation in the amount
Time (ns) of photons collected by the photodetector. It is known
Fig. 11: Power delay profile for 4 LED transmitters in a that the photon counting follows a poisson distribution
cubic room with plaster or plastic walls; reproduced which means that if the mean of number of photons
from [27]. collected by the photodetector in a unit time is x, then
the√standard deviation of number of photons collected
Fig. 11 shows the power delay profile in a realistic is x. This also results in poisson distributed variation
scenario where four LED luminaires are deployed in in photoelectrons generated by the photodetector. Based
a square topology on a ceiling of a cubic room with on this, the variance of shot noise can be calculated [23],
either plaster or plastic walls [27]. It can be observed [31] as below
that the first peak is due to the direct received signal (σshot )2 = 2qPRE B + 2qIB I2 B (22)
(LOS) from the LED. The other peaks are due to multiple
reflections from the wall as calculated using Equ. (16). The variance of thermal noise [23], [31] is
As expected, the received power due to reflection multi- 8πκTk 16π 2 κTk η 2 2
path is relatively lesser compared to the LOS power. (σthermal )2 = Cpd AI2 B 2 + Cpd A I3 B 3
Gol gm
Most of the power delay profiling [27]–[30] of visible (23)
light communication rely on simulations. However, de- where B Hz is the bandwidth of the photodetector, κ
tailed measurement-based studies in realistic scenarios is the Boltzmann’s constant, IB is the photocurrent due
(such as indoor places with many different reflecting to background radiation, Gol is the open-loop voltage
objects, different LED arrangements etc.) are necessary gain, Tk is the absolute temperature, Cpd is capacitance
for improved understanding of multi-path in VLC and of the photodetector per unit area, η is the FET channel
developing the techniques to combat it. noise factor, gm is the FET transconductance, and I2 and
I3 are the noise-bandwidth factors with values 0.562
3.1.4 Receiver Noise and SNR and 0.0868 respectively. Shot noise and thermal noise
There are three major sources of noise in indoor visible are dependent on the area of the photodetector, and
light optical link (1) ambient light noise due to solar depending on factors such as room temperature, ambient
radiation from windows, doors etc. and noise due to light etc. either of them can dominate the overall noise
other illumination sources such as incandescent and [23] observed by the VLC receiver.
fluorescent lamps, (2) shot noise induced in the pho-
todetector by the signal and the ambient light and (3) 3.1.5 Shadowing
electrical pre-amplifier noise (also known as thermal The receiver of a visible light communication link can be
noise) of the photodetector. shadowed by different objects or humans in the indoor
The ambient noise of solar radiation and artificial environment. For example, if a receiver photodiode is
illumination sources such as lamps results in ambient positioned on a desk, it is possible that movement of
noise floor which is a DC interference. The effect of the nearby chair can result in shadowing of the receiver.
such noise can be mitigated by using a electrical high Similarly, if a human passes by frequently between the

transmitter and the receiver, the link performance is 100

affected by the frequent shadowing. Authors in [32] 90
studied such case of human mobility using simula- 80

Measured Light (%)

tions and suggested that in multiple spatially sepa- 70
rated LED sources should be used in order to mitigate 60
the frequent disconnections due to human shadowing. 50
Apart from this preliminary work, shadowing in indoor 40
VLC networks is not studied in literature. Given that 30
visible light exhibits significantly different propagation 20
characteristics compared to RF (such as no penetration 10
through walls etc.), it is crucial to characterize and model 0
0 10 20 30 40 50 60 70 80 90 100
visible light shadowing in indoor environment. This
understanding can also provide insights on deployment Perceived Light (%)
aspects of indoor VLC networks and how they should Fig. 12: Human eye perceives the actual measured light
be different than current deployment of LEDs which are differently due to enlargement/contraction of pupil.
primarily used for illumination purposes.

measured light as
3.2 Modulation Methods r
Measured light(%)
Perceived light(%) = 100 × (24)
With the understanding of path-loss, noise and SNR, 100
we now discuss various modulation methods used in This means that a lamp that is dimmed 1% of its mea-
VLC. The most striking difference between VLC and sured light is perceived to be 10% dimmed by the human
RF is that in VLC, the data can not be encoded in eye. This is important in terms of VLC because a user
phase or amplitude of the light signal [10]. This means may choose an arbitrary level of dimming depending
that phase and amplitude modulation techniques can on the application or desired energy savings, but the
not be applied in VLC and the information has to be communication should not be affected by the dimming.
encoded in the varying intensity of the emitting light In other words, the data should be modulated in such a
wave. The demodulation depends on direct detection way that any desired level of dimming is supported.
at the receiver. These set of modulation techniques are (2) Flicker mitigation: An additional requirement for
referred as IM/DD (Intensity Modulated/Direct Detec- any VLC modulation scheme is that it should not result
tion) modulations. In this section, we will discuss the in human-perceivable fluctuations in the brightness of
IM/DD modulation techniques used for visible light the light. It was shown in [34] that flickering can cause
communication. serious detrimental physiological changes in humans.
Different from other types of communications, any For this reason, it is necessary that changes in the light
modulation scheme for VLC should not only achieve intensity should happen at a rate faster than human eye
higher data rate but should also meet the requirements can perceive. IEEE 802.15.7 standard [8] suggests that
of perceived light to humans. These requirements about flickering (or change in light intensity) should be faster
perceived light can be characterized by following two than 200 Hz to avoid any harmful effects. This means
properties - that any modulation scheme for VLC should mitigate
(1) Dimming: It was suggested in [17] that different flickering while providing higher data rate.
levels of illuminance is required when performing differ- The most common cause of flickering is long runs of 0s
ent types of activities. As an example, an illuminance in or 1s which can reduce the rate at which light intensity
the range of 30-100 lux is often enough for simple visual changes and cause the flickering effect. Run Length
tasks performed in most public places. On the other Limited (RLL) codes are used to mitigate long runs of
hand, office or residential applications require higher 0s or 1s. RLL codes ensure that the output symbols have
level of illuminance in the range of 300-1000 lux. With balanced repetition of 0s and 1s. Examples of commonly
the advancements in LED driver circuits, it has become used RLL codes include Manchester, 4B6B and 8B10B
possible to dim an LED to an arbitrary level depending coding. In Manchester coding, a “0” is replaced with
on the application requirement to save energy. a “down” transition (“10”) and “1” is replaced with
If an LED can be dimmed to an arbitrary level, it is an “up” transition (“01”). 4B6B coding maps a 4 bits
also necessary to understand its impact on the human symbol to a 6 bits symbol that has balanced repetition.
perceived light. It was first shown in [33] that the relation Similarly, 8B10B maps a 8 bits symbol to 10 bits symbol.
between the measured light and the perceived light is The number of additional bits added is the highest in
non-linear. This property is shown in Fig. 12. In other the Manchester coding making it a suitable choice for
words, a human eye adapts to lower illumination by low data rate services that require better balancing. On
enlarging the pupil to allow more light to enter the the other hand, 8B10B reduces the number of additional
eye. The perceived light can be calculated [33] from the bits added (high data rate), however, it performs poorly

in terms of the DC balancing. as ON/OFF levels causes the LEDs to be operated

We next discuss four types of modulation schemes at lower driving currents which in turn has shown
used in VLC (1) On-Off Keying, (2) Pulse modulation, to incur changes in color rendering (change in
(3) Orthogonal Frequency Division Modulation (OFDM) emitted color of LEDs) [41].
and (4) Color Shift Modulation (CSK). We describe each 2) Compensation periods: In this solution, the ON and
of them along with a discussion on how they provide the OFF levels of the modulation remain the same
the dimming support. but additional compensation periods are added
when the LED source is fully turned on (called
3.2.1 On-Off Keying (OOK) ON periods) or off (OFF periods). The duration
In OOK, the data bits 1 and 0 are transmitted by turning of the compensation periods is determined based
the LED on and off respectively. In the OFF state, the on the desired level of dimming. Specifically, ON
LED is not completely turned off but rather the light periods are added if the desired level of dimming
intensity is reduced. The advantages of OOK include is more than 50% and OFF periods are added if the
its simplicity and ease of implementation. OOK-like desired level of dimming is less than 50%. Authors
modulation is widely used in wireline communication. in [42] proposed a way to calculate the percentage
Most of the early work on using OOK modulation for time of active data transmission (γ) within the
VLC utilize while LED. As we discussed in Section 2, transmission interval T to obtain a dimming level
such LED produces white light by combining the blue of D as
emitter with yellow phosphor. The major limitation of (
(2 − 2D) × 100 : D > 0.5
the white LED is its limited bandwidth (few megahertz γ= (25)
[35]) due to slow time response of the yellow phos- 2D × 100 : D ≤ 0.5
phor. It was first proposed by [36] to use NRZ (Non-
Return-to-Zero) OOK with the white LED and a data When the desired dimming level is D with OOK,
rate of 10 Mbps was demonstrated over a VLC link. the maximum communication efficiency ED can be
To further improve the performance, [35] used a blue calculated [42] using information theoretic entropy
filter to remove the slow-responding yellow component, as
resulting in a datarate of 40 Mbps. Similarly, [37] and [38] ED = −D log2 D − (1 − D) log2 (1 − D) (26)
proposed to combine the blue-filtering with analogue
equalization at the receiver to achieve data rates of This means that communication efficiency is a
100 Mbps and 125 Mbps respectively. Authors in [39] triangular function of the dimming level with max-
showed that the performance can be further improved imum efficiency at dimming level of 50%. The
by using an avalanche photodiode as the receiver instead efficiency drops linearly when dimming level de-
of the P-I-N photodiode. The achievable data rate with creases to 0% or increases to 100%. The dimming
avalanche photodiode and NRZ-OOK was shown to support using compensation periods reduces the
be 230 Mbps. Newly available white LEDs combine data rate, however, since the modulated ON/OFF
the RGB frequencies to produce the white light. The signals have unchanged intensity, the communi-
advantage of such LEDs is that they do not have the cation range remains unchanged. To address the
slow-responding yellow phosphor layer. However, such problem of lower data rate with compensation
RGB white LEDs require three separate driver circuits periods, [43] proposed to use inverse source coding
to realize the white light. A different approach was to maintain the high data rate while achieving the
presented in [40] where RGB white LED was used but desired level of dimming.
only the red LED is modulated for data transmission
while the other two are provided constant current for 3.2.2 Pulse Modulation Methods
illumination. The proposed system can achieve a data Although OOK provides various advantages such as
rate of 477 Mbps with simple NRZ-OOK modulation and simplicity and ease of implementation, a major limita-
a P-I-N photodiode receiver. tion is its lower data rates especially when supporting
There are two ways proposed in the Standard IEEE different dimming levels. This has motivated the design
802.15.7 [8] to provide the dimming support when using of alternative modulation schemes based on pulse width
OOK as the modulation scheme: and position which are described next.
1) Redefine ON and OFF levels: To achieve the desired Pulse Width Modulation (PWM): An efficient way to
level of dimming, the ON and the OFF levels can be achieving modulation and dimming is through the use
assigned different light intensities. The advantage PWM. In PWM, the widths of the pulses are adjusted
of this scheme is that required level of dimming can based on the desired level of dimming while the pulses
be obtained without any additional communication themselves carry the modulated signal in the form of a
overhead. It can retain the data rate achievable by square wave. The modulated signal is transmitted dur-
NRZ-OOK modulation, however, the communica- ing the pulse, and the LED operates at the full brightness
tion range decreases at lower dimming levels. One during the pulse. The data rate of the modulated signal
major disadvantage is that using lower intensities should be adjusted based on the dimming requirement.

0 1
Bitstream
Mapping bits to DMT x(t) y(t)=x(t)p(t) PWM
QAM symbols modulation
LED
p(t) 0 1
PPM
PWM

50% dimming DMT+PWM 0 1

with PWM signal
TPWM VPPM

TON p(t) y(t)=x(t)p(t) S1 S2 S3 S4

OPPM

t t S1 S2 S3
...
MPPM
Fig. 13: Transmitter block diagram of DMT transmitter
with dimming control (top); An example of how 50%
PWM-controlled dimming signal can be combined with Fig. 14: Schematic diagram showing difference between
a DMT signal as proposed in [44] (bottom); Figures Pulse Width Modulation (PWM), Pulse Position
reproduced from [44] Modulation (PPM), Variable Pulse Position Modulation
(VPPM), Overlapping Pulse Position Modulation
(VPPM) and Multipulse Pulse Position Modulation
Authors in [45] showed that any dimming level from (MPPM); Sn refers to nth symbol.
0% to 100% can be obtained with high PWM frequency.
One benefit of PWM is that it achieves the dimming
without changing the intensity level of pulses, hence it of poor channel conditions. Authors in [50] designed
does not incur the color shift (like OOK with redefined a rate-variable punctured convolutional coded PPM for
ON/OFF levels) in the LED. The limitation of PWM infrared communication. Such a scheme adapts the mod-
is its limited data rate (4.8 kbps in [45]). To overcome ulation order of PPM and the code rate of punctured
this limitation, [44] proposed to combine PWM with convolutional codes based on the channel conditions.
Discrete Multitone (DMT) for joint dimming control and For even worse channel conditions, [51] proposed to
communication. The approach decouples the dimming use rate adaptive PPM transmission with both repeated
based on PWM and communication based on DMT on and punctured convolutional codes to achieve higher bit
the transmitter side. As shown in Fig. 13, the bitstream rate.
is divided and mapped to symbols using Quadrature Due to the limitations of lower spectral efficiency and
Amplitude Modulation (QAM). These QAM symbols data rate of PPM (only one pulse per symbol duration),
are transmitted on different DMT subcarriers that are other variants of pulse position-based modulation have
spaced by 1/T in frequency where T is the duration been proposed over time. A generalization of PPM is
of one symbol. The DMT signal x(t) is combined with referred as Overlapping PPM (OPPM) which allows
PWM square wave signal p(t) where the duty cycle is more than one pulse to be transmitted during the symbol
dependent on desired level of dimming. The resultant duration [48] and the different pulse symbols can be
signal y(t) = x(t)p(t) is shown in Fig. 13. It was also overlapping (see Fig. 14). [52] showed that OPPM can
shown that dimming constraint limits the achievable not only achieve a higher spectral efficiency compared
throughput due to high Bit Error Rate (BER). Authors to PPM and OOK but a wide range of dimming levels
in [46] also used QAM on DMT subcarriers to achieve a can be obtained along with the high data rate. Another
link rate of 513 Mbps, however, it does not address the generalization of PPM was proposed by [53] which is
issue of LED dimming. a scheme referred as Multipulse PPM (MPPM). Like
Pulse Position Modulation (PPM): Another pulse OPPM, it allows multiple pulses to be transmitted dur-
modulation method in visible light communication is ing the symbol duration, however, the pulses within a
based on the pulse position. In PPM, the symbol du- symbol duration do not have to be continuous (Fig. 14).
ration is divided into t slots of equal duration, and It was shown in [48] that MPPM can achieve a higher
a pulse is transmitted in one of the t slots. The po- spectral efficiency compared to OPPM.
sition of the pulse identifies the transmitted symbol. Authors in [54] proposed a variation of PPM that com-
Due to its simplicity, many early designs [47], [48] of bines OPPM and MPPM in a scheme called Overlapping
optical wireless systems adapted PPM for modulation. MPPM (OMPPM). In OMPPM, more than one pulse
In some of the early works of using PPM for infrared positions are allowed for each optical pulse. It shows
communication, authors in [49] proposed the use of rate that OMPPM can improve the spectral efficiency of
adaptive transmission scheme where repetition coding is MPPM without the expansion of bandwidth in noiseless
applied to gracefully reduce the throughput in presence photon counting channel. Further performance analysis

for noisy channels was presented in [55]. It was shown in OFDM generates complex-valued bipolar signals which
[56] that OMPPM with fewer pulse slots and more pulses need to converted to real-valued signals. This can be
per symbol duration has better cutoff rate performance. achieved by enforcing Hermitian symmetry constraint
Moreover, Trellis-coded OMPPM was studied in [57], on the sub-carriers and then converting the time-domain
[58] to show its effectiveness in direct detection channels signals to unipolar signals.
with background noise. In another set of modulation Depending on how the bipolar signals are converted
scheme, Differential PPM (DPPM) was proposed in [59]. to unipolar, there are two types of OFDM techniques: (1)
DPPM is similar to PPM except that the OFF symbols Asymmetrically-Clipped Optical OFDM (ACO-OFDM)
after the pulse in a PPM symbol are deleted and the and (2) DC-biased Optical OFDM (DCO-OFDM). In
next symbol starts right after the pulse of the previous ACO-OFDM, only odd subcarriers are modulated [66]
symbol. It was shown in [60] that DPPM requires signif- which automatically leads to symmetric time domain
icantly less average power than PPM for a given band- signal. While in DCO-OFDM [65], [67], [68], all sub-
width in an optical communication channel. Authors in carriers are modulated but a positive direct current is
[61] proposed Differential Overlapping PPM (DOPPM) added to make the signal unipolar. [69] presented a
where differential deletion of OFF symbols is applied comparison of both the OFDM schemes and showed
to OPPM, and showed that it achieves better spectral that LED clipping distortion is more significant in DCO-
efficiency and cutoff performance than PPM, DPPM and OFDM compared to ACO-OFDM. The biggest challenge
OPPM. in OFDM VLC system is the non-linearity of LED [70]
Authors in [62] proposed EPPM (Expurgated PPM) which is that the relationship between the current and
where symbols in the MPPM are expurgated to maxi- the emitted light of the LED is non-linear. This especially
mize the inter-symbol distance. EPPM achieves the same affects the OFDM-based VLC systems which have higher
spectral efficiency as PPM, however, it can be used in Peak-to-Average Power Ratio (PAPR). The effect of this
VLC to provide dimming support as it can achieve non-linearity was studied in [71], [72] and a solution
arbitrary level of illumination by changing the number was proposed to combat it by operating the LED in a
of pulses per symbol (code-weight) and the length of the small range where the driving current and optical power
symbol (code-length) [63]. With many pulses in a sym- are quasi-linear. Apart from the non-linearity, there is
bol, EPPM can also mitigate the flickering as compared only a limited support for dimming [73] in OFDM-based
to PPM. MEPPM (Multi-level EPPM) [64] extends the modulation schemes. Despite these challenges, OFDM
EPPM design with support to multiple amplitude levels for VLC holds great potential with achievable link rates
in order to increase the constellation size and spectral in the scale of multiple gbps [74], [75] using only single
efficiency. MEPPM can also support the dimming and LED.
provides flicker-free communication. IEEE 802.15.7 [8]
standard proposes a pulse modulation scheme called 3.2.4 Color Shift Keying (CSK)
Variable PPM (VPPM) which is a hybrid of PPM and To overcome the lower data rate and limited dim-
PWM. In VPPM, the bits are encoded by choosing dif- ming support issues of other modulation schemes, IEEE
ferent position of pulse as in PPM, however, the width of 802.15.7 standard [8] proposed CSK modulation which
the pulse can also be modified as needed. VPPM retains is specifically designed for visible light communication.
the simplicity and robustness of PPM while allowing CSK has attracted increasing amount of attention from
different dimming levels by altering the pulse width. research community in last couple of years [76]–[81].
As we discussed before, generating white light using
3.2.3 Orthogonal Frequency Division Multiplexing blue LED and yellow phosphorus slows down the fast
(OFDM) switching ability of LED and hinders high data rate
One limitation of previously discussed single-carrier communication. An alternative way to generating white
modulation schemes is that they suffer from high inter- light which is recently becoming more and more popular
symbol interference due to non-linear frequency re- is to utilize three separate LEDs - Red, Green and Blue
sponse of visible light communication channels. OFDM (RGB). This combined source with RGB LEDs is often
has been widely adopted in the RF communication referred as TriLED (TLED). CSK modulates the signal
due to its ability to effectively combat the inter-symbol using the intensity of the three colors in the TLED source.
interference and multipath fading. Authors of [65] first CSK modulation relies on the color space chromaticity
proposed the use of OFDM for visible light communi- diagram as defined by CIE 1931 [18] (see Fig. 15). The
cation. In OFDM, the channel is divided into multiple chromaticity diagram maps all colors perceivable by
orthogonal subcarriers and the data is sent in parallel human eye to two chromaticity parameters - x and y. The
sub-streams modulated over the subcarriers. OFDM for entire human visible wavelength is divided into seven
VLC can reduce the inter-symbol interference and does bands as shown in Table 3 and their centers are marked
not require complex equalizer, however, there are multi- in Fig. 15. Based on the diagram, the CSK modulation
ple challenges in realizing its implementation. First, the [8], [81] is performed as follows
OFDM technique for RF needs to be adapted for appli- 1) Determine RGB constellation triangle: The con-
cation in IM/DD systems such as VLC. This is because stellation triangle is decided based on the cen-

001
0.8 4CSK Symbols 0.8 8CSK Symbols 0.8 16CSK Symbols
0.7 0.7 0.7
010
0.6 0.6 0.6
Constellation
triangle 0.5 0.5 0.5

y
011 0.4 0.4 0.4
0.3 0.3 0.3
100
0.2 0.2 0.2
101
110 0.1 0.1 0.1
0 0 0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
000 x x x

(a) (b) (c) (d)

Fig. 16: (a) RGB constellation triangle (110, 010, 000) (b-d) Symbols of 4-CSK, 8-CSK and 16-CSK.

Band i Band j Band k

1 110 010 000
001 2 110 001 000
3 101 010 000
010 4 101 001 000
5 100 010 000
6 100 001 000
7 011 010 000
011 8 011 001 000
9 010 001 000
100 TABLE 4: Valid color band combinations that can be
101
110 chosen for building the constellation triangle for CSK.

000 be represented using the symbols for 4CSK, 8CSK

and 16CSK. Determining the position of the sym-
Fig. 15: CIE 1931 Chromaticity Diagram; The seven bols in the constellation design requires solving an
color codes correspond to the centers of seven bands optimization problem where the distance between
dividing the visible spectrum as shown in Table 3; the symbols should be maximized to minimize
reproduced from [82]. the inter-symbol interference. Note that there is an
additional constraint in the problem which ensures
Band (nm) Code Center (nm) (x, y) that the symbols should be equally distributed in
380-478 000 429 (0.169, 0.007)
478-540 001 509 (0.011, 0.733)
the triangle so that the combined light emitted
540-588 010 564 (0.402, 0.597) when transmitting different symbols is perceived
588-633 011 611 (0.669, 0.331) by the human eye to be white light only. The
633-679 100 656 (0.729, 0.271) optimization problem has been studied in [76]–[79]
679-726 101 703 (0.734, 0.265)
726-780 110 753 (0.734, 0.265) as we discuss next. Once the symbol coordinates
are decided, each symbol is assigned a bit sequence
TABLE 3: The seven bands used in CSK and their code, (e.g. in 4CSK, the 4 symbols are assigned 00, 01, 10
center and chromaticity coordinates. and 11 respectively), which is then used to map the
incoming bits to the symbols.
3) Determine the intensities of RGB LEDs: The
ter wavelength of the three RGB LEDs used in symbols are transmitted by varying the intensities
the TLED source. Table 4 shows the valid color of the RGB LEDs. The individual intensities of
band combinations as proposed by [81] that can the three LEDs (Pi , Pj and Pk ) for each symbol is
be chosen as the constellation triangle depending calculated by solving the following equations:
on the central wavelength of the RGB LEDs. For
the purpose of illustration, let us assume that we x s = Pi x i + Pj x j + P k x k (27)
choose the CSK constellation triangle to be (110, ys = Pi yi + Pj yj + Pk yk (28)
010, 000) as shown in Fig. 16a (example adapted Pi + Pj + PK = 1 (29)
from [81]).
2) Mapping data bits to chromaticity values: De- where xs and ys are the chromaticity values of the
pending on 4CSK, 8CSK or 16CSK being used, symbol (Fig. 16), and (xi , yi ), (xj , yj ) and (xk , yk ) are
the chromaticity values of symbols can be derived the chromaticity values of the central wavelength of the
from the constellation triangle. For our example, RGB LEDs being used (three points of the constellation
Figs. 16b, 16c and 16d show how data bits can triangle). The receiver uses the R, G and B intensities to

decode the transmitted signal. data rates from 1.25 Mbps upto 96 Mbps. PHY III can
Dimming support in CSK is simply amplitude dim- yield data rates starting from 12 Mbps upto 96 Mbps.
ming where the driving current of the LEDs is varied Further details of physical layer of IEEE 802.15.7 are
to change the brightness of resultant white light. Also, provided in [80].
different from OOK and pulse modulations, flickering Table 8 provides a comparison between four major
is not a problem with CSK since no amplitude varia- modulations schemes proposed for VLC. It can be ob-
tion is employed. Due to these advantages, researchers served that OFDM and CSK are more suitable for high
have recently attempted to improve the CSK scheme of data-rate applications in VLC access networks. As we
IEEE 802.15.7 by designing its generalized forms with will discuss next, OFDM is also more suitable for VLC
arbitrary constellation. Authors in [76] presented a CSK MIMO design, however, more research is necessary to
constellation design technique based on Billards equiv- ensure dimming support in OFDM. Another advantage
alent disk packing algorithm. Similarly, [77] and [78] of CSK is that it can provide multi-user access through
developed similar techniques with the use of different wavelength multiplexing as we will discuss in Section 4.
optimization algorithms such as interior point methods. Increasing demand of higher data-rates is likely to drive
All the constellation design techniques are designed to further research and development of OFDM and CSK
meet the color balance requirement where the TLED for VLC-based access networks.
source is required to produce any desired color for illu-
mination. The use of four LEDs (blue, cyan, yellow and
3.3 Multiple Input Multiple Output (MIMO)
red) was suggested in [79]. With four LEDs, it is possible
to achieve a quadrilateral constellation shape that allows In order to provide sufficient illumination, most of the
QAM-like constellation design. The presented system is luminaires typically contain multiple LEDs. These mul-
shown to be more energy efficient as well as reliable (less tiple LEDs can be treated as multiple transmitters that
inter-symbol noise) compared to the conventional CSK can enable visible light MIMO communication. In RF
with 3 LEDs. communications, MIMO systems are commonly used
The RGB tri-LED can also be used to implement Wave- (in IEEE 802.11n, Long-Term Evolution - LTE) to obtain
length Division Multiplexing (WDM) - a multiplexing higher data rates. Similarly, multiple LEDs can be used
technique commonly used in fiber optics communica- for higher spectral efficiency in VLC. Note that there are
tion. Authors in [83] proposed modulating separate data certain similarities between the VLC MIMO systems dis-
streams on three colors which together multiplex to cussed in this section and screen-camera links (discussed
white light. With the use of DMT, an aggregate data rate in Section 6.2) as both of them can use an image sensor
of 803 Mbps was shown to be achievable using single as a MIMO receiver. The difference is that unlike smart-
RGB LED in [83]. Authors in [84] proposed the use of phone screens, the multiple LED transmitters considered
carrier-less amplitude and phase modulation on WDM here are also used for the illumination. We will provide
VLC system with RGB LED to achieve a data rate of 3.22 further details of the screen-camera links in Section 6.2.
Gbps. MIMO systems in VLC are difficult to realize com-
IEEE 802.15.7 Physical Layer IEEE 802.15.7 [8] stan- pared to RF communications. In RF MIMO systems, the
dard has specified three PHY layers for VLC with a throughput gains are largely attributed to spatial diver-
total of 30 MCS (Modulation and Coding Scheme) in- sity (existence of multiple spatial paths that are diverse
dexes. These MCS levels are shown in Tables 5, 6 and in nature). However, such diversity gains are limited
7. Both PHY I and PHY II utilize OOK and VPPM in VLC MIMO because paths between the transmitter
for modulation. PHY I utilizes Reed Solomon (RS) and and receiver are very similar (less diverse) especially
Convolutional Codes (CC) for Forward Error Correction in indoor scenarios. This limits the available spatial
(FEC), while PHY II and III mostly reply for RS codes diversity of VLC MIMO systems. The other challenge in
only for FEC. VLC MIMO is the design of the receiver as we discuss
As described in [80], “optical clock rate” is an impor- next.
tant parameter for the performance of the PHY layers.
PHY 1 utilizes lower optical rate of ≤ 400KHz. This is 3.3.1 MIMO Receiver
because PHY 1 is designed to be usable in outdoor sce- As we discussed in Section 2, there can be two types
narios as well where the LED transmitters are typically of receivers in VLC MIMO systems - photodiode and
high-power and can switch the intensity at a slower rate. image sensor. The performance of the system depends on
PHY II is designed to be used indoors where the optical whether imaging (image sensor) or non-imaging (photo-
switching rate can be as high as 120 MHz. The optical diode) receiver is used [85].
rate is 24 MHz for PHY III which is the current feasible Non-imaging receiver in a MIMO system is a set of
switching rate for white TriLED. independent photodiodes each with its individual con-
Depending on the choice of modulation, RLL code, centrator optics. The advantage of such a receiver is
optical clock rate, FEC code, the three PHY modes can that a very high gain can be achieved due to narrow
provide different data rates. PHY I can provide data rates FOV of each photodiode. The disadvantage, however, is
from 11.67 Kbps upto 266.6 Kbps. PHY II can achieve that such a receiver requires careful alignment with the

Optical FEC
MCS Modulation RLL code Data rate (kbps)
clock rate Outer code (RS) Inner code (CC)
0 (15,7) 1/4 11.67
1 (15,11) 1/3 24.44
2 OOK Manchester 200 KHz (15,11) 2/3 48.89
3 (15,11) none 73.3
4 none none 100
5 (15,2) none 35.56
6 (15,4) none 71.11
VPPM 4B6B 400 KHz
7 (15,7) none 124.4
8 none none 266.6

TABLE 5: 802.15.7 PHY I operating mode specifications and achievable throughput.

Optical Data rate

MCS Modulation RLL code FEC
clock rate (Mbps)
16 RS(64,32) 1.25
3.75 MHz
17 RS(160,128) 2
18 VPPM 4B6B RS(64,32) 2.5
19 7.5 MHz RS(160,128) 4
20 none 5
21 RS(64,32) 6
15 MHz
22 RS(160,128) 9.6
23 RS(64,32) 12
30 MHz
24 RS(160,128) 19.2
25 OOK 8B10B RS(64,32) 24
60 MHz
26 RS(160,128) 38.4
27 RS(64,32) 48
28 120 MHz RS(160,128) 76.8
29 none 96

TABLE 6: 802.15.7 PHY II operating mode specifications and achievable throughput.

MCS Modulation Optical clock rate FEC Data rate (Mbps)

32 4 CSK RS(64,32) 12
12 MHz
33 8 CSK RS(64,32) 18
34 4 CSK RS(64,32) 24
35 8 CSK RS(64,32) 36
36 16 CSK 24 MHz RS(64,32) 48
37 8 CSK none 72
38 16 CSK none 96

TABLE 7: 802.15.7 PHY III operating mode specifications and achievable throughput.

Modulation Data Rate Dimming Support Flickering Issue Comments

OOK Low to moderate Yes High Low-complexity transceiver design
PPM Moderate Yes Low Maximum spectral efficiency with MEPPM
OFDM High No Low Complex design due to LED non-linearity, MIMO support
CSK High Yes Low Requires RGB tri-LED, improved multi-user access

TABLE 8: Major modulation schemes and their characteristics

transmitters because of the narrow FOV, and the capacity of LOS paths using narrow FOV like photodiodes and
can reduce dramatically even with minor misalignment. can be robust by leveraging non-LOS paths whenever
Imaging Receiver: Since an image sensor contains a needed like an image sensor. Authors in [87] proposed
projection lens and a large matrix of photodiodes, it the design of a spherically-shaped receiver that is made
has the potential to create a high data-rate MIMO link. of a large number of photodiodes. Each of the photodi-
The projection lens ensures a large FOV which nearly ode has a narrow FOV and points in different direction
eliminates the alignment requirement. The disadvantage in the room. The photodiodes pointing to transmitter
of such as a receiver is that individual photodiodes have LED can receive the signal with high gain while other
limited gain and advance image processing is required photodiodes pointing to other directions can establish
to create an efficient MIMO channel. Also, the sampling non-LOS channels to increase spatial diversity. However,
rate of the image sensor is comparatively lower further using such a receiver incurs cost for additional hardware.
reducing the achievable throughput. Instead, authors in [88] proposed a way to improve the
The channel models of both imaging and non-imaging lower sampling rate of the image sensor. A token-based
receiver MIMO, and their relative benefits and limita- pixel selection method was proposed where instead of
tions were presented in [85]. It was shown in [86] that conventional row-scanning approach, only the pixels of
an “ideal” MIMO receiver can be a hybrid of imaging interest are selectively scanned to improve the sampling
and non-imaging sensors which can achieve high gains rate.

3.3.2 VLC MIMO Techniques constellation from the receiver’s perspective. Researchers
There are three types of VLC MIMO techniques pro- investigated the performance of spatial modulation in
posed in literature [89]. [101] when only partial channel state information (CSI)
Repetition Coding (RC): This is the simplest tech- is available and concluded that highly accurate CSI
nique where the same signal is transmitted from all the estimation is necessary to realize the full potential of SM.
transmitters. The transmitted signal from all LEDs meet The use of generalized spatial modulation was proposed
constructively at the receivers increasing the overall gain. in [102], [103]. Such modulation extends the original
Spatial Multiplexing (SMP): In SMP, different data scheme by allowing more than one transmitter to be
is transmitted from each transmitter to a receiver pho- active during the a symbol duration. It was shown that
todiode. With multiple transmitters and receivers, this due to additional flexibility of activating multiple LEDs,
type of MIMO creates multiple parallel SISO streams. the generalized scheme can achieve higher spectral effi-
The challenge is that receiver photodiodes have to be ciency compared to the conventional scheme, however,
accurately aligned to the transmitters to avoid any inter- at the cost of additional complexity in constellation
channel interference. SMP MIMO for optical channels design.
has been studied in some of the early works [90]–[92]. Optical MIMO for non-LOS diffuse links has not re-
In [90], [92], authors proposed optical wireless MIMO ceived much attention. Authors in [104] showed how
communication with subcarrier multiplexing where zero backward spatial filter can be used for optical wireless
forcing was utilized to cancel the interference from other MIMO in diffuse channels (no precise alignment of
transmit antennas. It was shown that for the transmitter transmitter and receiver). With user movements, such
semi-angle more than 20o , the transmitter-receiver sepa- diffuse channel are more likely in practical scenarios and
ration should be more than 1.5 meters for lower BER. The optimizing MIMO performance for such channels should
impact of optical beat interference on OMIMO scheme be investigated further.
of [90] was studied in [91]. Optical beat interference is
the signal degradation caused by multiple transmitters
3.3.3 Optical Beamforming
transmitting simultaneously on nearby wavelengths.
Spatial Modulation (SM): This MIMO technique was Beamforming allows multiple transmitters to concen-
proposed by [93]–[96] where only one transmitter trans- trate their signal in a specific direction based on the
mits data at any point of time. The constellation dia- receiver location. This type of transmit beamforming is
gram is extended to include the spatial dimension. Each well studied in RF communication and also utilized by
transmitter LED is assigned a specific symbol and when recent WLAN standards such as IEEE 802.11ac. Similar
data bits to be transmitted matches the symbol, the to RF beamforming, emitting light from multiple LEDs
LED is activated. The receiver estimates which LED was can be focused towards the receiver to create optical
activated based on the received signal, and uses this to beamforming. Recently, it was shown in [105] how light
decode the transmitted data. Since the data is encoded emitted from a single LED can be focused in a specific
in both spatial and signal domain, SM achieves much target direction using Spatial Light Modulator (SLM).
higher spectral efficiency compared to other techniques. SLM is an additional device that is required to modu-
A comparison of all the three MIMO techniques were late the phase or amplitude of the visible light signal.
provided in [89]. It was shown that RC is less restrictive It was shown that significant SNR improvements can
in terms of its requirement for transmitter-receiver align- be achieved by using the optical beamforming with
ment but provides only a limited spectral efficiency. SMP, any modulation technique. Authors in [106] derived
on the other hand, requires more careful alignment of the transmit beamforming vectors when multiple LEDs
transmitter-receiver but also provides higher data rates are used to perform the optical beamforming. Optical
compared to RC. SM achieves the best of both worlds beamforming can improve the performance of a visible
by being robust to correlated channels and providing light communication link significantly, however, there
higher spectral efficiency. Also, it was shown in [97] is only a limited amount of research done towards
that imaging receivers can obtain much higher SNR this. Performing optical beamforming while meeting the
when using SM or SMP technique compared to the non- illumination constraints is an important direction for
imaging receivers. research in VLC MIMO systems.
Due to its advantages over other MIMO techniques,
SM has been studied further in recent years. It was
shown in [98], [99] that power imbalance between the 4 L INK L AYER
transmitter LEDs can improve the performance of spatial
modulation especially when optical paths between the When there exists multiple transmitter LEDs and re-
transmitter and receiver are highly correlated. Authors ceiver devices connected to them, it is essential to control
in [100] studied the performance of spatial modulation the medium access, device association and device mobil-
using an implementation of 4 × 4 MIMO system and ity. In this section, we provide an overview of different
showed that the challenge in achieving higher through- techniques proposed in literature to manage link layer
put with SM is to maintain symbol separation in the services.

LED Photodiode

Beacon

Beacon
CAP CFP

GTS

GTS
Inactive

(a) Frame structure

Beacon
(b) Peer-to-peer
Peer-to-peer Star Broadcast
CAP
Fig. 17: VLC link layer topologies; reproduced from [8]
(c) Star Uplink GTS

Downlink GTS
4.1 Medium Access Control (MAC)
The application scenarios of VLC can be used to identify (d) Broadcast
the link layer topologies that need to be supported by the
Fig. 18: (a) IEEE 802.15.7 frame structure includes
MAC protocols. IEEE 802.15.8 [8] proposes three types
beacon, Contention-based Access Periods (CAP) and
of link layer topologies for VLC as shown in Fig. 17 -
Contention Free Periods (CFP), (b-d) Example usage of
1) Peer-to-peer: The peer-to-peer topology involves frame structure in different topologies; reproduced
one device acting as a coordinator (or master) for from [8]
the link between two devices. Both devices can
communicate with each other since the client has
an uplink to the master. This topology is typically off slot and then waits for a random number of back-
more suitable for high-speed Near-Field Commu- off slots before performing Clear Channel Assessment
nication (NFC). (CCA). If the channel is found to be idle, the device
2) Star: In a start topology, there can be many client starts to transmit. If the medium is found to be busy,
devices connected to a master device which acts as the device waits for additional random number of back-
the coordinator. A typical use case of this topology off slots before performing the CCA again.
is VLC wireless access networks. The MAC design The beacon-enabled random access also contains con-
is especially challenging in the star topology due tention free period which consists of multiple Guaran-
to many bi-directional links in the same collision teed Time Slots (GTS). This period is used by the coor-
domain. dinator to ensure medium access to devices with delay
3) Broadcast: Different from the star topology, the or bandwidth constrained applications. Depending on
client devices in a broadcast topology can only the requirement, a coordinator can also assign multiple
receive data from the master LED transmitter with- time slots to one GTS. Figs. 18(b), 18(c) and 18(d) show
out forming any uplink. Such topology can be used how different types of slots are used for beacon-disabled
for broadcasting information via LEDs throughout access in peer-to-peer topology, beacon-enabled access in
the network. Since there is no explicit association star and broadcast topologies respectively.
needed, the broadcast topology simplifies the MAC A preliminary version of CSMA/CA-based MAC was
design. implemented on a testbed by [107] with OOK modula-
Three types of multiple access control (MAC) schemes tion. The CSMA/CA implementation was extended in
are proposed for VLC - Carrier Sense Multiple Ac- [108] for cases where different LED transmitters have
cess (CSMA), Orthogonal Frequency Division Multiple different FOV. It shows that when transmitters are het-
Access (OFDMA) and Code Division Multiple Access erogeneous, avoiding collisions especially due to hidden
(CDMA). terminals is a challenging problem. It provides design
CSMA: There are two types of random channel access and implementation of Request-To-Send/Clear-To-Send
mechanisms proposed by IEEE 802.15.7 standard. In the (RTS/CTS) on OpenVLC platform (discussed in Sec-
first type, the beacons from the coordinator are disabled. tion 5). Authors observe that when an LED OFF symbol
Such beacon-disabled random access uses an unslotted (logic “0”) is being transmitted by an LED transceiver
random channel access with CSMA. Here, if a device A, the receiver node B can use the same optical channel
wishes to transmit, it first waits for a random back-off to transmit information to the LED A at the same time,
period and then senses the channel to be busy or not, enabling a bidirectional communication.
before transmitting. If the channel is found to be busy, OFDMA: OFDMA is a multi-carrier multiple access
the device waits for another random period before trying scheme where different users are assigned separate re-
to access the channel again. In the second type where source blocks (set of subcarriers in time) for communica-
the beacons are enabled, the time is divided into beacon tion. Application of OFDMA for multiple access in VLC
intervals. A superframe within the beacon interval con- is a natural extension to utilizing OFDM for modulation
tains Contention Access Periods (CAP) and Contention in physical layer. Two variations of OFDMA were com-
Free Periods (CFP) as shown in Fig. 18(a). If a device pared in [109] and it was shown that power efficiency
wishes to transmit, it first locates the start of a next back- and decoding complexity are two main challenges while

applying OFDMA to VLC. Authors in [110] proposed a The performance of the random codes was studied in
heuristic solution to subcarrier allocation problem in the [119] where the limits of their spectral efficiency was
case of interfering transmitters. Considering the spectral also demonstrated. It was shown in [120] that the re-
efficiency of OFMA, further research is necessary to flected light inside a room can increase the inter-symbol
design power-efficient and interference-aware resource interference of the OCDMA system, and the performance
allocation schemes for OFDMA. Authors in [111] pro- degrades furthermore as the number of users increase in
posed to use joint transmission from multiple LEDs the system. When utilizing optical CDMA for multi-user
using OFDMA to improve the SINR of the edge users in access, authors in [121] proposed a centralized power
a room. It was shown that due to intensity modulation allocation algorithm that maximizes the minimum SINR
of VLC systems, it is possible to achieve much better (Signal to Noise and Interference Ratio) of all devices.
coordinated multi-point transmission compared to RF This centralized algorithm achieves improved BER per-
systems. formance, however, it requires all the LED transmitters
CDMA: Optical CDMA (OCDMA) relies on optically to communicate with each other, which might not be
orthogonal codes to provide access to the same channel scalable due to its computational complexity in a large
by multiple users. The principle of optically orthogonal indoor environment with many LEDs and receivers.
codes is well-studied for the optical fiber networks [112], Recently, CSK modulation with RGB tri-LED transmit-
[113]. In the OCDMA for VLC, each device is assigned ter has been combined with OCDMA to achieve multi-
a code (binary sequence) such that the data can be en- user VLC. [122] and [123] showed how different CSK
coded in time domain by turning the LED ON and OFF symbols can be combined with the CDMA codes of users
[114]. These codes are referred as Optical Orthogonal to simultaneously transmit data to multiple devices.
Codes (OOC) [112]. It was shown in [114] synchronous Lastly, authors in [124] proposed a hybrid CDMA and
OCDMA can be implemented using the OOC codes and OFDM scheme that can achieve the advantages of both
OOK modulation with LED transmitters. A limitation as previously suggested for RF communication in [125].
of this technique is that long OOC codes are needed to Multi-User MIMO (MU-MIMO): Advanced MIMO
ensure optimality, which in turn reduces the achievable techniques such as Multi-user MIMO (MU-MIMO) are
data rate of devices. Authors in [115] proposed to ad- still to be designed and developed for VLC. Some early
dress this issue using Code Cycle Modulation (CCM) work such as [126] studied the multi-user MISO (Multi-
where different cyclic shifts of the sequence assigned ple Input Single Output) problem where multiple trans-
to devices are used to transmit an M-ary information. mitter LEDs (connected via power-line network in in-
Since any cyclic shift of an OOC code (with length L) is door environment) can coordinate to transmit the data to
considered a symbol, the spectral efficiency increases by different users while canceling the inter-user interference
a factor of log2 L. using zero-forcing pre-coding. Authors in [127] designed
Authors in [116], [117] showed that the CCM OOC a precoding MU-MIMO VLC system where block diag-
codes can provide higher spectral efficiency, however, onalization algorithm was adopted to eliminate multi-
they are not suitable for providing dimming support. user interference. It showed that such precoding reduces
This is because to achieve a different dimming level the receiver-side computational complexity and power
(different Peak-to-Average-Power-Ratio (PAPR)), a new consumption. The SNR and BER performance of the
set of OOC codes are required to be calculated. Two tech- scheme were studied in [128] and the impact of re-
niques have been proposed to address this issue in [116] ceiver’s FOV was analyzed. It was discussed in [129] that
and [117]. In the first method, user’s encoded binary the block diagonalization algorithm for precoding results
sequence is multiplied by BIBD (Balanced Incomplete in performance uncertainty and requires the number of
Block Designs) codewords, and the results are added receiving antennas (photodiodes) to be no more than the
to generate a Multi-level MEPPM signal. This way, the transmitting antennas (LEDs in an array). [129] proposed
dimming level can be changed by changing the ratio of to use Tomlinson-Harashima Precoding algorithm which
code-length to code-weight of the BIBD code, without is shown to achieve better BER performance compared
changing the OOC codes. In the second technique, the to the block diagonalization algorithm.
BIBD codes are partitioned in different subsets for dif-
ferent devices. The MEPPM scheme is used to generate
multi-level signals based on the assigned subsets. Since 4.2 Cell Design and Coordination
this technique provides larger constellation size, it can Cell Design: The requirement of managing access to
ensure higher data rate for each user. multiple devices in VLC is different than other types
It was also identified in [118] that when there are of networks. This is because the size of a cell can
a large number of devices sharing the channel access, vary depending on how illumination is provided. For
the OOC codes are difficult to generate. This has led to example, it is possible that one multi-LED luminaire on
design of random optical codes [118]. It was shown that the ceiling provides illumination to an entire room. In
the random codes are not optimal, however, their ease this case, the luminaire transmits data to multiple users
of generation and ability to support a large number of each possibly with multiple devices. We refer to this
users in multiple access make them a valid alternative. type of cell as an attocell [130]. Based on the required

Central
Controller

25 25
SNR (dB)

SNR (dB)
20 20
15 15
10 10
5 5
6 6
4 5 4 5
4 4
2 2 3 2 2 3
Width (m) 0 0 1 Width (m) 0 0 1
Length (m) Length (m)

Fig. 19: Rearranging LEDs can result in lower variance

of SNR in an indoor space; reproduced from [131] Cell 1 Cell 2 Cell 3

illumination, it is not difficult to see that radius of an Fig. 20: Device mobility can be managed by the central
attocell is no more than 10 meters. The other type of controller that coordinates the operations of VLC cells;
cell is even smaller in size where the luminaire provides reproduced from [8]
illumination mostly for personal usage. Such type of
luminaires are commonly used to brighten small spaces
in homes and spaces such as desk lamps etc. We refer proposed to use 12 LED luminaires in a circle and 4
to this type of cell as a zeptocell. The typical radius of luminaires on the corner (with the same total emitted
a zeptocell is no more than 5 meters and comparatively power) as shown in Fig. 19. It was shown that such an
fewer devices connect to a zeptocell. arrangement minimizes the variance of received power
Since multiple bright attocells are typically used for at different locations in the room. The choice of the
illumination of large indoor spaces, inter-cell interfer- radius of the circle of LEDs determines the delay spread
ence is more severe for attocells. For zeptocell, each of the received signal that can be optimized based on the
luminaire has lesser brightness compared to the lumi- spatial distribution of the receivers. In general, further
naire of an attocell. Also, the zeptocells are deployed research is necessary to design LED arrangements that
relatively sparsely which means that the inter-zeptocell can optimize communication performance while meet-
interference is not as severe as inter-attocell interference. ing the illumination constraints for a variety of indoor
The cell design for outdoor scenarios is discussed in layouts such as homes, hospitals, shopping malls etc.
Section 6.3. In indoor spaces, a common scenario that Cell Coordination: IEEE 802.15.7 [8] provides sugges-
can emerge in future is when multiple zeptocells are tions for managing cell design and techniques to reduce
deployed within the range of one or more attocells. inter-cell interference. It assumes that the LED transmit-
Such a heterogeneous network (referred as hetnet) is ters in an indoor environment are connected to a central
analogous to today’s cellular networks where many controller entity that can coordinate the cell operations
femtocells are deployed within the range of a macrocell and device mobility. A device associates with a cell for
to meet the traffic requirement. Important problems in which the signal strength of the beacon is maximum
such hetnet scenario are managing interference between among all the nearby cells. This is shown in Fig. 20. The
attocells and zeptocells, and determining resource-aware mobility management framework is similar to that of
association bias towards the zeptocells. Also, since the WiFi or other cellular networks where a device handover
primary purpose of installing luminaires is to provide occurs when the device moves from one cell to the other.
illumination, it is not clear that whether the resultant For managing the inter-cell interference, the transmitters
cell topology is interference-optimal for the communica- use frequency hopping where the controller ensures that
tion purpose or not. Further investigation is needed to interfering cells do not use the same frequency band
determine interference-optimal cell topology which can at the same time. Note that the support of frequency
maximize the throughput while meeting the illumination hopping depends on the capabilities of the LEDs, e.g.
requirements. if the LEDs are RGB LEDs, it is possible to provide
In some initial efforts to design VLC cells that provide many color bands that can be used for hopping. The cell
improved communication performance and also meet management, mobility and inter-cell interference can be
the illumination constraints, [132] and [131] proposed simplified with the use of central controller, however, it
a novel LED arrangement design. It showed that SNR is expensive to implement such a controller in practice
variation in a room is significantly high when one LED due to higher deployment cost of interconnects.
luminaire is placed at the center of the ceiling [23]. Inter-cell inference is known to be a challenging prob-
Although this is a common practice for illumination, lem in any type of cellular network design. From the
the differences in SNR can cause serious performance perspective of VLC, the inter-cell interference can cause
degradation for the users at the edge of the room. The much severe degradation of SINR for the cell edge users.
SNR variation is shown in Fig. 19. Authors in [131], [132] For example, in a room that has two LED transmitters

that can be used to evaluate newly design protocols

LED array
or techniques. We survey the existing evaluation plat-
forms that provide reprogrammability and flexibility in
VLC system design with the use of commercial off-the-
shelf hardware or software-defined radio platforms.
Fig. 22 shows a schematic diagram of a VLC sys-
tem with various components of the transmitter and
the receiver. The transmitter communication module is
responsible for modulation and digital to analog con-
version of the data. The dimming module maintains the
desirable dimming level for the illumination. The driver
Single point circuit combines the analog input data for communica-
transmission Multi-point Joint tion and dimming control signal (DC power level), and
Transmission
superimposes them to drive the LED. The visible light
Fig. 21: Multiple LED transmitters can jointly transmit signals emitted by the LED are then received by the pho-
to receivers located in the multi-point joint todiode. Note that both LED and photodiode typically
transmission region; reproduced from [133] employ a lens to achieve a specific FOV. The received
signal is filtered using an optical filter of specific wave-
length and amplified. The receiving communication
in two parts of the room, the users in the middle of module converts the received analog signals to digital
the room can experience very low SINR, which in turn and demodulates them. If the communication modules
results in low data rate. One of the early solution to are software-defined, various modulation/demodulation
the problem was proposed by [134] for infrared optical and MAC modules can be programmed and evaluated.
communication where the cells were partitioned into We next survey programmable VLC platforms that are
clusters. Different cells in a cluster used different fre- used in some of the recent research.
quency resources that are orthogonal to the neighboring
cells. The frequency resources were reused between the Input data
(Software-defined)
clusters. The limitation of this approach is that band- Tx communication
width available to each individual cell is limited which module
reduces the achievable data rate. To improve on such Driver circuit LED (array)
Dimming
static resource partitioning, authors in [135] proposed a control
dynamic resource allocation scheme where a channel re- module
source is dynamically assigned to devices based on cur-
rent inference. This scheme, however, incurs an overhead
of additional uplink communication that is necessary to (Software-defined)
Filter and
acquire a channel resource. Authors in [133] showed how Rx communication
amplifier
Photodiode
module
multiple LEDs can transmit simultaneously to the same
receiver using Joint Transmission (JT). It was shown Fig. 22: A block diagram showing various modulates of
that in order to synchronize the signals from multiple VLC transmitter and receiver.
LEDs, the transmitters can use different delay values
such that the signal is constructively at the receiver. The
We first discuss the low-cost solutions for VLC pro-
JT technique was further improved by [136] where it
totyping. We then provide how software-defined radios
was shown that if the LED transmitters are made of an
can be used to create VLC transceivers which provide
array of 7 LEDs each pointing to different directions, the
more flexibility in design at a higher cost.
edge users can benefit from joint transmission from LEDs
of different luminaires. The joint transmission region is
shown in an example in Fig. 21. Due to relatively smaller 5.1 Low Cost VLC Prototyping using Commodity
coverage of typical VLC cells, it is imperative to design Hardware
interference avoidance techniques that can ensure high The objective of such VLC system design is rapid pro-
data rate communication even with dense deployment totyping with low-cost off-the-shelf hardware.
of receivers. OpenVLC is an open-source implementation for net-
worked VLC research [137]. It consists of one BeagleBone
Black (BBB) board [138] as the communication module
5 S YSTEM DESIGN AND REPROGRAMMABLE
which implements PHY and MAC layer on Linux. The
TESTBEDS OpenVLC utilizes a bidirectional front-end which can
In this section, we introduce the details of VLC system act as an LED and a photodiode as well for transmitting
architecture. Detailed understanding of the system com- and receiving the light signals respectively. The trans-
ponents can allow researchers to build a VLC platform mitter and receiver mode can be switched using a tri-

state buffer in the front-end circuit. In its current form, much greater flexibility. They are more suitable for high
OpenVLC uses OOK modulation and Manchester RLL data-speed applications and realistic evaluation of vari-
coding at the transmitter. At the receiver, direct detection ous PHY protocols.
is implemented to demodulate the OOK signals. The Software-Defined VLC with WARP utilizes WARP
platform has been used for evaluation of CSMA/CD platform boards [143] developed at Rice University
MAC protocol in [108]. Currently, OpenVLC can operate as the software-defined communication module. The
with limited wattage LEDs reducing its communication WARP platform has been used widely in RF system
range to be less than a meter and throughput in range implementation and evaluation due to its flexibility and
of tens of kbps. extensibility. A WARP-based VLC platform which was
recently proposed by [144] is depicted in Fig. 24. Here,
the OOK modulation and demodulation modules are
implemented on the WARP boards. The ADC/DAC
module is interfaced with the WARP board. On the
transmitter side, the analog signal and the DC input are
combined using a Bias Tee implemented on the driver
circuit. On the receiver side, the driver circuit imple-
ments the filter and amplifier, while the ADC converts
the signal to digital domain and inputs them to the
WARP module. Similarly, authors of [145] demonstrated
the implementation of ACO-OFDM and DCO-OFDM for
VLC using the WARP boards. The WARP boards can
be an ideal platform for developing a hybrid VLC-RF
Fig. 23: VLC transmitter front-end with 20 LEDs system as shown in [146]. With the use of different FEC
providing 360o coverage; Image from [139] codes, [144] showed to achieve a data rate of 4 mbps
using OOK.
Similar to OpenVLC, a low-cost embedded evaluation
board (Atmel ATmega328P [140]) was used in the design
of LED-to-LED communication in [107]. A 2-PPM based OOK
Filter
Bias
Amp
mod. DAC Tee LED
PHY and CSMA based MAC was developed to run on
WARP DC Driver
the embedded boards. In the testbed used in [137] and
[107], the FOV of the LED is limited to the direction of
LED’s central axis. This limits its coverage to a specific
OOK
direction. This limitation was overcome by [139] where demod.
Filter
ADC
Filter Amp
Photodiode
authors designed an LED front end with 20 LEDs as WARP Driver
shown in Fig. 23. In the design, the LEDs are equally
spaced to provide coverage in 360o . The advantage of Fig. 24: Architecture of the VLC software-radio
such front-end is that a VLC node with multiple LEDs prototype; reproduced from [144].
can communicate with multiple other nodes in different
directions, enabling a multi-hop network of visible light The combination of GNU Radio [147] and USRP (Uni-
communication nodes. The developed front-end is based versal Software Radio Platform) [148] is another popular
on a bread-board which can interconnect with popular software platform widely used for RF research. [149]
low-cost evaluation boards such as BeagleBone [138], developed a VLC system utilizing the USRP and GNU
Raspberry Pi [141], Arduino [142] etc. Radio software. BPSK, QPSK and 2048 FFT length OFDM
As mentioned before, these prototyping allow faster modulation schemes were implemented in the testing,
development of VLC system at a lower cost which is and 2 Mbps data rate was achieved in the OFDM case.
sufficient in many research and commercial applications. Further, authors in [150] extended the VLC system by
However, their performance is often limited by the introducing an adaptive modulation for dynamic illu-
hardware (processor, ADC-DAC speeds, LED frequency mination that can dynamically modify the modulation
etc.), which makes them more suitable for low data- schemes in order to meet both the data communications
rate applications. The reconfigurability of such platforms and illumination requirements of a dual-use VLC sys-
is also limited as implementation of different PHY or tem.
MAC requires significant modifications to hardware and
The advantage of using software-defined radios is that
software design.
they provide improved flexibility because various PHY
and MAC modules can be implemented in software.
5.2 VLC Prototyping with Software-Defined Radios: Also, higher processing capacity of the software-defined
The use of software-defined radios allows redesigning platforms can provide high data-rates which is necessary
various communication modules (PHY, MAC etc.) with for evaluating VLC in access network scenarios.

6 S ENSING AND OTHER APPLICATIONS CT

TRB CT

The image sensor receiver available on today’s mobile

devices enables a new direction of research where VLC
can be combined with mobile computing to realize novel TRB
Code Area TRB

forms of sensing and applications. This sections provides

an overview of such sensing and communication appli- Corner
Tracker
cations. (CT) Timing Reference Blocks (TRB) CT

Fig. 25: Example of 2D COBRA barcode; reproduced

6.1 Indoor localization
from [162].
Location-based services have observed tremendous
growth in last few years. Mobile device localization in
outdoor scenarios largely depend on Global Position- used for triangulation to localize the receiver device.
ing System (GPS). However, the GPS does not work Luxapose to shown to achieve localization accuracy of
indoors, requiring alternative ways of localizing devices. ≈ 0.1 meters. Similar to Luxapose, it was shown in
Among the alternatives, WiFi-based indoor localization [155] how an imaging sensor can be used to receive
has proven to be most attractive where existing WiFi from multiple luminaires each of which creates a visual
AP deployment is leveraged to identify client’s location. landmark. Other variations of visible light localization
Although low-cost, WiFi-based indoor localization offers include [156] where both RSS and AoA are used for
lower accuracy (refer to [151] for overview), and complex two-phase hybrid localization. Authors in [157] showed
multi-path cancellation techniques [152] are required to that 3-D localization with sub-centimeter accuracy is also
improve the accuracy. feasible using multiple tilted receivers.
Similar to WiFi-based localization, indoor visible light With very high accuracy and ability to leverage the
communication system can also be leveraged for accu- existing lighting infrastructure, visible light localization
rate localization. The advantage of using VLC over WiFi will on the forefront of future location-based services.
for localization is that typically there are many more Commercial products such as ByteLight [158] are already
LED luminaires in a building compared to the number being available for retail markets.
of WiFi APs. It was observed in [153] that there are 10
times more LED luminaires than the number of WiFi
6.2 Screen-Camera Communication
APs in a typical indoor building. This higher density can
allow more accurate triangulation of the mobile device In this section, we take a look at a special application
resulting in higher accuracy. Epsilon [153] presented the of VLC where an LCD screen and a camera sensor
first practical visible light localization system. In Epsilon, can communicate for device-to-device communication.
the mobile device performs receiver side localization by LCD screens and cameras are widely used in today’s
receiving the light beacons from LED sources. Each LED mobile devices such as smartphones, laptops, etc. In
source broadcasts a beacon with identity and location infrastructure-to-device communication which we dis-
information. To avoid collisions between the beacons cussed in previous sections, LED luminaire serves a dual
of uncoordinated LED sources, a distributed channel purpose of illumination and communication. On the
hopping is utilized. The receiver (a photodiode) on a other hand, screen-camera communication is a form of
mobile device receives the beacon from multiple sources. device-to-device communication where information can
It utilizes the RSS values of received beacon to estimate be encoded in display screens of smartphone, laptop,
the distance from the received to the LED source. Based advertisement boards etc., and another device with a
on the distance estimates, the receiver uses trilateration camera sensor can capture the screen and decode the
to obtain its own location. Additionally, if the receiver data using image analysis. Due to the short wavelengths
can see less than 3 LED sources, the user can actively and narrow beams of visible light, LCD screen - camera
move the device to increase the visibility. It was shown links are highly directional, low-interference and secure.
that Epsilon can achieve the location accuracy of ≈ 0.4 It was first identified in [159], [160] through analysis
meters - compared to 3 – 6 meters accuracy achievable and experiments that such links are capable of achieving
in WiFi-based schemes [151]. hundreds of kbps to mbps of data rates. However, there
Another practical visible light localization approach are three main challenges in such links as described in
was presented in Luxapose [154]. Different from Epsilon, [161].
in Luxapose, the receiver is considered to an imaging • Perspective distortion is a common phenomenon
sensor such as the smartphone camera. The user takes an in daily life. When we take a look at an image
image of LED luminaires using the camera. The image is on a rectangular screen from a certain angle, the
then analyzed to detect the beacon information broadcast image on the screen appears more like a trapezoid.
by the luminaires and the angle-of-arrival (AoA) of the In particular, we observe that some pixels shrink,
beacon. Based on the orientation of the smartphone while others expand. This same phenomenon is also
camera, angle-of-arrival from the luminaires is then observed in screen-camera links where some pixels

have better visibility than others improving their as reference for detecting the data blocks in code area.
communication reliability. In the code area, the data (header, payload and CRC
• Blurring occurs when the camera moves while cap- checksum) is encoded by a sequence of color blocks in
turing the display. The result of blurring is out- the code area. For example, assuming red, green, blue
of-focus images where some pixels are blended and white colors represent 00, 01, 10 and 11 respectively.
together. In the frequency domain, blur can be In this way, a byte with value of 10110001 can be encoded
considered a low-pass filter where high frequency as ”blue white red green”. At the receiver, COBRA
attenuates much more than the low frequency [161]. includes a pre-processing component which selects the
• Ambient light is a source of noise which changes best quality for each barcode for further processing and
the luminance of the received pixels. This can cause a code extraction component which decodes the original
errors in the information encoded in the pixels, information. The COBRA is successfully implemented
resulting in information loss at the receiver. In fre- on Android smartphone, and experimental results show
quency domain, since ambient light changes the that COBRA can achieve a throughput of upto 172 Kbps
overall luminance, it can be considered the DC (much higher than PixNet’s throughput of < 10 Kbps).
component. The limited available throughput in screen-camera
links was further improved by LightSync. [163]. Authors
To solve these problems, inspired by traditional OFDM identify that improving the frame synchronization be-
modulation scheme in RF, PixNet [161] proposed to tween the transmitter and receiver can nearly double
encode the information in two-dimensional spatial fre- the achievable throughput. In LightSync, linear erasure
quencies. The main components of PixNet include a coding is used to recover the lost frames and color track-
perspective correction algorithm, blur-adaptive OFDM ing is used to decode the data correctly from imperfect
coding and a ambient light filter. The blur-adaptive frames. Another coding scheme referred as Styrofoam
OFDM coding is introduced at the transmitter where bits [164] addressed the problem of inter-symbol interference
are first modulated into complex numbers and then bro- due to lack of synchronization by inserting blank frames
ken down to symbols, then the symbols are arranged in in the code pattern. Further, authors in Hilight [165]
a two dimensional Hermitian matrix that guarantees the introduced a new scheme for screen-camera communica-
output is real. The transmitter treats different frequencies tion without any coded images. Leveraging the proper-
differently. Since the blur attenuates the high frequencies ties of orthogonal transparency (alpha) channel, HiLight
of an image, the information is transmitted through low “hides” the bits by changing the pixel translucency
frequency and protected with a Reed Solomon error cor- instead of modifying the RGB color. Experimental results
recting code. The RS code operates on a block size of 255 demonstrate the feasibility of HiLight’s by using the
and 8 bits elements in one block. Ambient light filter at off-the-shelf smartphones. Because screen-camera link is
the receiver can directly filter the zero frequency caused inherently unidirectional, [166] added reliability to the
by the ambient light. Perspective correction algorithm is communication using tri-level error correction through
partially implemented at the transmitter and partially at packet-frame-block structure. Similarly, authors in [167]
the receiver. It allows the PixNet system to work with proposed variable rate screen-camera links where in-
an irregular Sampling Frequency Offset (SFO) caused by stead of all-or-nothing detection, various intermediate
the perspective distortion, and use the SFO to re-sample resolutions of camera can allow lower data rate (anal-
the information at right frequencies to correctly recover ogous to rate adaption of RF links). The proposed lay-
the bits at the receiver. ered coding is shown to achieve improved reliability
Another approach for screen-camera communication at larger distances with minor penalties in throughput.
was presented in COlor Barcode stReaming for smArt- With increasing popularity of screen-camera links for
phones (COBRA) [162]. COBRA is designed to achieve near-field communication, it has become necessary to
one-way communication between small size screens and address the security aspects of the channel. Authors in
low-speed cameras of smartphones using 2-D color [168] showed why is it difficult to secure the screen-
barcodes. At the transmitter, a header and the CRC camera communication, and proposed enhancements by
checksum are added to the original data. Then each manipulating the screen viewing angles and user motion
byte of the data block is mapped to a certain color tracking.
to generate a color barcode that can be displayed on
the screen. The novel 2-D color barcode is shown in
Fig. 25. The barcode contains three types of areas - corner 6.3 Vehicular communication
trackers, timing reference blocks and code area in the In this section, we review the application of VLC in
center. The corner trackers are used to locate the barcode the vehicular communication. As the VLC based vehic-
on the screen using green and red blocks at top-left ular communication systems are used in the outdoor
and bottom-right corner respectively, and blue blocks scenario, they have one distinguishing characteristic
at the other two corners. The corner trackers can then compared to the indoor applications, namely the non-
be used by the camera to detect the black and white negligible ambient light interference due to background
timing reference blocks. These timing blocks are used solar radiation and other light sources, such as the road

light, the building lights, etc. Most of the VLC vehic- ter. To utilize these channel characteristics, the authors
ular communication systems address the problem and proposed a hierarchical transmission scheme which in-
present ways to mitigate the effects of intense ambient jects high priority data to low frequency components
interference. and low priority data to high frequency components
The VLC applications for vehicular communication to guarantee the high priority data can be received
fall into two categories: Vehicle to Infrastructure (V2I) even when the camera is far away from the LED traffic
and Vehicle to Vehicle (V2V). For the V2I applications lights. When the car approaches the LED traffic light, it
[169]–[171], they focus on utilizing the traffic related can receive the low priority data in the low frequency
infrastructure, such as traffic light, street light etc. to components.
communicate useful information. There are two types of To enable the VLC vehicular communication (V 2 LC),
cells in V2I communications. In the first type, the street two primary key elements should be examined: (i) the
lights whose primary purpose is to provide illumination feasibility of V 2 LC networks in the working condition
can be used for data communication with cars or pedes- which experiences noise and interference from solar
trians. Such VLC cells can typically provide coverage in radiation and other light sources; (ii) the capability of
50 - 100 meters range. The other type of outdoor LEDs V 2 LC networks to provide efficient services to sup-
are traffic signaling LEDs that can communicate with port vehicular applications. Liu et al. [172] suggested
cars. Since their primary purpose is not illumination additional services feasible with V 2 LC such as vehicle-
and because they are always ON (even when there is to-vehicle broadcasting, infrastructure-to-vehicle broad-
sunlight), they are more suitable for applications such casting, etc. After examining the ability of V 2 LC to
as vehicle safety, traffic information broadcast etc. On satisfy the requirements of vehicular applications, they
the other hand, the illumination LEDs are available on find V 2 LC could achieve efficient communication in
streets even where there are no traffic lights, making the dense vehicular traffic condition. V 2 LC is more
them more suitable for high-speed Internet access type of resilient to the background noise from solar radiation
applications. For the V2V applications [172]–[175], they (the diffused sunlight), but is much more susceptible to
mainly work on exploiting the headlights and taillights the direct sunlight. Besides, the nocturnal noise coming
on automobile as transmitter, and the photodiode or from idle VLC transmitters as well as legacy lights with
image sensor as the receiver to provide reliable commu- no data transmission abilities has very limited impact
nications between vehicles. on V 2 LC, which means V 2 LC is robust to this kind of
For the VLC based vehicular communication systems, noise.
both types of receivers - the photodiode [172], [174], [175] One of the most important purposes of vehicular com-
and the image sensor [169]–[171], [176], [177] are used. munication systems is to provide road safety. Dedicated
Compared with the photodiode, if there are different Short-Range Communication (DSRC, which utilizes the
light sources and each signal is modulated individually, 5.9 GHz radio spectrum) is usually regarded as the most
the image sensor can recognize all of them simultane- promising technology to support V2V communication. A
ously and achieve the parallel data transmission. An- comparison between DSRC and VLC was presented in
other advantage of the image sensor is that it is more [175], and following advantages of VLC over DSRC were
resistant to the light interference. Since most mobile outlined
phones have an integrated camera sensor which makes it • Lower complexity and cost: Due to the much
an attractive choice for vehicular applications. However, smaller multipath effect, the design of VLC trans-
most image sensors are limited by the rolling shutter mitter is much easier than the RF transmitter. Also,
scheme which means they can not capture the whole the LED lights already exist in auotomobiles, while
scene instantaneously but scan it vertically or horizon- RF based systems incur additional cost of deploying
tally. This limits the sampling rate for receiving data. On the equipment.
the other hand, the photodiode can support higher data • Scalability: The RF based V2V communication
rates at lower cost, which enables it as receiving devices scales poorly when the vehicle density increases in
for economic systems. the communication range, but for VLC, only the
A V2I system was presented in [169] which adopts vehicles in the LOS are in the same contention
the high-speed camera deployed on the automobile to domain, which means they experience much less
efficiently and accurately receive the signal from traffic interference.
lights. If the camera is far from LED traffic lights, it is • Positioning capability: RF based positioning
hard to differentiate each individual data pattern sent schemes can not provide sub-meter accuracy. The
out by the LED because of reduction of pixel size and VLC provides a promising way to perform relative
de-focusing of the LED data pattern. Hence, the high positioning with sub-meter accuracy due to the
spatial frequency components of data pattern are often high directivity of visible light.
lost. However, the low frequency components remain in • Security: In the VLC scenario, if an attacker tries
the pixels, which means that the high-speed camera can launch an attack, it must be within the LOS range of
receive the low-frequency LED data pattern contained in the victim, which means the attacker will be exposed
these pixels, even if the camera is far from the transmit- with higher possibility. Hence, compared to the RF

communication, VLC also provide a better security deployed in indoor environment, visible light sensing
in vehicular communication. and gesture recognition have the potential to solve many
With the communication and positioning capabilities contemporary HCI-related challenges.
of VLC, it is possible to accurately construct a map of
vehicle’s surrounding as shown in [174], [175]. Based on 7 S UMMARY AND C HALLENGES
the map, a car can not only obtain the distance from Visible light communication has the potential to provide
other cars, but it can also broadcast its real-time speed to high speed data communication with improved energy
the neighbors through the headlight and taillight. Other efficiency and communication security/privacy. With
vehicles receiving this information can adjust their speed looming crisis of RF spectrum shortage, VLC can become
accordingly (especially useful for self-driving cars) to a practical augmentation technology for the existing RF
maximize the fuel efficiency and minimize the chances networks. Increasing interest from research community
of collisions. and industries as well the standardization efforts such
as VLCA and IEEE 802.15.7 show that VLC can be
successfully commercialized in coming years.
6.4 Human-Computer Interaction (HCI) using Visible In this survey, we provided an overview of literature
Light covering visible light communication, networking and
There is a growing interest in utilizing the wireless sensing. We first discussed various components of a
communication systems for enabling improved HCI. visible light communication system including LED de-
Recently, RF communication systems, especially WiFi, sign and type of receivers. We then provided a compre-
has been extended to perform motion detection [178], hensive survey of VLC communication channel model
gesture recognition [179] and efficient input detection and its propagation characteristics. This included a dis-
[180]. Visible light-based interaction systems are well- cussion of path loss, multi-path, SNR and shadowing.
studied in research and many similar commercial prod- With this understanding of channel propagation, we
ucts are available already. For example, optical mouse provided a survey of VLC physical layer modulation
utilizes LEDs and photodiodes to detect fine-grained techniques. We discussed how different modulation tech-
motion. Similarly, Kinect [181] system uses a combina- niques should be able to provide dimming support and
tion of infrared and visible light to perform accurate minimize flickering effect while maintaining higher spec-
3D-gesture recognition. However, the problem with such tral efficiency. This included a discussion of four major
3D-gesture recognition systems is that they are expensive modulation techniques (OOK, PPM, OFDM and CSK). It
mostly because they require sophisticated image sensors was shown that due to their higher data rate capacity,
along with advanced graphics techniques to process the OFDM and CSK are likely to be play an important role
captured images. in future VLC broadband access networks. Additionally,
Some recent research has proposed inexpensive means feasibility of VLC MIMO as shown in literature ensures
of providing richer HCI using visible light. Authors in further data rate enhancements. This was followed by
[182] showed that human presence or motion causes a survey of link layer protocols for VLC. An elaborate
changes in the electromagnetic field around the flu- discussion on current CSMA, OFDMA, CDMA and MU-
orescent lamp. This changes result in variations in MIMO protocols was provided. We covered a variety
home/office power-line network, hence the gestures can of inter-cell interference management techniques, and
be recognized by another pluggable module anywhere highlighted the importance of techniques such as LED
on the power-line network. Similarly, it was demon- rearrangement and joint transmission. A review of exist-
strated in [183] how LEDs can be used to receive the ing VLC system platforms that provide reprogrammabil-
visible light (like a photodiode) and applied it to differ- ity and flexibility of implementation was then provided
ent sensing applications. PICOntrol [184] showed how with a discussion on low-cost platforms as well as
a pico-projector can be used to emit visible light and software-defined platforms. We then provided a detailed
enable remote control to any device that has a simple survey of an interesting upcoming field of VLC sensing
embedded control unit. The projected light on the sensor where literature on visible light localization, screen-
unit provides a GUI where user can provide various camera NFC, vehicular networking and VLC-assisted
commands to control the physical object. Okuli [185] HCI techniques was reviewed.
presented a system where user’s finger can be precisely Based on the review of current literature on VLC, we
located within a small workspace using an LED and now outline the important challenges that need to be
two photodiodes, allowing the user to interact with mo- addressed in near future. Solving these challenges are
bile/wearable devices with small form-factors. Authors essential so that VLC can be deployed in practice as a
in [186]–[188] designed techniques where user’s gesture high-speed mobile networking technology.
and complete skeleton posture can be reconstructed
using her shadow (detected via photodiodes on the 7.1 FOV Alignment and Shadowing
floor) to enable a variety of applications in smart-spaces The techniques of achieving high data rates in VLC links
and person identification. As LEDs become prevalent, primarily assume an LOS channel where the transmit-
and more and more photodiodes/image sensors are ter and the receiver have aligned their field of views

to maximize the channel response. However, in more The cost of deploying wired infrastructure (e.g. Ethernet,
practical scenarios, receiver movement and orientation fiber etc.) can be very high, which in turn can cancel the
changes are common. For example, in VLC-based indoor benefits of reusing the LEDs for communication. Provid-
access network, user’s smartphone equipped with a light ing wireless connectivity is feasible, however, with dense
sensor can move and rotate based on user’s actions. This deployment of LEDs, the interference of wireless connec-
means that receiver’s FOV cannot always be aligned tions can be a limiting factor. This can reduce the achiev-
with the transmitter. As we saw in Section 3.1, the drop able Internet data rate for the LEDs. In some recent work,
in received optical power can be significant due to such power-line communication has been proposed in [189] as
misalignment. It is necessary to design techniques that a way to interconnect LEDs. The power-line communi-
can ensure high data rates even in the presence of FOV cation provides an attractive choice as it can reuse the
misalignment cases. This requires design and develop- existing power line network for communication without
ment of methods that can provide graceful degradation additional cost of cable deployment. However, the use
in data rate using the optical power of reflected light. of power line incurs cost overheads of using Ethernet-
Designing such techniques is extremely challenging and to-power modem and power-to VLC modems. Apart
is an important direction of future research. from this cost, the performance and coverage issues [190]
Apart from the FOV alignment, another critical limi- of power line communications should be addressed for
tation is imposed by shadowing events. When an object successful power-line and VLC integration. There is a
or human blocks the LOS, the observed optical power scope for designing novel techniques that can provide
degrades substantially, resulting in severe data rate re- high-speed Internet connectivity to the LEDs at low cost.
duction. As discussed in Section 3.1.5, limited research
is done to understand/model the effect of shadowing 7.4 Inter-cell Interference
events on VLC. When the LOS is blocked by a shad-
The dense deployments of LEDs can provide higher
owing event, it is not only necessary to exploit the
capacity due to smaller cell radius. However, as dis-
reflected optical power of the diffuse channels but it also
cussed in Section 4.2, the VLC small cell architecture puts
necessary to do so in a timely manner as typical blockage
forward a challenge of managing inter-cell interference.
events can be of very short duration (e.g. human passing
On one hand, visible light communication provides less
by). Thus, it is imperative to design techniques that can
interference since the visible light is blocked by walls
quickly react to changes in received power due to FOV
which naturally restricts its propagation to rooms in
misalignment and shadowing.
indoor spaces. On the other hand, the LEDs inside a
room (in the same collision domain) can cause severe
7.2 Receiver Design and Energy Efficiency interference to each other, and the performance can
Current VLC receivers either use a photodiode or an degrade due to low SINR. The problem can be solved by
imaging sensor for receiving the VLC signals. The use of employing various techniques such as network MIMO,
photodiode is more suitable for stationary clients where joint transmission and LED rearrangement. In network
its FOV can be aligned to the LED fixture for high MIMO and joint transmission, the interfering LEDs can
received optical power. On the mobile devices, the imag- coordinate their transmissions via interference nulling or
ing sensor can be used since they have comparatively synchronization to ensure high SINR at the receivers.
larger FOV (due to wider concentration lens), making Another method to combat the interference is to rear-
the mobile device a little more robust to movements and range the LEDs such that their mutual interference is
FOV misalignment. However, due to a large number less. For this method, systematic design and analysis are
of photodiodes, operating imaging sensor is slow and required for optimizing the communication performance
energy expensive. This can significantly slow down the while meeting the illumination constraints.
overall achievable data rate. This is natural given that
imaging sensor was primarily designed for image and 7.5 Uplink and RF Augmentation
video capture, and not for the communication. Thus, it is
Almost all current research in visible light networking
challenging to design a receiver that can provide robust-
focuses on downlink (from LED luminaire to photodi-
ness to device movements and FOV misalignment. The
ode/image sensor receiver) traffic without taking into
receiver should also operate with low energy consump-
consideration how the uplink can operate. Although
tion to be useful on battery-powered mobile devices
efficient LEDs are incorporated in today’s mobile devices
while providing high-speed visible light communication.
as camera flashlight or notification indicator, they can
not be used directly for communication. This is because
7.3 LED to Internet Connectivity constantly turning on the LED not only consumes signif-
In order to create a VLC-based broadband access net- icant energy for mobile devices but it also causes visual
work, it is necessary to connect the LEDs to Internet. disturbance to users while using the devices. Also, VLC
Because the deployment of LEDs (for illumination pur- uplink requires that user’s mobile device maintains a
poses) is likely to be very dense, it becomes a challenging directional beam towards the receiver which can result
problem to connect the large number of LEDs to Internet. in significant throughput reductions when the mobile

device is constantly moving/rotating. To address these in the design. Most of the current research in VLC
challenges, use of other types of communication has has focused on physical and MAC layer performance
been proposed where RF [191]–[193] or infrared [194] enhancements with stationary devices. Equipped with
can be used for transmitting uplink data. these techniques, one of the most important direction of
RF-based uplink transmission is an attractive alter- future research in VLC is the urgent need to address the
native considering that WiFi is already omni-present issues that arise when utilizing mobile devices in VLC
especially in indoor environments. Operating VLC small access networks. As discussed before, researchers have
cells under the coverage of WiFi cells (larger range) already started addressing these challenges, and more
also ensures that clients have uninterrupted connec- and more research is likely to follow the direction in
tivity when VLC communication is not available (e.g. near future.
night time, blockage etc.) Utilizing 3G and 4G cellular
networks such as LTE is also feasible, however, they
impose additional challenges when the indoor wireless R EFERENCES
network provider is different than the cellular network [1] United States Department of Energy. Energy Savings Forecast
provider. Utilizing different communication technologies of Solid-State Lighting in General Illumination Applications.
[Online]. Available: http://apps1.eere.energy.gov/buildings/
for uplink and downlink gives rise to heterogeneous publications/pdfs/ssl/energysavingsforecast14.pdf
networks (HetNets) [195]–[197]. Such networks impose [2] Y. Tanaka, S. Haruyama, and M. Nakagawa, “Wireless optical
additional practical challenges such as complex network transmissions with white colored LED for wireless home links,”
in Personal, Indoor and Mobile Radio Communications, 2000. PIMRC
management for multi-homed clients [198], throughput 2000. The 11th IEEE International Symposium on, vol. 2, 2000, pp.
asymmetry issues for transport layer [199], link-layer 1325–1329 vol.2.
packet loss management and reliable data delivery etc. [3] Visible Light Communications Consortium (VLCC). [Online].
Available: http://www.vlcc.net/
There is only a limited amount of research done in [4] Japan Electronics and Information Technology Industries
asymmetric and heterogeneous networking. In order to Association (JEITA). [Online]. Available: http://www.jeita.or.jp/
build robust high-speed HetNet of VLC and RF, it is [5] Infrared Data Association (IrDA). [Online]. Available: http:
//www.irda.org
crucial that these challenges are resolved. [6] Home Gigabit Access (OMEGA) Project. [Online]. Available:
7.6 Mobility and Coverage http://www.ict-omega.eu/
[7] Visible Light Communications Association (VLCA). [Online].
In order for VLC to be a ubiquitous mobile technology, it Available: http://vlca.net/
[8] Institute of Electrical and Electronics Engineers (IEEE), “802.15.7
is necessary that it can provide uninterrupted and high- IEEE Standard for Local and Metropolitan Area Networks–Part
speed connectivity in presence of user mobility within 15.7: Short-Range Wireless Optical Communication Using Visible
a VLC cell and between the VLC cells. User mobility Light,” 2011.
introduces novel issues for VLC that are significantly dif- [9] D. Karunatilaka, F. Zafar, V. Kalavally, and R. Parthiban, “LED
Based Indoor Visible Light Communications: State of the Art,”
ferent than RF. For example, as shown in [144], even in a Communications Surveys Tutorials, IEEE, vol. PP, no. 99, pp. 1–1,
small VLC cell, the client SNR varies dramatically (many 2015.
times within the frame transmission duration) when [10] D. Tsonev, S. Videv, and H. Haas, “Light fidelity (Li-Fi): towards
all-optical networking,” in SPIE OPTO. International Society
user moves within the cell. These fast variations have for Optics and Photonics, 2013, pp. 900 702–900 702.
to be accounted for when designing various link-layer [11] A. Sevincer, A. Bhattarai, M. Bilgi, M. Yuksel, and N. Pala,
techniques such as rate adaptation, frame aggregation “LIGHTNETs: Smart LIGHTing and Mobile Optical Wireless
NETworks - A Survey,” Communications Surveys Tutorials, IEEE,
etc. In many indoor places, LED deployment is inten- vol. 15, no. 4, pp. 1620–1641, Fourth 2013.
tionally made sparser to leverage the existing sunlight. [12] M. Khalighi and M. Uysal, “Survey on Free Space Optical
When using LEDs for communication, it is necessary Communication: A Communication Theory Perspective,” Com-
munications Surveys Tutorials, IEEE, vol. 16, no. 4, pp. 2231–2258,
that sufficient coverage is available in all areas of an Fourthquarter 2014.
indoor space. When VLC is used as augmentation to RF [13] F. Demers, H. Yanikomeroglu, and M. St-Hilaire, “A Survey of
networks, user mobility requires seamlessly managing Opportunities for Free Space Optics in Next Generation Cellu-
lar Networks,” in Communication Networks and Services Research
horizontal (VLC to VLC) as well as vertical (VLC to RF) Conference (CNSR), 2011 Ninth Annual, May 2011, pp. 210–216.
handover of user devices. [14] N. Kumar and N. R. Lourenco, “LED-based visible light com-
munication system: a brief survey and investigation,” Journal of
Apart from these challenges, some recent research Engineering and Applied Sciences, vol. 5, no. 4, pp. 296–307, 2010.
[200] has shown the feasibility of eavesdropping on [15] Deviant Art. [Online]. Available: http://flexpoint.deviantart.
visible light communication from outside a room us- com/art/Small-Living-Room-87530751
ing the light signals leaked through the gap between [16] C. Danakis, M. Afgani, G. Povey, I. Underwood, and H. Haas,
“Using a CMOS camera sensor for visible light communication,”
floor and door, keyhole and even partially covered in Globecom Workshops (GC Wkshps), 2012 IEEE, Dec 2012, pp.
windows. These early insights demonstrate that further 1244–1248.
investigations are required to evaluate security/privacy [17] A. Zukauskas, M. Shur, and R. Gaska, Introduction to solid-state
lighting. J. Wiley, 2002.
of VLC in access networks. The challenges described [18] CIE, “Commission Internationale de lEclairage Proc.” Cambridge
above primarily stem from the use of mobile devices as University Press, 1931.
receivers. To successfully build and operate a VLC access [19] Cree XLamp XP-E High-Efficiency White. [Online].
Available: http://www.cree.com/∼/media/Files/Cree/
network, it is imperative to address the issues such as LEDComponentsandModules/XLamp/DataandBinning/
device movement, user mobility and energy efficiency XLampXPEHEW.pdf

[20] G. Wyszecki and W. Stiles, “Color science: concepts and meth- Conference 2013. Optical Society of America, 2013, p.
ods, quantitative data and formulae,” Wiley classics library, Au- JTh2A.73. [Online]. Available: http://www.opticsinfobase.org/
gust 2000. abstract.cfm?URI=NFOEC-2013-JTh2A.73
[21] Cree XLAMP XR-E. [Online]. Available: http://www.cree.com/ [41] S. Muthu and J. Gaines, “Red, green and blue LED-based white
∼/media/Files/Cree/LEDComponentsandModules/XLamp/ light source: implementation challenges and control design,” in
DataandBinning/XLamp7090XRE.pdf Industry Applications Conference, 2003. 38th IAS Annual Meeting.
[22] Cree LMH6. [Online]. Available: http://www.cree.com/∼/ Conference Record of the, vol. 1, Oct 2003, pp. 515–522 vol.1.
media/Files/Cree/LEDComponentsandModules/Modules/ [42] S. Kaur, W. Liu, and D. Castor, VLC Dim-
DataSheets/LEDModules LMH6.pdf ming Support IEEE 802.15-09-0641-00-0007,
[23] T. Komine and M. Nakagawa, “Fundamental analysis for visible- https://mentor.ieee.org/802.15/dcn/09/15-09-0641-00-0007-
light communication system using LED lights,” Consumer Elec- vlc-dimming-proposal.pdf.
tronics, IEEE Transactions on, vol. 50, no. 1, pp. 100–107, Feb 2004. [43] J. K. Kwon, “Inverse Source Coding for Dimming in Visible Light
[24] K. Cui, G. Chen, Z. Xu, and R. Roberts, “Line-of-sight vis- Communications Using NRZ-OOK on Reliable Links,” Photonics
ible light communication system design and demonstration,” Technology Letters, IEEE, vol. 22, no. 19, pp. 1455–1457, Oct 2010.
in Communication Systems Networks and Digital Signal Processing [44] G. Ntogari, T. Kamalakis, J. Walewski, and T. Sphicopoulos,
(CSNDSP), 2010 7th International Symposium on, July 2010, pp. “Combining Illumination Dimming Based on Pulse-Width Mod-
621–625. ulation With Visible-Light Communications Based on Discrete
[25] Spectral Response of a Photodiode. [Online]. Available: http: Multitone,” Optical Communications and Networking, IEEE/OSA
//en.wikipedia.org/wiki/Photodiode Journal of, vol. 3, no. 1, pp. 56–65, January 2011.
[26] B. Xie, G. Tan, Y. Liu, M. Lu, K. Chen, and T. He, “LIPS: A Light [45] H. Sugiyama, S. Haruyama, and M. Nakagawa, “Brightness Con-
Intensity Based Positioning System For Indoor Environments,” trol Methods for Illumination and Visible-Light Communication
arXiv preprint arXiv:1403.2331, 2014. Systems,” in Wireless and Mobile Communications, 2007. ICWMC
[27] K. Lee, H. Park, and J. Barry, “Indoor Channel Characteristics for ’07. Third International Conference on, March 2007, pp. 78–78.
Visible Light Communications,” Communications Letters, IEEE, [46] J. Vucic, C. Kottke, S. Nerreter, K.-D. Langer, and J. Walewski,
vol. 15, no. 2, pp. 217–219, February 2011. “513 Mbit/s Visible Light Communications Link Based on DMT-
[28] K. Cui, “Physical Layer Characteristics and Techniques for Vis- Modulation of a White LED,” Lightwave Technology, Journal of,
ible Light Communications,” Ph.D. dissertation, UNIVERSITY vol. 28, no. 24, pp. 3512–3518, Dec 2010.
OF CALIFORNIA, RIVERSIDE, 2012. [47] C. Georghiades, “Modulation and coding for throughput-
[29] S. Lee, J. K. Kwon, S.-Y. Jung, and Y.-H. Kwon, “Evaluation of efficient optical systems,” Information Theory, IEEE Transactions
visible light communication channel delay profiles for automo- on, vol. 40, no. 5, pp. 1313–1326, Sep 1994.
tive applications,” EURASIP Journal on Wireless Communications [48] D. shan Shiu and J. Kahn, “Differential pulse-position modula-
and Networking, vol. 2012, no. 1, pp. 1–8, 2012. tion for power-efficient optical communication,” Communications,
[30] L. Yuan-jian, S. Qin-jian, M. Xue, and Z. Ye-rong, “Simulation IEEE Transactions on, vol. 47, no. 8, pp. 1201–1210, Aug 1999.
and Analysis of Indoor Visible Light Propagation Characteristics
[49] F. Gfeller, W. Hirt, M. de Lange, and B. Weiss, “Wireless infrared
Based on the Method of SBR/Image,” International Journal of
transmission: How to reach all office space,” in Vehicular Technol-
Antennas and Propagation, vol. 2014, 2014.
ogy Conference, 1996. Mobile Technology for the Human Race., IEEE
[31] Z. Ghassemlooy, W. Popoola, and S. Rajbhandari, Optical Wireless
46th, vol. 3, Apr 1996, pp. 1535–1539 vol.3.
Communications: System and Channel Modelling with MATLAB,
[50] T. Ohtsuki, “Rate adaptive transmission scheme using punctured
1st ed. Boca Raton, FL, USA: CRC Press, Inc., 2012.
[32] T. Komine and M. Nakagawa, “A study of shadowing on indoor convolutional codes in a fixed channel reuse strategy with PPM
visible-light wireless communication utilizing plural white LED CDMA,” in Global Telecommunications Conference, 1998. GLOBE-
COM 1998. The Bridge to Global Integration. IEEE, vol. 1, 1998, pp.
lightings,” in Wireless Communication Systems, 2004, 1st Interna-
tional Symposium on, Sept 2004, pp. 36–40. 207–212 vol.1.
[33] M. S. Rea, “The IESNA lighting handbook: reference & applica- [51] ——, “Rate-adaptive indoor infrared wireless communication
tion,” Illuminating Engineering Society of North America, 2000. systems using repeated and punctured convolutional codes,”
[34] S. M. Berman, D. S. Greehouse, I. L. Bailey, R. D. Clear, and vol. 4, no. 2, Feb 2000, pp. 56–58.
T. W. Raasch, “Human electroretinogram responses to video [52] B. Bai, Z. Xu, and Y. Fan, “Joint LED dimming and high capacity
displays, fluorescent lighting, and other high frequency sources,” visible light communication by overlapping PPM,” in Wireless
Optometry & Vision Science, vol. 68, no. 8, pp. 645–662, 1991. and Optical Communications Conference (WOCC), 2010 19th Annual,
[35] J. Grubor, S. C. J. Lee, K.-D. Langer, T. Koonen, and J. W. May 2010, pp. 1–5.
Walewski, “Wireless High-Speed Data Transmission with Phos- [53] H. Sugiyama and K. Nosu, “MPPM: a method for improving the
phorescent White-Light LEDs,” in Optical Communication - Post- band-utilization efficiency in optical PPM,” Lightwave Technology,
Deadline Papers (published 2008), 2007 33rd European Conference Journal of, vol. 7, no. 3, pp. 465–472, March 1989.
and Exhibition of, Sept 2007, pp. 1–2. [54] T. Ohtsuki, I. Sasase, and S. Mori, “Overlapping multi-pulse
[36] S. Park, D. Jung, H. Shin, D. Shin, Y. Hyun, K. Lee, and pulse position modulation in optical direct detection channel,”
Y. Oh, “Information broadcasting system based on visible light in Communications, 1993. ICC ’93 Geneva. Technical Program, Con-
signboard,” Proc. Wireless Opt. Commun, pp. 311–313, 2007. ference Record, IEEE International Conference on, vol. 2, May 1993,
[37] H. L. Minh, D. O’Brien, G. Faulkner, L. Zeng, K. Lee, D. Jung, and pp. 1123–1127 vol.2.
Y. Oh, “High-speed visible light communications using multiple- [55] ——, “Performance analysis of overlapping multi-pulse pulse
resonant equalization,” Photonics Technology Letters, IEEE, vol. 20, position modulation (OMPPM) in noisy photon counting chan-
no. 14, pp. 1243–1245, July 2008. nel,” in Information Theory, 1994. Proceedings., 1994 IEEE Interna-
[38] J. Vucic, C. Kottke, S. Nerreter, K. Habel, A. Buttner, K.-D. Langer, tional Symposium on, Jun 1994, pp. 80–.
and J. Walewski, “125 Mbit/s over 5 m wireless distance by [56] T. Ohtsuki and I. Sasase, “Capacity and Cutoff Rate of Overlap-
use of OOK-Modulated phosphorescent white LEDs,” in Optical ping Multi-Pulse Pulse Position Modulation (OMPPM) in Op-
Communication, 2009. ECOC ’09. 35th European Conference on, Sept tical Direct-Detection Channel: Quantum-Limited Case,” IEICE
2009, pp. 1–2. TRANSACTIONS on Fundamentals of Electronics, Communications
[39] J. Vucic, C. Kottke, S. Nerreter, K. Habel, A. Buttner, K.-D. Langer, and Computer Sciences, vol. 77, no. 8, pp. 1298–1308, 1994.
and J. Walewski, “230 Mbit/s via a wireless visible-light link [57] T. Ohtsuki, I. Sasase, and S. Mori, “Trellis coded overlapping
based on OOK modulation of phosphorescent white LEDs,” in multi-pulse pulse position modulation in optical direct detection
Optical Fiber Communication (OFC), collocated National Fiber Optic channel,” in Communications, 1994. ICC ’94, SUPERCOMM/ICC
Engineers Conference, 2010 Conference on (OFC/NFOEC), March ’94, Conference Record, ’Serving Humanity Through Communica-
2010, pp. 1–3. tions.’ IEEE International Conference on, May 1994, pp. 675–679
[40] N. Fujimoto and H. Mochizuki, “477 Mbit/s visible light vol.2.
transmission based on OOK-NRZ modulation using a [58] T. Ohtsuki and I. Sasase, “Error Performance of Overlapping
single commercially available visible LED and a practical Multi-Pulse Pulse Position Modulation (OMPPM) and Trellis
LED driver with a pre-emphasis circuit,” in Optical Coded OMPPM in Optical Direct-Detection Channel,” IEICE
Fiber Communication Conference/National Fiber Optic Engineers transactions on communications, vol. 77, no. 9, pp. 1133–1143, 1994.

[59] D. Zwillinger, “Differential PPM has a higher throughput than [80] S. Rajagopal, R. Roberts, and S.-K. Lim, “IEEE 802.15.7 visible
PPM for the band-limited and average-power-limited optical light communication: modulation schemes and dimming sup-
channel,” Information Theory, IEEE Transactions on, vol. 34, no. 5, port,” Communications Magazine, IEEE, vol. 50, no. 3, pp. 72–82,
pp. 1269–1273, Sep 1988. March 2012.
[60] D. shan Shiu and J. Kahn, “Differential pulse-position modula- [81] A. Yokoi, J. Son, and T. Bae, CSK constellation in all color band
tion for power-efficient optical communication,” Communications, combinations, http://mentor.ieee.org/ 802.15/dcn/11/15-11-
IEEE Transactions on, vol. 47, no. 8, pp. 1201–1210, Aug 1999. 0247-00-0007-csk-constellati on-in-all-color-bandcombinations.
[61] T. Ohtsuki, I. Sasase, and S. Mori, “Differential overlapping pdf.
pulse position modulation in optical direct-detection channel,” [82] CIE 1931 Chromaticity diagram. [Online].
in Communications, 1994. ICC ’94, SUPERCOMM/ICC ’94, Con- Available: http://commons.wikimedia.org/wiki/File:Cie
ference Record, ’Serving Humanity Through Communications.’ IEEE chromaticity diagram wavelength.png
International Conference on, May 1994, pp. 680–684 vol.2. [83] J. Vucic, C. Kottke, K. Habel, and K.-D. Langer, “803 Mbit/s
[62] M. Noshad and M. Brandt-Pearce, “Expurgated PPM Using visible light WDM link based on DMT modulation of a single
Symmetric Balanced Incomplete Block Designs,” Communications RGB LED luminary,” in Optical Fiber Communication Conference
Letters, IEEE, vol. 16, no. 7, pp. 968–971, July 2012. and Exposition (OFC/NFOEC), 2011 and the National Fiber Optic
[63] ——, “Application of Expurgated PPM to Indoor Visible Light Engineers Conference, March 2011, pp. 1–3.
Communications; Part I: Single-User Systems,” Lightwave Tech- [84] F.-M. Wu, C.-T. Lin, C.-C. Wei, C.-W. Chen, Z.-Y. Chen, and
nology, Journal of, vol. 32, no. 5, pp. 875–882, March 2014. K. Huang, “3.22-Gb/s WDM Visible Light Communication of a
[64] ——, “Multilevel pulse-position modulation based on balanced Single RGB LED Employing Carrier-Less Amplitude and Phase
incomplete block designs,” in Global Communications Conference Modulation,” in Optical Fiber Communication Conference/National
(GLOBECOM), 2012 IEEE, Dec 2012, pp. 2930–2935. Fiber Optic Engineers Conference 2013. Optical Society of
[65] M. Afgani, H. Haas, H. Elgala, and D. Knipp, “Visible light America, 2013, p. OTh1G.4. [Online]. Available: http://www.
communication using OFDM,” in Testbeds and Research Infras- osapublishing.org/abstract.cfm?URI=OFC-2013-OTh1G.4
tructures for the Development of Networks and Communities, 2006. [85] L. Zeng, D. O’brien, H. Minh, G. Faulkner, K. Lee, D. Jung,
TRIDENTCOM 2006. 2nd International Conference on, 2006, pp. 6 Y. Oh, and E. T. Won, “High data rate multiple input multiple
pp.–134. output (MIMO) optical wireless communications using white
[66] J. Armstrong and A. Lowery, “Power efficient optical OFDM,” led lighting,” Selected Areas in Communications, IEEE Journal on,
Electronics Letters, vol. 42, no. 6, pp. 370–372, March 2006. vol. 27, no. 9, pp. 1654–1662, December 2009.
[67] H. Elgala, R. Mesleh, H. Haas, and B. Pricope, “OFDM Visible [86] D. O’Brien, “Multi-input multi-output (MIMO) indoor optical
Light Wireless Communication Based on White LEDs,” in Ve- wireless communications,” in Signals, Systems and Computers,
hicular Technology Conference, 2007. VTC2007-Spring. IEEE 65th, 2009 Conference Record of the Forty-Third Asilomar Conference on,
April 2007, pp. 2185–2189. Nov 2009, pp. 1636–1639.
[87] D. Tsonev, S. Sinanovic, and H. Haas, “Practical MIMO Capacity
[68] Y. Tanaka, T. Komine, S. Haruyama, and M. Nakagawa, “In-
for Indoor Optical Wireless Communication with White LEDs,”
door visible light data transmission system utilizing white LED
in Vehicular Technology Conference (VTC Spring), 2013 IEEE 77th,
lights,” IEICE transactions on communications, vol. 86, no. 8, pp.
June 2013, pp. 1–5.
2440–2454, 2003.
[88] J. C. Chau and T. D. Little, “Scalable imaging VLC receivers
[69] R. Mesleh, H. Elgala, and H. Haas, “Performance analysis
with token-based pixel selection for spatial multiplexing,” in
of indoor OFDM optical wireless communication systems,” in
Proceedings of the 1st ACM MobiCom workshop on Visible light
Wireless Communications and Networking Conference (WCNC), 2012
communication systems. ACM, 2014, pp. 21–26.
IEEE, April 2012, pp. 1005–1010.
[89] T. Fath and H. Haas, “Performance Comparison of MIMO Tech-
[70] H. Burchardt, N. Serafimovski, D. Tsonev, S. Videv, and H. Haas,
niques for Optical Wireless Communications in Indoor Environ-
“VLC: Beyond point-to-point communication,” Communications
ments,” Communications, IEEE Transactions on, vol. 61, no. 2, pp.
Magazine, IEEE, vol. 52, no. 7, pp. 98–105, July 2014.
733–742, February 2013.
[71] H. Elgala, R. Mesleh, and H. Haas, “A study of LED nonlinearity [90] D. Takase and T. Ohtsuki, “Optical wireless MIMO communica-
effects on optical wireless transmission using OFDM,” in Wireless tions (OMIMO),” in Global Telecommunications Conference, 2004.
and Optical Communications Networks, 2009. WOCN ’09. IFIP GLOBECOM ’04. IEEE, vol. 2, Nov 2004, pp. 928–932 Vol.2.
International Conference on, April 2009, pp. 1–5. [91] ——, “Performance analysis of optical wireless MIMO with
[72] ——, “Practical considerations for indoor wireless optical system optical beat interference,” in Communications, 2005. ICC 2005.
implementation using OFDM,” in Telecommunications, 2009. Con- 2005 IEEE International Conference on, vol. 2, May 2005, pp. 954–
TEL 2009. 10th International Conference on, June 2009, pp. 25–29. 958 Vol. 2.
[73] Z. Wang, C. Yu, W.-D. Zhong, J. Chen, and W. Chen, “Perfor- [92] ——, “Spatial multiplexing in optical wireless MIMO communi-
mance of variable M-QAM OFDM visible light communication cations over indoor environment,” IEICE transactions on commu-
system with dimming control,” in Opto-Electronics and Commu- nications, vol. 89, no. 4, pp. 1364–1371, 2006.
nications Conference (OECC), 2012 17th, July 2012, pp. 741–742. [93] Y. Chau and S.-H. Yu, “Space modulation on wireless fading
[74] A. Khalid, G. Cossu, R. Corsini, P. Choudhury, and E. Ciaramella, channels,” in Vehicular Technology Conference, 2001. VTC 2001 Fall.
“1-Gb/s Transmission Over a Phosphorescent White LED by IEEE VTS 54th, vol. 3, 2001, pp. 1668–1671 vol.3.
Using Rate-Adaptive Discrete Multitone Modulation,” Photonics [94] R. Mesleh, H. Elgala, and H. Haas, “Optical Spatial Modulation,”
Journal, IEEE, vol. 4, no. 5, pp. 1465–1473, Oct 2012. Optical Communications and Networking, IEEE/OSA Journal of,
[75] D. Tsonev, H. Chun, S. Rajbhandari, J. McKendry, S. Videv, E. Gu, vol. 3, no. 3, pp. 234–244, March 2011.
M. Haji, S. Watson, A. Kelly, G. Faulkner, M. Dawson, H. Haas, [95] C. Vladeanu, “Turbo trellis-coded spatial modulation,” in Global
and D. O’Brien, “A 3-Gb/s Single-LED OFDM-Based Wireless Communications Conference (GLOBECOM), 2012 IEEE, Dec 2012,
VLC Link Using a Gallium Nitride LED,” Photonics Technology pp. 4024–4029.
Letters, IEEE, vol. 26, no. 7, pp. 637–640, April 2014. [96] T. Fath, H. Haas, M. Di Renzo, and R. Mesleh, “Spatial Modula-
[76] R. Drost and B. Sadler, “Constellation design for color-shift tion Applied to Optical Wireless Communications in Indoor LOS
keying using billiards algorithms,” in GLOBECOM Workshops Environments,” in Global Telecommunications Conference (GLOBE-
(GC Wkshps), 2010 IEEE, Dec 2010, pp. 980–984. COM 2011), 2011 IEEE, Dec 2011, pp. 1–5.
[77] E. Monteiro and S. Hranilovic, “Constellation design for color- [97] P. Butala, H. Elgala, and T. Little, “Performance of optical spatial
shift keying using interior point methods,” in Globecom Work- modulation and spatial multiplexing with imaging receiver,” in
shops (GC Wkshps), 2012 IEEE, Dec 2012, pp. 1224–1228. Wireless Communications and Networking Conference (WCNC), 2014
[78] ——, “Design and implementation of color-shift keying for IEEE, April 2014, pp. 394–399.
visible light communications,” Lightwave Technology, Journal of, [98] T. Fath, M. Di Renzo, and H. Haas, “On the performance of
vol. 32, no. 10, pp. 2053–2060, May 2014. Space Shift Keying for optical wireless communications,” in
[79] R. Singh, T. O’Farrell, and J. David, “An Enhanced Color Shift GLOBECOM Workshops (GC Wkshps), 2010 IEEE, Dec 2010, pp.
Keying Modulation Scheme for High-Speed Wireless Visible 990–994.
Light Communications,” Lightwave Technology, Journal of, vol. 32, [99] M. Di Renzo and H. Haas, “On the performance of Space Shift
no. 14, pp. 2582–2592, July 2014. Keying MIMO systems over correlated Rician fading channels,”

in Smart Antennas (WSA), 2010 International ITG Workshop on, Feb munications Conference (GLOBECOM), 2013 IEEE, Dec 2013, pp.
2010, pp. 72–79. 2436–2441.
[100] S. Videv and H. Haas, “Practical space shift keying VLC system,” [118] J. Martin-Gonzalez, E. Poves, and F. Lopez-Hernandez, “Random
in Wireless Communications and Networking Conference (WCNC), optical codes used in optical networks,” Communications, IET,
2014 IEEE, April 2014, pp. 405–409. vol. 3, no. 8, pp. 1392–1401, August 2009.
[101] M. Di Renzo and H. Haas, “Space Shift Keying (SSK) Modulation [119] M. Guerra-Medina, B. Rojas-Guillama, O. Gonzalez, J. Martin-
with Partial Channel State Information: Optimal Detector and Gonzalez, E. Poves, and F. Lopez-Hernandez, “Experimental op-
Performance Analysis over Fading Channels,” Communications, tical code-division multiple access system for visible light com-
IEEE Transactions on, vol. 58, no. 11, pp. 3196–3210, November munications,” in Wireless Telecommunications Symposium (WTS),
2010. 2011, April 2011, pp. 1–6.
[102] W. Popoola, E. Poves, and H. Haas, “Generalised space shift [120] J. Liu, Z. Pan, W. Noonpakdee, and S. Shimamoto, “Impact
keying for visible light communications,” in Communication of light reflection on indoor wireless optical CDMA systems,”
Systems, Networks Digital Signal Processing (CSNDSP), 2012 8th Optical Communications and Networking, IEEE/OSA Journal of,
International Symposium on, July 2012, pp. 1–4. vol. 4, no. 12, pp. 989–996, Dec 2012.
[103] T. Fath, M. Di Renzo, and H. Haas, “On the performance of [121] J. Lian, M. Noshad, and M. Brandt-Pearce, “Multiuser MISO
Space Shift Keying for optical wireless communications,” in indoor visible light communications,” in Signals, Systems and
GLOBECOM Workshops (GC Wkshps), 2010 IEEE, Dec 2010, pp. Computers, 2014 48th Asilomar Conference on, Nov 2014, pp. 1729–
990–994. 1733.
[104] D. Takase and T. Ohtsuki, “Optical Wireless MIMO (OMIMO) [122] S.-H. Chen and C.-W. Chow, “Color-Shift Keying and Code-
with Backward Spatial Filter (BSF) in Diffuse Channels,” in Division Multiple-Access Transmission for RGB-LED Visible
Communications, 2007. ICC ’07. IEEE International Conference on, Light Communications Using Mobile Phone Camera,” Photonics
June 2007, pp. 2462–2467. Journal, IEEE, vol. 6, no. 6, pp. 1–6, Dec 2014.
[105] S.-M. Kim and S.-M. Kim, “Performance improvement of visible [123] J. Luna-Rivera, R. Perez-Jimenez, J. Rabadan-Borjes, J. Rufo-
light communications using optical beamforming,” in Ubiquitous Torres, V. Guerra, and C. Suarez-Rodriguez, “Multiuser scheme
and Future Networks (ICUFN), 2013 Fifth International Conference for indoor visible light communications using RGB LEDs,” in
on, July 2013, pp. 362–365. Bio-inspired Intelligence (IWOBI), 2014 International Work Confer-
[106] L. Wu, Z. Zhang, and H. Liu, “Transmit Beamforming for MIMO ence on, July 2014, pp. 119–123.
Optical Wireless Communication Systems,” Wireless Personal [124] C. Yang, Y. Wang, Y. Wang, X. Huang, and N. Chi,
Communications, vol. 78, no. 1, pp. 615–628, 2014. [Online]. “Demonstration of high-speed multi-user multi-carrier CDMA
Available: http://dx.doi.org/10.1007/s11277-014-1774-3 visible light communication,” Optics Communications, vol. 336,
no. 0, pp. 269 – 272, 2015. [Online]. Available: http://www.
[107] S. Schmid, G. Corbellini, S. Mangold, and T. R. Gross,
“LED-to-LED Visible Light Communication Networks,” in sciencedirect.com/science/article/pii/S0030401814009018
Proceedings of the Fourteenth ACM International Symposium on [125] L. L. Hanzo, M. Münster, B. Choi, and T. Keller, OFDM and
MC-CDMA for broadband multi-user communications, WLANs and
Mobile Ad Hoc Networking and Computing, ser. MobiHoc ’13.
New York, NY, USA: ACM, 2013, pp. 1–10. [Online]. Available: broadcasting. John Wiley & Sons, 2005.
http://doi.acm.org/10.1145/2491288.2491293 [126] Z. Yu, R. Baxley, and G. Zhou, “Multi-user MISO broadcasting
for indoor visible light communication,” in Acoustics, Speech and
[108] Q. Wang and D. Giustiniano, “Communication Networks of
Signal Processing (ICASSP), 2013 IEEE International Conference on,
Visible Light Emitting Diodes with Intra-Frame Bidirectional
May 2013, pp. 4849–4853.
Transmission,” in Proceedings of the 10th ACM International on
[127] J. Chen, Y. Hong, and Z. Wang, “Performance of precoding multi-
Conference on Emerging Networking Experiments and Technologies,
user MIMO indoor visible light communications,” in Proc. of
ser. CoNEXT ’14. New York, NY, USA: ACM, 2014, pp. 21–
IEEE Photonics Conference (IPC), 2013, pp. 541–542.
28. [Online]. Available: http://doi.acm.org/10.1145/2674005.
[128] Y. Hong, J. Chen, Z. Wang, and C. Yu, “Performance of a
2675000
Precoding MIMO System for Decentralized Multiuser Indoor
[109] J. Dang and Z. Zhang, “Comparison of optical OFDM-IDMA Visible Light Communications,” Photonics Journal, IEEE, vol. 5,
and optical OFDMA for uplink visible light communications,” no. 4, pp. 7 800 211–7 800 211, Aug 2013.
in Wireless Communications Signal Processing (WCSP), 2012 Inter- [129] J. Chen, N. Ma, Y. Hong, and C. Yu, “On the performance of MU-
national Conference on, Oct 2012, pp. 1–6. MIMO indoor visible light communication system based on THP
[110] D. Bykhovsky and S. Arnon, “Multiple Access Resource Al- algorithm,” in Communications in China (ICCC), 2014 IEEE/CIC
location in Visible Light Communication Systems,” Lightwave International Conference on, Oct 2014, pp. 136–140.
Technology, Journal of, vol. 32, no. 8, pp. 1594–1600, April 2014. [130] H. Haas, PureLiFi Attocell, www.purelifi.com.
[111] C. Chen, D. Tsonev, and H. Haas, “Joint transmission in indoor [131] Z. Wang, C. Yu, W.-D. Zhong, J. Chen, and W. Chen,
visible light communication downlink cellular networks,” in “Performance of a novel LED lamp arrangement to reduce
Globecom Workshops (GC Wkshps), 2013 IEEE, Dec 2013, pp. 1127– SNR fluctuation for multi-user visible light communication
1132. systems,” Opt. Express, vol. 20, no. 4, pp. 4564–4573,
[112] J. Salehi, “Code division multiple-access techniques in opti- Feb 2012. [Online]. Available: http://www.opticsexpress.org/
cal fiber networks. I. Fundamental principles,” Communications, abstract.cfm?URI=oe-20-4-4564
IEEE Transactions on, vol. 37, no. 8, pp. 824–833, Aug 1989. [132] Z. Wang, W.-D. Zhong, C. Yu, and J. Chen, “A novel LED
[113] J. Salehi and C. Brackett, “Code division multiple-access tech- arrangement to reduce SNR fluctuation for multi-user in visible
niques in optical fiber networks. II. Systems performance anal- light communication systems,” in Information, Communications
ysis,” Communications, IEEE Transactions on, vol. 37, no. 8, pp. and Signal Processing (ICICS) 2011 8th International Conference on,
834–842, Aug 1989. Dec 2011, pp. 1–4.
[114] M. Guerra-Medina, O. Gonzalez, B. Rojas-Guillama, J. Martin- [133] G. Prince and T. Little, “On the Performance Gains of Cooper-
Gonzalez, F. Delgado, and J. Rabadan, “Ethernet-OCDMA sys- ative Transmission Concepts in Intensity Modulated Direct De-
tem for multi-user visible light communications,” Electronics tection Visible Light Communication Networks,” in Wireless and
Letters, vol. 48, no. 4, pp. 227–228, February 2012. Mobile Communications (ICWMC), 2010 6th International Conference
[115] P. Kumar, R. Omrani, J. Touch, A. Willner, and P. Saghari, “A on, Sept 2010, pp. 297–302.
Novel Optical CDMA Modulation Scheme: Code Cycle Modula- [134] G. Marsh and J. Kahn, “Channel reuse strategies for indoor
tion,” in Global Telecommunications Conference, 2006. GLOBECOM infrared wireless communications,” in Global Telecommunications
’06. IEEE, Nov 2006, pp. 1–5. Conference, 1996. GLOBECOM ’96. ’Communications: The Key to
[116] M. Noshad and M. Brandt-Pearce, “Application of Expurgated Global Prosperity, vol. 3, Nov 1996, pp. 1597–1602 vol.3.
PPM to Indoor Visible Light Communications—Part II: Access [135] B. Ghimire and H. Haas, “Self-organising interference coordina-
Networks,” J. Lightwave Technol., vol. 32, no. 5, pp. 883–890, Mar tion in optical wireless networks,” EURASIP Journal on Wireless
2014. [Online]. Available: http://jlt.osa.org/abstract.cfm?URI= Communications and Networking, vol. 2012, no. 1, 2012. [Online].
jlt-32-5-883 Available: http://dx.doi.org/10.1186/1687-1499-2012-131
[117] ——, “High-speed visible light indoor networks based on optical [136] C. Chen, D. Tsonev, and H. Haas, “Joint transmission in indoor
orthogonal codes and combinatorial designs,” in Global Com- visible light communication downlink cellular networks,” in

Globecom Workshops (GC Wkshps), 2013 IEEE, Dec 2013, pp. 1127– vol. 32, no. 14, pp. 2480–2485, Jul 2014. [Online]. Available:
1132. http://jlt.osa.org/abstract.cfm?URI=jlt-32-14-2480
[137] Q. Wang, D. Giustiniano, and D. Puccinelli, “OpenVLC: [158] Bytelight - Scalable Indoor Localization with LED,
Software-defined Visible Light Embedded Networks,” in www.bytelight.com.
Proceedings of the 1st ACM MobiCom Workshop on Visible [159] A. Ashok, M. Gruteser, N. Mandayam, J. Silva, M. Varga, and
Light Communication Systems, ser. VLCS ’14. New York, K. Dana, “Challenge: Mobile Optical Networks Through Visual
NY, USA: ACM, 2014, pp. 15–20. [Online]. Available: MIMO,” in Proceedings of the Sixteenth Annual International
http://doi.acm.org/10.1145/2643164.2643167 Conference on Mobile Computing and Networking, ser. MobiCom
[138] BeagleBone Black. [Online]. Available: http://beagleboard.org/ ’10. New York, NY, USA: ACM, 2010, pp. 105–112. [Online].
[139] L. P. Klaver, “Design of a network stack for directional visible Available: http://doi.acm.org/10.1145/1859995.1860008
light communication,” Ph.D. dissertation, Delft University of [160] A. Ashok, S. Jain, M. Gruteser, N. Mandayam, W. Yuan, and
Technology, 2014. K. Dana, “Capacity of pervasive camera based communica-
[140] Atmel ATmega328P. [Online]. Available: http://www.atmel. tion under perspective distortions,” in Pervasive Computing and
com/tools/mega328p-xmini.aspx Communications (PerCom), 2014 IEEE International Conference on,
[141] Raspberry Pi. [Online]. Available: https://www.raspberrypi. March 2014, pp. 112–120.
org/ [161] S. D. Perli, N. Ahmed, and D. Katabi, “PixNet: Interference-
[142] Arduino. [Online]. Available: https://www.arduino.cc/ free Wireless Links Using LCD-camera Pairs,” in Proceedings
[143] WARP project. [Online]. Available: http://warpproject.org of the Sixteenth Annual International Conference on Mobile
[144] G. W. Jialiang Zhang, Xinyu Zhang, “Dancing with Light: Pre- Computing and Networking, ser. MobiCom ’10. New York,
dictive In-frame Rate Selection for Visible Light Networks,” in NY, USA: ACM, 2010, pp. 137–148. [Online]. Available:
IEEE Conference on Computer Communications (INFOCOM), 2015. http://doi.acm.org/10.1145/1859995.1860012
[145] E. K. Yijun Qiao, Harald Haas, “A Software-defined Visible Light [162] T. Hao, R. Zhou, and G. Xing, “COBRA: Color Barcode
Communications System with WARP,” in 1st ACM Workshop on Streaming for Smartphone Systems,” in Proceedings of the 10th
Visible Light Communication Systems, 2014. International Conference on Mobile Systems, Applications, and
[146] Q. Wang, D. Giustiniano, and D. Puccinelli, “An Open-Source Services, ser. MobiSys ’12. New York, NY, USA: ACM, 2012, pp.
Research Platform for Embedded Visible Light Networking,” 85–98. [Online]. Available: http://doi.acm.org/10.1145/2307636.
arXiv preprint arXiv:1410.0328, 2014. 2307645
[147] WikiStart - GNU Radio. [Online]. Available: http://gnuradio. [163] W. Hu, H. Gu, and Q. Pu, “Lightsync: Unsynchronized visual
org/redmine/projects/gnuradio/wiki communication over screen-camera links,” in Proceedings of the
[148] Universal Softwaere Radio Platform. [Online]. Available: http: 19th annual international conference on Mobile computing & network-
//www.ettus.com/ ing. ACM, 2013, pp. 15–26.
[149] M. Rahaim, A. Miravakili, T. Borogovac, T. Little, and V. Joyner, [164] R. LiKamWa, D. Ramirez, and J. Holloway, “Styrofoam: a tightly
“Demonstration of a software Defined Visible Light Communica- packed coding scheme for camera-based visible light communi-
tion System,” in the 17th Annual International Conference on Mobile cation,” in Proceedings of the 1st ACM MobiCom workshop on Visible
Computing and Networking, Mobicom2011, 2011. light communication systems. ACM, 2014, pp. 27–32.
[150] M. Rahaim, A. Miravakili, S. Ray, V. Koomson, M. Hella, and [165] T. Li, C. An, A. T. Campbell, and X. Zhou, “Hilight: Hiding bits in
T. Little, “Software defined visible light communication,” in pixel translucency changes,” ACM SIGMOBILE Mobile Computing
Wireless Innovation Forum Conference on Communications Technolo- and Communications Review, vol. 18, no. 3, pp. 62–70, 2015.
gies and Software Defined Radio (WInnComm SDR), 2014. [166] A. Wang, S. Ma, C. Hu, J. Huai, C. Peng, and G. Shen,
[151] K. Chintalapudi, A. Padmanabha Iyer, and V. N. Padmanabhan, “Enhancing Reliability to Boost the Throughput over Screen-
“Indoor Localization Without the Pain,” in Proceedings of the camera Links,” in Proceedings of the 20th Annual International
Sixteenth Annual International Conference on Mobile Computing Conference on Mobile Computing and Networking, ser. MobiCom
and Networking, ser. MobiCom ’10. New York, NY, USA: ACM, ’14. New York, NY, USA: ACM, 2014, pp. 41–52. [Online].
2010, pp. 173–184. [Online]. Available: http://doi.acm.org/10. Available: http://doi.acm.org/10.1145/2639108.2639135
1145/1859995.1860016 [167] W. Hu, J. Mao, Z. Huang, Y. Xue, J. She, K. Bian, and G. Shen,
[152] S. Sen, J. Lee, K.-H. Kim, and P. Congdon, “Avoiding “Strata: Layered Coding for Scalable Visual Communication,”
Multipath to Revive Inbuilding WiFi Localization,” in Proceeding in Proceedings of the 20th Annual International Conference on
of the 11th Annual International Conference on Mobile Systems, Mobile Computing and Networking, ser. MobiCom ’14. New
Applications, and Services, ser. MobiSys ’13. New York, York, NY, USA: ACM, 2014, pp. 79–90. [Online]. Available:
NY, USA: ACM, 2013, pp. 249–262. [Online]. Available: http://doi.acm.org/10.1145/2639108.2639132
http://doi.acm.org/10.1145/2462456.2464463 [168] B. Zhang, K. Ren, G. Xing, X. Fu, and C. Wang, “SBVLC: Secure
[153] L. Li, P. Hu, C. Peng, G. Shen, and F. Zhao, “Epsilon: barcode-based visible light communication for smartphones,” in
A Visible Light Based Positioning System,” in 11th USENIX INFOCOM, 2014 Proceedings IEEE, April 2014, pp. 2661–2669.
Symposium on Networked Systems Design and Implementation [169] S. Arai, S. Mase, T. Yamazato, T. Endo, T. Fujii, M. Tanimoto,
(NSDI 14). Seattle, WA: USENIX Association, Apr. 2014, K. Kidono, Y. Kimura, and Y. Ninomiya, “Experimental on Hier-
pp. 331–343. [Online]. Available: https://www.usenix.org/ archical Transmission Scheme for Visible Light Communication
conference/nsdi14/technical-sessions/presentation/li using LED Traffic Light and High-Speed Camera,” in Vehicular
[154] Y.-S. Kuo, P. Pannuto, K.-J. Hsiao, and P. Dutta, “Luxapose: Technology Conference, 2007. VTC-2007 Fall. 2007 IEEE 66th, Sept
Indoor Positioning with Mobile Phones and Visible Light,” 2007, pp. 2174–2178.
in Proceedings of the 20th Annual International Conference on [170] S. Arai, Y. Shiraki, T. Yamazato, H. Okada, T. Fujii, and T. Yendo,
Mobile Computing and Networking, ser. MobiCom ’14. New “Multiple LED arrays acquisition for image-sensor-based I2V-
York, NY, USA: ACM, 2014, pp. 447–458. [Online]. Available: VLC using block matching,” in Consumer Communications and
http://doi.acm.org/10.1145/2639108.2639109 Networking Conference (CCNC), 2014 IEEE 11th, Jan 2014, pp. 605–
[155] N. Rajagopal, P. Lazik, and A. Rowe, “Visual Light Landmarks 610.
for Mobile Devices,” in Proceedings of the 13th International [171] C. Li and S. Shimamoto, “An Open Traffic Light Control Model
Symposium on Information Processing in Sensor Networks, ser. for Reducing Vehicles CO2 Emissions Based on ETC Vehicles,”
IPSN ’14. Piscataway, NJ, USA: IEEE Press, 2014, pp. 249–260. Vehicular Technology, IEEE Transactions on, vol. 61, no. 1, pp. 97–
[Online]. Available: http://dl.acm.org/citation.cfm?id=2602339. 110, Jan 2012.
2602367 [172] C. B. Liu, B. Sadeghi, and E. W. Knightly, “Enabling
[156] G. Prince and T. Little, “A two phase hybrid RSS/AoA algorithm Vehicular Visible Light Communication (V2LC) Networks,”
for indoor device localization using visible light,” in Global in Proceedings of the Eighth ACM International Workshop
Communications Conference (GLOBECOM), 2012 IEEE, Dec 2012, on Vehicular Inter-networking, ser. VANET ’11. New York,
pp. 3347–3352. NY, USA: ACM, 2011, pp. 41–50. [Online]. Available: http:
[157] S.-H. Yang, H.-S. Kim, Y.-H. Son, and S.-K. Han, “Three- //doi.acm.org/10.1145/2030698.2030705
Dimensional Visible Light Indoor Localization Using AOA and [173] L.-C. Wu and H.-M. Tsai, “Modeling vehicle-to-vehicle visible
RSS With Multiple Optical Receivers,” J. Lightwave Technol., light communication link duration with empirical data,” in

Globecom Workshops (GC Wkshps), 2013 IEEE, Dec 2013, pp. 1103– [189] H. Ma, L. Lampe, and S. Hranilovic, “Integration of indoor
1109. visible light and power line communication systems,” in Power
[174] S.-H. You, S.-H. Chang, H.-M. Lin, and H.-M. Tsai, “Visible Line Communications and Its Applications (ISPLC), 2013 17th IEEE
Light Communications for Scooter Safety,” in Proceeding of International Symposium on, March 2013, pp. 291–296.
the 11th Annual International Conference on Mobile Systems, [190] M. Yousuf and M. El-Shafei, “Power Line Communications: An
Applications, and Services, ser. MobiSys ’13. New York, Overview - Part I,” in Innovations in Information Technology, 2007.
NY, USA: ACM, 2013, pp. 509–510. [Online]. Available: IIT ’07. 4th International Conference on, Nov 2007, pp. 218–222.
http://doi.acm.org/10.1145/2462456.2465707 [191] M. Rahaim, A. Vegni, and T. Little, “A hybrid Radio Frequency
[175] S.-H. Yu, O. Shih, H.-M. Tsai, N. Wisitpongphan, and R. Roberts, and broadcast Visible Light Communication system,” in GLOBE-
“Smart automotive lighting for vehicle safety,” Communications COM Workshops (GC Wkshps), 2011 IEEE, Dec 2011, pp. 792–796.
Magazine, IEEE, vol. 51, no. 12, pp. 50–59, December 2013. [192] S. Shao, A. Khreishah, M. Rahaim, H. Elgala, M. Ayyash, T. Little,
[176] I. Takai, S. Ito, K. Yasutomi, K. Kagawa, M. Andoh, and and J. Wu, “An Indoor Hybrid WiFi-VLC Internet Access Sys-
S. Kawahito, “LED and CMOS Image Sensor Based Optical tem,” in Mobile Ad Hoc and Sensor Systems (MASS), 2014 IEEE
Wireless Communication System for Automotive Applications,” 11th International Conference on, Oct 2014, pp. 569–574.
Photonics Journal, IEEE, vol. 5, no. 5, pp. 6 801 418–6 801 418, Oct [193] C. Lee, C. Tan, H. Wong, and M. Yahya, “Performance evaluation
2013. of hybrid VLC using device cost and power over data through-
[177] T. Yamazato, I. Takai, H. Okada, T. Fujii, T. Yendo, S. Arai, put criteria,” in SPIE Optical Engineering+ Applications. Inter-
M. Andoh, T. Harada, K. Yasutomi, K. Kagawa, and S. Kawahito, national Society for Optics and Photonics, 2013, pp. 88 451A–
“Image-sensor-based visible light communication for automotive 88 451A.
applications,” Communications Magazine, IEEE, vol. 52, no. 7, pp. [194] K.-D. Langer and J. Grubor, “Recent Developments in Optical
88–97, July 2014. Wireless Communications using Infrared and Visible Light,” in
[178] Y. Zeng, P. H. Pathak, C. Xu, and P. Mohapatra, “Your Transparent Optical Networks, 2007. ICTON ’07. 9th International
AP Knows How You Move: Fine-grained Device Motion Conference on, vol. 3, July 2007, pp. 146–151.
Recognition Through WiFi,” in Proceedings of the 1st ACM [195] Y.-S. Kuo, P. Pannuto, and P. Dutta, “System Architecture
Workshop on Hot Topics in Wireless, ser. HotWireless ’14. New Directions for a Software-defined Lighting Infrastructure,” in
York, NY, USA: ACM, 2014, pp. 49–54. [Online]. Available: Proceedings of the 1st ACM MobiCom Workshop on Visible
http://doi.acm.org/10.1145/2643614.2643620 Light Communication Systems, ser. VLCS ’14. New York,
[179] B. Kellogg, V. Talla, and S. Gollakota, “Bringing Gesture NY, USA: ACM, 2014, pp. 3–8. [Online]. Available: http:
Recognition to All Devices,” in 11th USENIX Symposium //doi.acm.org/10.1145/2643164.2643166
on Networked Systems Design and Implementation (NSDI 14). [196] D. O’brien, L. Zeng, H. Le-Minh, G. Faulkner, J. W. Walewski,
Seattle, WA: USENIX Association, Apr. 2014, pp. 303– and S. Randel, “Visible light communications: Challenges and
316. [Online]. Available: https://www.usenix.org/conference/ possibilities,” in Personal, Indoor and Mobile Radio Communications,
nsdi14/technical-sessions/presentation/kellogg 2008. PIMRC 2008. IEEE 19th International Symposium on. IEEE,
2008, pp. 1–5.
[180] J. Wang, D. Vasisht, and D. Katabi, “RF-IDraw: Virtual Touch
[197] D. O’Brien, “Cooperation in optical wireless communications,”
Screen in the Air Using RF Signals,” in Proceedings of the
in Cognitive Wireless Networks, F. Fitzek and M. Katz, Eds.
2014 ACM Conference on SIGCOMM, ser. SIGCOMM ’14. New
Springer Netherlands, 2007, pp. 623–634. [Online]. Available:
York, NY, USA: ACM, 2014, pp. 235–246. [Online]. Available:
http://dx.doi.org/10.1007/978-1-4020-5979-7 32
http://doi.acm.org/10.1145/2619239.2626330
[198] J. Ylitalo, T. Jokikyyny, T. Kauppinen, A. Tuominen, and J. Laine,
[181] Microsoft Kinect Sensor. [Online]. Available: http://www.xbox.
“Dynamic network interface selection in multihomed mobile
com/en-US/xbox-one/accessories/kinect-for-xbox-one
hosts,” in System Sciences, 2003. Proceedings of the 36th Annual
[182] S. Gupta, K.-Y. Chen, M. S. Reynolds, and S. N. Patel, Hawaii International Conference on, Jan 2003, pp. 10–18.
“LightWave: Using Compact Fluorescent Lights As Sensors,” [199] H. Balakrishnan and V. Padmanabhan, “How network asymme-
in Proceedings of the 13th International Conference on Ubiquitous try affects TCP,” Communications Magazine, IEEE, vol. 39, no. 4,
Computing, ser. UbiComp ’11. New York, NY, USA: ACM, pp. 60–67, April 2001.
2011, pp. 65–74. [Online]. Available: http://doi.acm.org/10. [200] J. Classen, J. Chen, D. Steinmetzer, M. Hollick, and E. Knightly,
1145/2030112.2030122 “The Spy Next Door: Eavesdropping on High Throughput Vis-
[183] P. Dietz, W. Yerazunis, and D. Leigh, “Very Low-Cost Sensing ible Light Communications,” in Proceedings of the 2nd ACM
and Communication Using Bidirectional LEDs,” in UbiComp MobiCom Workshop on Visible Light Communication Systems, ser.
2003: Ubiquitous Computing, ser. Lecture Notes in Computer VLCS ’15. New York, NY, USA: ACM, 2015.
Science, A. Dey, A. Schmidt, and J. McCarthy, Eds. Springer
Berlin Heidelberg, 2003, vol. 2864, pp. 175–191. [Online].
Available: http://dx.doi.org/10.1007/978-3-540-39653-6 14
[184] D. Schmidt, D. Molyneaux, and X. Cao, “PICOntrol: Using
a Handheld Projector for Direct Control of Physical Devices
Through Visible Light,” in Proceedings of the 25th Annual ACM
Symposium on User Interface Software and Technology, ser. UIST
’12. New York, NY, USA: ACM, 2012, pp. 379–388. [Online].
Available: http://doi.acm.org/10.1145/2380116.2380166
[185] C. Zhang, J. Tabor, J. Zhang, and X. Zhang, “Extending Mo-
bile Interaction Through Near-Field Visible Light Sensing,” in Parth H. Pathak is a postdoctoral scholar in
Proceedings of the Twenty First Annual International Conference on Computer Science Department at University of
Mobile Computing and Networking, ser. MobiCom ’15. New York, California, Davis. He earned his doctoral degree
NY, USA: ACM, 2015. in Computer Science at North Carolina State
[186] X. Zhou and A. T. Campbell, “Visible light networking and University, Raleigh in 2012. His current research
sensing,” in Proceedings of the 1st ACM Workshop on Hot Topics interests are in design and development of next-
in Wireless, ser. HotWireless ’14. New York, NY, USA: ACM, generation wireless networks, network analyt-
2014. ics and mobile computing. In the past, he has
[187] T. Li, C. An, Z. Tian, A. T. Campbell, , and X. Zhou, “Human worked on cross-layer protocol design and opti-
Sensing using Visible Light Communication,” in Proceedings of the mization of wireless networks.
Twenty First Annual International Conference on Mobile Computing
and Networking, ser. MobiCom ’15. New York, NY, USA: ACM,
2015.
[188] C. An, T. Li, Z. Tian, A. T. Campbell, and X. Zhou, “Visible Light
Knows Who You Are,” in Proceedings of the 2nd ACM MobiCom
Workshop on Visible Light Communication Systems, ser. VLCS ’15.
New York, NY, USA: ACM, 2015.

Xiaotao Feng is currently a Ph.D. student in the

Department of Electrical and Computer Engi-
neering at University of California, Davis. He re-
ceived his B.S. degree and M.S. degree in Elec-
trical Engineering from Beijing Jiaotong Univer-
sity, China, in 2009 and Peking University, China,
in 2012, respectively. His current research inter-
ests include wireless networking, game theory
and its applications to the field of multi-agent
systems and cyber security.

Pengfei Hu is a Ph.D. student in the Department

of Computer Science at University of Califor-
nia, Davis. He received his B.S. degree from
Changchun University of Science and Technol-
ogy in 2010 and his M.S. from University of
Science and Technology of China in 2013. In
2012, he received National Scholarship which
is the top award for graduate students in China.
His research interests include mobile computing,
visible light communication, and security and
privacy.

Dr. Prasant Mohapatra is a Professor in the

Department of Computer Science and is cur-
rently serving as the Associate Chancellor of
the University of California, Davis. He was the
Department Chair of Computer Science during
2007-13, and held the Tim Bucher Family En-
dowed Chair Professorship during that period.
He served as the Interim Vice-Provost and the
Campus CIO of UC Davis during 2013-14. In
the past, he has been on the faculty at Iowa
State University and Michigan State University.
Dr. Mohapatra is the Editor-in-Chief of the IEEE Transactions on Mobile
Computing. He has served on the editorial board of the IEEE Trans-
actions on Computers, IEEE Transactions on Mobile Computing, IEEE
Transaction on Parallel and Distributed Systems, ACM WINET, and Ad
Hoc Networks. He has served as the Program Chair and the Gen-
eral Chair and has been on the program/organizational committees of
several international conferences.x Dr. Mohapatra received his doctoral
degree from Penn State University in 1993, and received an Outstanding
Engineering Alumni Award in 2008. Dr. Mohapatra’s research interests
are in the areas of wireless networks, mobile communications, cyberse-
curity, and Internet protocols.

A 12-m 2.5-Gb/s Lighting Compatible Integrated Receiver For OOK Visible Light Communication Links
No ratings yet
A 12-m 2.5-Gb/s Lighting Compatible Integrated Receiver For OOK Visible Light Communication Links
8 pages
Final Project Uthm (Group 2)
No ratings yet
Final Project Uthm (Group 2)
2 pages
D LIGHTpbrvits
No ratings yet
D LIGHTpbrvits
9 pages
VLC Challeng and Possibilities
No ratings yet
VLC Challeng and Possibilities
5 pages
40-2022-High-Speed Visible Light Communication Based On micro-LED - A Technology With Wide Applications in Next Generation Communication
No ratings yet
40-2022-High-Speed Visible Light Communication Based On micro-LED - A Technology With Wide Applications in Next Generation Communication
24 pages
Data Transmission Using LI-FI
No ratings yet
Data Transmission Using LI-FI
5 pages
Performance Investigation of Visible Light Communication System Using Optisystem Simulation Tool
No ratings yet
Performance Investigation of Visible Light Communication System Using Optisystem Simulation Tool
8 pages
A 40 Mbps VLC System Reusing An Existing Large LED Panel in An Indoor Office Environment
No ratings yet
A 40 Mbps VLC System Reusing An Existing Large LED Panel in An Indoor Office Environment
16 pages
A Dimmable Offline LED Driver With OOK-M-FSK Modulation For VLC Applications
No ratings yet
A Dimmable Offline LED Driver With OOK-M-FSK Modulation For VLC Applications
11 pages
Siddiqui 2015
No ratings yet
Siddiqui 2015
6 pages
Visible Light Communication Systems For Optical Video Transmission
No ratings yet
Visible Light Communication Systems For Optical Video Transmission
5 pages
BER Tool
No ratings yet
BER Tool
8 pages
An Overview of Visible Light
No ratings yet
An Overview of Visible Light
12 pages
Fin Irjmets1637943871
No ratings yet
Fin Irjmets1637943871
6 pages
Wireless Communication Using Li-Fi
No ratings yet
Wireless Communication Using Li-Fi
9 pages
Integrated Sensing, Lighting and Communication
No ratings yet
Integrated Sensing, Lighting and Communication
21 pages
Dataset Paper
No ratings yet
Dataset Paper
35 pages
L. Matheus - 2019 - Visible Light Communication Concepts, Applications and Challenges
No ratings yet
L. Matheus - 2019 - Visible Light Communication Concepts, Applications and Challenges
59 pages
Technical Seminar Report VTU
No ratings yet
Technical Seminar Report VTU
30 pages
Visible Light Communications: Challenges and Possibilities: October 2008
No ratings yet
Visible Light Communications: Challenges and Possibilities: October 2008
6 pages
Li-Fi Technology Seminar Report
No ratings yet
Li-Fi Technology Seminar Report
39 pages
VLC for ECE Students
No ratings yet
VLC for ECE Students
3 pages
World's Largest Science, Technology & Medicine Open Access Book Publisher
No ratings yet
World's Largest Science, Technology & Medicine Open Access Book Publisher
31 pages
Beee Working Model TOPIC: Li-Fi: Submitted To:-Sundeep Sir Submitted By
No ratings yet
Beee Working Model TOPIC: Li-Fi: Submitted To:-Sundeep Sir Submitted By
28 pages
Drp+mcee 117 125
No ratings yet
Drp+mcee 117 125
9 pages
A Survey On Visible Light Communication
No ratings yet
A Survey On Visible Light Communication
4 pages
World's Largest Science, Technology & Medicine Open Access Book Publisher
No ratings yet
World's Largest Science, Technology & Medicine Open Access Book Publisher
14 pages
Design and Outage Analysis For VLP-Assisted Indoor Laser Communication Systems REF11
No ratings yet
Design and Outage Analysis For VLP-Assisted Indoor Laser Communication Systems REF11
17 pages
Mini Project Report (P)
No ratings yet
Mini Project Report (P)
16 pages
A Small Step in VLC Systems - A Big Step in Li-Fi Implementation
No ratings yet
A Small Step in VLC Systems - A Big Step in Li-Fi Implementation
14 pages
Seminar Report: Department: Computer Science SESSION YEAR: 2018-19
No ratings yet
Seminar Report: Department: Computer Science SESSION YEAR: 2018-19
20 pages
Li Fi Light Fidelity Technology
No ratings yet
Li Fi Light Fidelity Technology
7 pages
Technical Seminar Report
No ratings yet
Technical Seminar Report
31 pages
Concept of Visible Light Communication and VIDAS: Navin Kumar, Luis A Nero, Rui L Aguiar
No ratings yet
Concept of Visible Light Communication and VIDAS: Navin Kumar, Luis A Nero, Rui L Aguiar
4 pages
Visible Light Communication OOK
No ratings yet
Visible Light Communication OOK
6 pages
On Optimizing VLC Networks For Downlink Multi-User Transmission: A Survey
No ratings yet
On Optimizing VLC Networks For Downlink Multi-User Transmission: A Survey
30 pages
Ajunraj M: Cochin University of Science & Technology School of Engineering, Kochi Division of Electronics
No ratings yet
Ajunraj M: Cochin University of Science & Technology School of Engineering, Kochi Division of Electronics
45 pages
M. Saadi, L. Wattisuttikulkij, Y. Zhao, P. Sangwongngam
No ratings yet
M. Saadi, L. Wattisuttikulkij, Y. Zhao, P. Sangwongngam
11 pages
Technical Seminar of Light Fidelity
No ratings yet
Technical Seminar of Light Fidelity
25 pages
Led Based Visible Light Communication Technology, Applications and Challenges - A Survey
No ratings yet
Led Based Visible Light Communication Technology, Applications and Challenges - A Survey
14 pages
Current Challenges For Visible Light Communications Usage in Vehicle Applications A Survey
No ratings yet
Current Challenges For Visible Light Communications Usage in Vehicle Applications A Survey
23 pages
LED-Based Visible Light Communications
100% (1)
LED-Based Visible Light Communications
256 pages
Visible Light Communication
No ratings yet
Visible Light Communication
19 pages
10.1007@s11277-019-06617-2 Optisystem
No ratings yet
10.1007@s11277-019-06617-2 Optisystem
18 pages
Unit-5 - S-7
No ratings yet
Unit-5 - S-7
38 pages
Li Fi PDF
100% (2)
Li Fi PDF
23 pages
Next Gen - Lifi
No ratings yet
Next Gen - Lifi
8 pages
Visible Light Communication, Sananda Nath 2124101
No ratings yet
Visible Light Communication, Sananda Nath 2124101
23 pages
Visible Light Communication
No ratings yet
Visible Light Communication
16 pages
Karanlifi Report
No ratings yet
Karanlifi Report
25 pages
VLC
No ratings yet
VLC
22 pages
Visible Light Communication
No ratings yet
Visible Light Communication
21 pages
Visible Light Communication For Wireless Data Transmission: Sathiya.T, Prof.E.Divya, Prof.S.Raja
No ratings yet
Visible Light Communication For Wireless Data Transmission: Sathiya.T, Prof.E.Divya, Prof.S.Raja
5 pages
Light Fidelity A New Prototype in Wireless
No ratings yet
Light Fidelity A New Prototype in Wireless
4 pages
Transmission of Data Using Li-Fi Technology
No ratings yet
Transmission of Data Using Li-Fi Technology
4 pages
Optimizing Indoor Visible Light Communication Systems: A Comparative Analysis of Multi-LED Configurations
No ratings yet
Optimizing Indoor Visible Light Communication Systems: A Comparative Analysis of Multi-LED Configurations
30 pages
Salian 2013
No ratings yet
Salian 2013
5 pages
AREA B - ALE Reviewer - Building Utilities
No ratings yet
AREA B - ALE Reviewer - Building Utilities
12 pages
Electrical Installation Guide 2007 - Schneider Electric
100% (6)
Electrical Installation Guide 2007 - Schneider Electric
432 pages
Solar Fan Project Lesson 2
No ratings yet
Solar Fan Project Lesson 2
14 pages
kjl-6000ss Operatorsmanual
No ratings yet
kjl-6000ss Operatorsmanual
4 pages
Airfield Lighting System: Runway End Identification Lights
No ratings yet
Airfield Lighting System: Runway End Identification Lights
7 pages
Fluorescent Lamp Guide & Troubleshooting
50% (2)
Fluorescent Lamp Guide & Troubleshooting
3 pages
Project On - Patent Infringement-Doctrines On Infringement: Hidayatullah National Law University Raipur, C.G
100% (1)
Project On - Patent Infringement-Doctrines On Infringement: Hidayatullah National Law University Raipur, C.G
18 pages
Roadway Lighting Design & Evaluation
No ratings yet
Roadway Lighting Design & Evaluation
12 pages
Electrical Equivalent of Heat: Instruction Manual and Experiment Guide For The PASCO Scientific Model TD-8552
No ratings yet
Electrical Equivalent of Heat: Instruction Manual and Experiment Guide For The PASCO Scientific Model TD-8552
14 pages
Physics Project Class 12th - Organized
100% (1)
Physics Project Class 12th - Organized
16 pages
Electronics q1 Week1 Mod1 Ligsay Jerry Lloyd
No ratings yet
Electronics q1 Week1 Mod1 Ligsay Jerry Lloyd
26 pages
Mil DTL 7788G
No ratings yet
Mil DTL 7788G
46 pages
16504-Home Automation System
No ratings yet
16504-Home Automation System
23 pages
Electrical Engineering
No ratings yet
Electrical Engineering
24 pages
22mm Mush PB
No ratings yet
22mm Mush PB
12 pages
Diagnostic Test-Electronics
100% (4)
Diagnostic Test-Electronics
4 pages
22530-2019-Winter-Model-Answer-Paper (Msbte Study Resources) (2) - Organized
No ratings yet
22530-2019-Winter-Model-Answer-Paper (Msbte Study Resources) (2) - Organized
13 pages
Vibgyor - Prelim 1 - Physics
No ratings yet
Vibgyor - Prelim 1 - Physics
13 pages
Sources of Light
No ratings yet
Sources of Light
6 pages
TZ-GK-10 Highbay Light
No ratings yet
TZ-GK-10 Highbay Light
1 page
Ohio TRM
No ratings yet
Ohio TRM
397 pages
Low-Cost Solar Emulator Design
No ratings yet
Low-Cost Solar Emulator Design
4 pages
CELMA ELC Guide On The Importance of Lighting June 2011
No ratings yet
CELMA ELC Guide On The Importance of Lighting June 2011
24 pages
LED Street Light - LU1: Brief
No ratings yet
LED Street Light - LU1: Brief
5 pages
Sapa's Innovations in Aluminum Profiles
No ratings yet
Sapa's Innovations in Aluminum Profiles
24 pages
Phy 2 Practical
No ratings yet
Phy 2 Practical
115 pages
Grade 7
No ratings yet
Grade 7
7 pages
GCSE Physics AQA Resistors Worksheet
No ratings yet
GCSE Physics AQA Resistors Worksheet
18 pages
Physics Project - Arush Final
No ratings yet
Physics Project - Arush Final
20 pages
Electricity and Electronics Ebook - January 2012 PDF
100% (1)
Electricity and Electronics Ebook - January 2012 PDF
113 pages

Visible Light Communication Survey

Uploaded by

Visible Light Communication Survey

Uploaded by

This article has been accepted for publication in a future issue of this journal, but has not been

Visible Light Communication, Networking and

Radio Microwave Infrared Visible Ultraviolet X-ray Gamma

105 m 1m 1 mm 750 nm 380 nm 10 nm 0.01 nm Wavelength

Red Orange Yellow Green Blue Violet

TABLE 2: Symbols and their meaning

Human eye sensitivity V(λ)

Normalized Radiant Power (%)

Luminous intensity (candela)

Luminous Intensity gt (θ): While luminous flux mea- Transmitter

sures the total amount of light emitted by an LED, the

of-view (FOV) can be expressed as

Normalized received power

Normalized received power

Normalized received power

1 N LEDs at time instance t as

transmitter and the receiver, the link performance is 100

Measured Light (%)

in terms of the DC balancing. as ON/OFF levels causes the LEDs to be operated

50% dimming DMT+PWM 0 1

TON p(t) y(t)=x(t)p(t) S1 S2 S3 S4

(a) (b) (c) (d)

Band i Band j Band k

000 be represented using the symbols for 4CSK, 8CSK

TABLE 5: 802.15.7 PHY I operating mode specifications and achievable throughput.

Optical Data rate

TABLE 6: 802.15.7 PHY II operating mode specifications and achievable throughput.

MCS Modulation Optical clock rate FEC Data rate (Mbps)

Modulation Data Rate Dimming Support Flickering Issue Comments

TABLE 8: Major modulation schemes and their characteristics

(a) Frame structure

Fig. 19: Rearranging LEDs can result in lower variance

that can be used to evaluate newly design protocols

6 S ENSING AND OTHER APPLICATIONS CT

The image sensor receiver available on today’s mobile

forms of sensing and applications. This sections provides

Fig. 25: Example of 2D COBRA barcode; reproduced

Xiaotao Feng is currently a Ph.D. student in the

Pengfei Hu is a Ph.D. student in the Department

Dr. Prasant Mohapatra is a Professor in the

You might also like