Optical Wireless Communications for In-Body and Transdermal Biomedical Applications

Abstract

This article discusses the fundamental architectures for optical-wireless-systems for biomedical-applications. After summarizing the main applications and reporting their requirements, we describe the characteristics of the transdermal and in-body optical-channels (OCs) as well as the challenges that they impose in the design of communication systems. Specifically, we provide three possible architectures for transdermal communications, namely electro-optical (EO) monitoring, optoelectrical (OE), and all-optical (AO) for neural stimulation, which are currently under investigation, whereas for in-body communications, we provide a nano-scale AO (NAO) concept. For each architecture, we discuss the main operation principles, the technology enablers, and research directions for their development. Finally, we highlight the necessity of designing an information-theoretic framework for the analysis and design of the physical (PHY) and medium access control (MAC) layers, which takes into account the channels’ characteristics.

Full text

1
Optical wireless communications for in-body
and transdermal biomedical applications
Alexandros-Apostolos A. Boulogeorgos (Senior Member, IEEE),
Stylianos E. Trevlakis (Student Member, IEEE), and Nestor D. Chatzidiamantis (Member, IEEE)
Abstract—This article discusses the fundamental architectures
for optical-wireless-systems for biomedical-applications. After
summarizing the main applications and reporting their require-
ments, we describe the characteristics of the transdermal and
in-body optical-channels (OCs) as well as the challenges that
they impose in the design of communication systems. Specifi-
cally, we provide three possible architectures for transdermal
communications, namely electro-optical (EO) monitoring, opto-
electrical (OE) and all-optical (AO) for neural stimulation,
which are currently under investigation, whereas for in-body
communications, we provide a nano-scale AO (NAO) concept.
For each architecture, we discuss the main operation principles,
the technology enablers and research directions for their de-
velopment. Finally, we highlight the necessity of designing an
information-theoretic framework for the analysis and design of
the physical (PHY) and medium access control (MAC) layers,
which takes into account the channels’ characteristics.
I. INTRODUCTION
Medical implants (MIs) and nano-scale wireless networks
(NWNs) have been advocated as an effective-solution to
numerous health-issues. Typical MIs consist of an out-of-
body (in-body) unit that captures the stimulus (bio-signal),
converts it into a RF signal and wirelessly transmits it into
the in-body (out-of-body) unit, which stimulates (monitors)
the corresponding nerve (bio-process). The main disadvantage
of this approach is its incapability to support high-data-rates
required for neural-prosthesis applications, under reasonable
transmission-power. Additionally, there is a lack of effec-
tiveness and flexibility in co-existing with the huge amount
of RF-devices that are expected to conquer the beyond 5G
wireless world. Finally, the use of RF-frequencies prohibit the
utilization of nano-scale biomedical-devices.
In view of the advances in optoelectronics and optogenetics,
very recent innovative designs have been reported, which, by
employing optical-wireless-communications (OWCs), deliver
more compact, reliable, and energy-efficient solutions, while,
at the same time, reducing the RF-radiation concerns [1].
The main advantages provided by OWCs are the abundant
available bandwidth, no interference from other devices, higher
achievable data-rates and increased safety. As a result, OWCs
are expected to be used for both transdermal and in-body
links establishment; hence, they will influence the fundamental
technology trends in biomedical-applications for the next 10
year and beyond. Due to the transdermal and in-body OC
particularities, the compactness and energy-limitations as well
The authors are with the Department of Electrical and Com-
puter Engineering in Aristotle University of Thessaloniki, Greece (E-
mails:al.boulogeorgos@ieee.org, {trevlakis, nestoras}@auth.gr).
as the directionality of the OWC links, the design and devel-
opment of OWC-based biomedical-applications needs to lever-
age breakthrough technological concepts. Indicative examples
are the co-design of communication and stimulation units,
the presentation of digital signal processing (DSP) schemes
for the end-to-end link from the device (biological unit) to
the biological unit (device), and the utilization of new EO,
OE, AO, and NAO interfaces. Likewise, new channel and
noise models that take into account the peculiarities of the
transdemal and in-body optical-medium need to be developed.
Building upon these, a novel information-theoretic framework
is required for the design of energy-efficient PHY and MAC
schemes, as well as the development of simultaneous light
information and power transfer (SLIPT) policies and resource
allocation strategies [2].
Motivated by the above, the aim of this article is to present
the vision and approach of delivering safe and high quality-
of-experience (QoE) in MIs and NWNs, identify the critical
technology gaps and the appropriate enables. In particular,
after presenting the biomedical-applications, which can be
OWC-enabled, together with their critical design requirements,
we introduce three possible architectures for transdermal com-
munications, which are currently under investigation, whereas
for in-body communications, we discuss the NAO concept.
These architectures are expected to drastically increase the
achievable data-rate and significantly reduce the in-body de-
vices energy-consumption. For each architecture, we report
the main operation principles, technology enablers and critical
technology gaps. We emphasize the need of developing novel
channel models that take into account the transdermal and
in-body optical-medium particularities as well as deriving
a corresponding theoretical framework for the performance
assessment and design of the PHY and MAC schemes.
II. BEYOND CLASSICAL BIOMEDICAL-APPLICATIONS
Optical-wireless MIs (OWMIs) and NWNs are envisioned
to enable a vast variety of novel biomedical-applications, such
as smart drug delivery and tissue recovery, as well as to
improve the QoE and safety of the conventional ones, not
only by targeting 10−100 times higher-data-rates, but also by
combining them with reliability and compact designs. Vital
signals, pathogens and allergens real-time continuous moni-
toring, detection of tissues and molecules abnormalities, as
well as smart drug delivery are only some examples of several
highly challenging applications. To support such scenarios, the
research world turns to the adaptation of THz technologies,

2
Fig. 1: RF vs OWCs.
which, by leveraging graphene designs, are expected to utilize
nano-scale transceivers. However, the safety of THz links is
still questionable and these technologies are yet immature. A
comparison between RF and OWCs is provided in Fig. 1. To
break these barrier, we need to directly research towards the
mature OWC concepts, identify and categorize the possible
applications and design suitable architectures and systems [3].
In this direction, we present the main application categories
based on the link nature and scenarios as well as their critical
design parameters. These applications are depicted in Fig. 2.
Biomedical-applications requiring the establishment of
transdemal links: The main representatives of this category
are cochlear, retinal, cortical and foot drop implants, gastric
stimulators, wireless capsule endoscope, insulin pumps, and
implantable orthopedic devices. Maybe the most successful
application is cochlear implants (CIs), which have restored
partial hearing to more than 350,000 people, half of whom
are pediatric users, who ultimately develop nearly normal
language capability. Conventional CIs exploit near-field mag-
netic communication technologies and typically operate in
low-RF frequencies, from 5to 49 MHz, while their transmit
power is in the order of tens of mW. As a result, they
cannot support high-data-rates (in the order of Mb/s), under
reasonable transmit power constraint, which are required to
achieve similar performance to the cochlea. Transdermal-
optical-link (TOL) is an important building-block to guaran-
tee high-speed-connectivity between the external and internal
units with low-energy-consumption [4]. Although, the need
for higher data-rates is detrimental for the performance of
retinal implants, it plays an important role in the quality of
the cochlear implants as well. One reason is the higher sound
quality that can be achieved with higher data-rates. A high
definition recording, for example 24 bits/192 kHz, requires
a minimum data rate of 4.6 Mbps for single channel audio
and 9.2 Mbps for dual channel audio, which is impossible
with current state-of-the-art cochlear implants. Furthermore,
another reason that justifies the requirement of higher data-
rates is the fact that higher transmission frequency that is
entangled with the communication data-rate plays an important
role in the precision of the stimulation of the cochlear neurons,
which results in the information to be conveyed faster to the
implanted part of the system and the stimulation of the targeted
neurons to be performed in a more timely manner. In this
scenario, apart from the high targeted data-rates (10 Mb/s),
the critical parameters are the latency, which should be in the
order of 0.1 ms, uninterrupted connectivity (i.e., outage prob-
ability lower than 10−5), and low error rate (BER ≤10−4).
Additionally, these devices should support SLIPT in order to
guarantee the internal device energy-autonomy.
After the successful clinical validation of CIs, neural-
prostheses were exploited in the treatment of visual impair-
ments, by developing retinal-implants (RIs). In this appli-
cation, the implanted device is an epiretinal prosthesis that
includes a receiver antenna, a DSP-unit, and an electrode array
attached in to the retina surface. The external unit consists of
a miniature-video-camera attached in glasses and a transmitter
coil, connected via a small video-processing-unit, which con-
verts the captured-video to electrical-pulse. These pulses are
transmitted wirelessly to the implanted device. Unfortunately,
the visual acuity of this approach is quite low, due to the
limitations of the stimulation unit and the achievable data-rate
of the RF-communication-link. To break through this barrier,
two solution has been reported, namely, the use of optical-
stimulation approaches in order to reduce the pixel dimensions
and OWCs to increase the date-rate [1]. The critical parameters
for this application are the pixel population, which should be
higher than 60 ×60 pixels, the data-rate (∼100 Mb/s), and
the latency (less than some decades of ms).
Gastric stimulators are used in the treatment of gastric
dysmotility disorders and obesity. They consists of two units,
namely out-of-body and in-body. The former’s main respon-
sibility is to wirelessly transfer power to the latter, which
consists of a received RF-antenna, a charge pump and a
peripheral interface controller that generates the electrical-
pulse and is connected to the electrodes. The transmission
medium is body tissue of 1to 5 cm thickness. This is
translated to an about 4 dB penetration-loss, when RF-signals
are employed. However, if optical-signals are employed, this
loss is further constrained. Except from the transmission range,
another critical parameter for this application is the in-body
unit compactness.
Biomedical-implants are also used for bio-signals and bio-
processes monitoring, such as neural implants, insulin pumps
and wireless capsule endoscopes. Neural implants, pacemak-
ers, and insulin pumps, except of the high-data-rates, demand
high-energy-autonomy and compactness of the in-body unit.
In other words, efficient-power-transfer that take into account
the space constraints needs to be utilized. On the other hand,
wireless capsule endoscope is an ad-hoc setup. Thus, its bat-
tery covers the energy-demands, while its main challenge is to
sustain continuous-connectivity between the in-body and out-
of-body units independently of the in-body unit orientation.
In-body biomedical-applications: Nanotechnology provides
a new set of tools to control matter at the atomic and
molecular scales, thus, enabling the development of nano-
scale biomedical-applications for nano-sensing and actuation.
Nano-devices utilized in such systems require the use of
bio-compatible materials and communication techniques with

3
Fig. 2: Biomedical-applications that can be supported
by OWCs.
limited electromagnetic radiation on the biological-tissues [5].
Moreover, several breakthroughs in the field of nano-photonic
devices enabled nano-scale PWCs. These applications in-
clude diagnostic and therapeutic techniques utilizing nano-
devices, healthcare monitoring solutions that combine in-body
nano-photonic bio-sensors and non-intrusive brain-machine
interfaces. Among others, nano-LEDs can be leveraged as
energy-efficient compact signal sources for optical-wireless-
links, which meet both size and power requirements of nano-
devices [6]. Similarly, sensitive nano-PDs were developed and
can be combined with nano-LEDs to form nano-sensors, which
can be utilized in compact nano-transceivers to forge fast,
short-distance wireless-communication-links. The aforemen-
tioned nano-transceivers when paired with nano-antennas, can
serve as nano-gateways to bridge the communication with out-
of-body devices.
Despite the latest advances, numerous challenges still exist
that need to be addressed. For example, in the case of wireless-
network on a chip, the main concern is to increase the capacity
and data-rate of the network, while dealing with a more
predictable and controllable system model. Therefore, new
challenges in terms of nano-photonic devices and communi-
cation protocols, such as energy-efficiency, link-capacity and
reliability, have to be addressed before NAO networks can be
established as a realistic solution.
III. CANDIDATE ARCHITECTURES
For transdermal-optical-links: As illustrated in Fig. 3, three
candidate architectures for transdermal OWCs have been iden-
tified, namely OE, EO, and AO. OE and AO are suitable
for translating external stimuli, such as audio and video, into
nerve stimulation signals. OE was introduced in [1] for next-
generation CI designs. The main difference between EO/OE
and AO is that the former’s in-body units performs DSP. Thus,
a energy-autonomy demand for the internal-device arises. To
satisfy this, a SLIPT policy is utilized. In particular, the DC
Fig. 3: Candidate architectures for transdermal-applications.
part of the received signal is used for energy-harvesting, while
the AC part conveys the neural stimulation message. A simple
AC and DC separator (ADS) is employed at the internal device
consisting of a capacitor that blocks the DC component of
the received signal and forwards it to the energy-harvesting
branch. At the same time, the AC component of the signal
at the output of the ADS is inserted to the DSP-unit, which
together with the STM translates the received message into
the appropriate neural-stimulation-signal.
On the other hand, in AO, no-SLIPT is required, since
the stimulation signal is directly emitted by the out-of-body-
device. The in-body-device consists of a fiber-coupling-unit,
which is usually a MEM and is responsible for capturing the
transmitted-signal and forwarding it to the optical-stimulator.
This concept may be utilized by employing commercially
available optics and optogenetics modules. In this scenario,
DSP for mitigating the impact of both the TOL impairments
and stimulation signal is required. The main challenge of AO
is modeling the link from external-unit to the corresponding
nerve. Another challenge is mapping the transmission signal
into the stimulation one.
The EO-architecture can be used for monitoring and sens-
ing bio-signals; hence, it can be employed in neural im-
plants and insulin pumps. The in-body-device consists of a
sensing/monitoring-unit, a DSP-unit and a light-source that
emits the modulated sensed message. To cover the energy-
requirements of the in-body-device an energy-harvesting-
branch is utilized. The out-of-body-device consists of a pho-
todetector that captures the light-signal and forwards it in a
DSP-unit, responsible for message detection. The detected-
signal is forwarded to the imaging-device. Moreover, the out-
of-body-unit has a DC signal-generator with a dimming con-
troller that feeds a light-source, responsible for energy-transfer.
The major challenge of EO is to present suitable dimming
control strategies that guarantee the energy-autonomy of the
in-body-device. Moreover, this concept allows the integration
of the internal device front-end into a single-module with
superior compactness and energy-efficiency.

4
Fig. 4: Candidate architectures for in-body optical-
applications.
For nano-scale systems: For the utilization of NAO net-
works, the two architectures depicted in Fig. 4 have been
identified. A common nano-node structure is considered in
both cases. Each nano-node is equipped with a nano-LED and
a nano-PD that enable communication with others, as well
as an energy-harvester that replenishes the energy stored in a
nano-capacitor. This enables circulating nano-nodes to over-
come the energy-limitations and even have practically infinite
lifetime. Depending on the application, nano-nodes can be
equipped with nano-sensors and nano-actuators for monitoring
and manipulation of biological processes, respectively.
The first architecture incorporates an intermediate nano-
device, called nano-gateway. This device is responsible for
forwarding the data collected from the nano-nodes to the exter-
nal interface. The link between them can be established on the
optical or the RF spectrum, depending on the application char-
acteristics. For example, in close-range communications the
optical-link has been proven to outperform the corresponding
RF, while for longer range applications the required optical-
beam intensity for achieving adequate signal quality can be
destructive to the human tissue due to increased temperature.
The second architecture consists of the nano-nodes and the
nano-router. The former move throughout the body with a
pattern that varies with the application, while the latter collects
the gathered information when the nano-nodes are in range [7].
For example, if we consider a scenario where the nano-nodes
are injected into the bloodstream and the nano-router is placed
on the outer surface of the skin, as the nano-node travel
through the circulatory system, they communicate with the
nano-router when they are positioned close enough.
The main difference between the two architectures is scal-
ability. The first architecture can be multiplied throughout
the human-body, in both deep-tissue and near the surface
alike. By placing the nano-gateways with appropriate distance
between them, the first architecture shown in Fig. 4 can be
repeated multiple times. However, both architectures share the
same concerns such as limited energy-consumption, limited
computational capabilities and stochastic network topology.
IV. DESIGN PRINCIPLES AND TECHNOLOGY ENABLERS
In order for the candidate architectures to provide solu-
tions to the OWC-based biomedical applications challenges,
rethinking of the fundamental design principles and adoption
of signal-transmission and neural-stimulation jointly design
approaches is necessary. Towards this direction, a generalized
optical-wireless-channel model is needed that accommodates
the particularities of such systems and can support the ex-
traction of their fundamental limits. Moreover, new neural-
stimulation approaches need to be examined in order to
countermeasure the pixel population requirement. To cover the
energy autonomy demand, SLIPT approaches are discussed.
Finally, the development of PHY and MAC schemes capable
of supporting the data-rate, range and latency limitations as
well as neural-stimulation functionalities becomes a necessity.
A. Optical-wireless-channel
In-body and transdermal optical-wireless-channel models
are considered to be the main building-blocks for developing
the aforementioned architectures. In comparison with the out-
of-body models, these ones are required to accommodate a
number of different characteristics and biological particulari-
ties. For instance, in TOLs, the dermis content in hemoglobin
can significantly influence its light absorption in the blue
and green-yellow regions. On the other hand, the dominant
absorber in the ultraviolet and infrared regions is the epider-
mis melanin. Finally, chromophores, like bilirubin, carotene,
lipids, cell nuclei, filamentous proteins, etc. can cause further
absorption in different wavelengths. Of note, in spite of the
abundance in all tissues, due to the short communication
distance, water is not a significant light absorber in these
systems. This is one of the differences between transdermal/in-
body and out-of-body optical-channels.
The impact of the optical-wireless-channel particularities
with regard to path-loss is presented in Fig. 5. This figure
summarizes the main contributing elements of any tissue
which is applicable to the communication paradigms under
investigation. Moreover, the path-loss that occurs when light
of a specific wavelength travels through a tissue is a function of
the absorption coefficient of the tissue, which can be calculated
based on the tissues concentrations and absorption coefficients
of water, blood, fat and melanin [9]. In order to calculate
the impact on the link’s performance only due to the path-
loss, we assume perfect alignment, negligible noise, perfectly
aligned receivers with divergence angle and field of view such
that they completely receive the transmitted beam, and the
received signal is normalized to unity. From Fig. 5, we observe
that the attenuation due to the existence of blood plays a
detrimental role in the performance of the communication link.
If we consider only the blood, it is evident that a transmission

5
Fig. 5: Optical-wireless-channel path-loss. Of note the pathloss
was evaluated based on the model presented in [8].
window exists after 600 nm. On the contrary, water attenuates
more intensively optical-signals with higher wavelength, while
the rest of the elements have a more consistent impact. As
a result, the appropriate transmission wavelength for each
use case has to be selected based on the composition of the
intermediate tissue.
Apart from path-loss, transdermal and in-body optical-
links experience wavelength-dependent particulate scattering.
In particular, within the skin, the main source of scatter is
filamentous proteins (like keratin in epidermis and collagen
in dermis). Note than since these particles are comparable
or larger than the transmission wavelength, scattering can
be approximated as a Mie solution to Maxwell’s equations.
On the other hand, in in-body applications, tissues, such
as membranes, striations in collagen fibrils, macromolecules,
lysosomes, vesicles, mytochondria, and nuclei, are the main
scatters. Notice that membranes are usually lower than 1/10 of
the wavelength, while all the other structures are comparable
to the wavelength. Thus, scattering in tissues can be modeled
as a mixture of Rayleigh and Mie processes.
Inhomogeneities in the body content in light absorbing and
scattering structures leads to a variation of the reflective index
along the transmission path, which cause random fluctuations
in both the amplitude and phase of the received signal, i.e.
channel fading. To provide the theoretical framework for the
performance assessment and design of optical-biomedical-
implants, it is of high importance to deliver a stochastic
model for the accommodation of this type of fading. In this
direction, stochastic geometry approaches have been employed
(see [10] and references therein). Finally, another source of
received power uncertainty, which arises from the directional
nature of the optical-links, is transceiver misalignment. As
reported in [4], this type of uncertainty can be modeled
through a stochastic process and can significantly affect the
system performance.
Another differentiation of in-body/transdermal-optical-
communications in comparison to other out-of-body appli-
cations is the existence of neural noise. Specifically, optical
approaches for tracking neural dynamics have modeled the
physical bounds on the detection of neural spikes as photon
counting statistics (shot noise). Besides the neural noise, other
noise sources come from the receiver, like thermal, background
and dark current noises, and can be modeled as zero-mean
Gaussian processes. Their power depends on the communi-
cation bandwidth, the detector’s responsivity, the background
optical-power and the intensity of the dark current, which is
generated by the photodetector in the absence of background
light. Depending on the architecture, different types of noise
determines the fundamental bounds. For example, in AO,
neural noise is expected to play the most detrimental role,
while, in OE and EO, its intensity tends to zero.
The above observations motivates the development of
generalized deterministic path-loss and statistical in-body-
particulate models for TOLs and in-body optical-links that
will aid to the appropriate transmission waveform design, the
development of the PHY and MAC, as well as the utilization
of the candidate architectures. These models should take into
account the patients particularities, such as the skin color, the
composition in different structures (collagen, melanin, etc.),
the transmission wavelength, and the misalignment intensity.
B. Electrical and optical neural stimulation
Based on the stimulation type, implants can be classified
into electrical and optical. Electrical-stimulation methods use
an external electrical current to manipulate the membrane
potential directly and have been extensively used for igniting
action potentials in various types of human cells. Numerous
medical implants, such as cochlear implants, pacemakers,
and more, take advantage of these techniques. However, the
performance of electrical-stimulation is hindered by poor
spectral coding and bandwidth scarcity of the RF spectrum.
These limitations motivated the research community to turn
its attention to the optical spectrum, which is capable of
providing large amounts of unexploited bandwidth and higher
safety for the human body [11]. As a result, optical-stimulation
has been greatly investigated over the past decade, due to
the various advantages it offers over the established electrical
one. Optogenetic techniques are based on the development
of several genetically modified proteins, called opsins, which
when illuminated generate a flow of ions through the cellular
membrane that alters the membrane potential [8]. This imple-
mentation has been experimentally proven to control the cell
dynamics with higher accuracy, frequency and greater spatial
resolution.
Regarding their performance comparison, as
indicated in Fig. 6, under the same power consumption
and stimulation frequency, optical-stimulation is proven to
outperform the equivalent electrical one in terms of accuracy
and power consumption, respectively. In this figure, the
stimulation success rate is calculated as the percentage of the
transmitted pulses that successfully generated action potential
at the targeted neurons. It should be noted that the vertical
doted line, which corresponds to the 15 Hz, marks the upper

6
Fig. 6: Optical versus electrical stimulation performance.
frequency limit for the electrical stimulation. It is evident
that the optical-stimulation is capable of generating accurate
action potentials with higher frequency while maintaining the
similar power consumption. This observation is in agreement
with several publications, which state that optogenetics offer
more precise control over the excited neurons due to both
the higher achievable frequency and the temporal accuracy
provided by the variety of existing photosensitive opsins.
In addition, the optical-cell-stimulation outperforms the
corresponding electrical in terms of the stimulation success
rate. This is expected due to the fact that the optical-spectrum
offers more bandwidth. However, if we consider the fact
that optical-signals attenuate faster as they travel through
the human body thus offering lower geometric spread, or in
other words, higher spatial resolution, we can deduct that
optical-stimulation is an overall more robust solution for cell
stimulation.
The safety and stability of radiating the human body with
light pulses with repetition rates of few hundreds of Hz need
to be evaluated. Such stimulations at high intensities have been
proven to cause phototoxicity or heating [12]. Pulses with
optical irradiance up to 75 mW/mm2are considered safe for
in vivo application, while tissue heating has been reported for
amplitude of approximately 200 mW/mm2and phototoxic
effect are absent up to 600 mW/mm2. Furthermore, these
limits were acquired for pulse width of 5 ms, while reliable
excitation can be achieved at about 1 ms and possibly even
shorter with future advances.
C. Energy-transfer and SLIPT
A critical system design parameters in several in-body
biomedical-devices is their lifetime, which depends on the
amount of available power. Replacing or removing and
recharging the battery of such device is impractical or even
impossible. To counterbalance this, energy-harvesting from the
body using thermoelectric, piezoelectric, and electromagnetic
generators attracted the eye of biomedical-engineers [13].
However, these approaches have two inherent disadvantages:
(i) low-harvesting-efficiency (∼20%); and (ii) require the
installation of extra modules that increase the implementation
cost and size. As a consequence, SLIPT policies seem to be
more appealing for biomedical-applications.
In biomedical-devices with SLIPT capabilities, the PD,
which is a common-choice, due to the high-data-rate that
supports, needs to be replaced with solar cells, which provide
apparent advantages in terms of harvesting-efficiency [14],
[15]. Solar cells dimensions can be different, depending on the
application for which they are used, from some nano-meters
to decades-of-centimeters. They are basically photoelectric
converters with zero bias and can be used to support both
energy-harvesting and information-decoding. From the SLIPT
utilization point of view, two fundamental policies can be
used, namely time and signal splitting. In the former, the
transmission period is divided into an energy-harvesting and
information-transfer one, while, in the latter, the DC and AC
part of the received photo-current are respectively used for
energy-harvesting and information-decoding.
Although SLIPT approaches and policies were extensively
investigated in different types of OWC systems, they cannot
be straightforwardly applied in biomedical-systems, because of
the receiver particularities. Receivers designed for out-of-body
OWC-applications with SLIPT capabilities are equipped with
solar panels. On the other hand, in biomedical-applications,
they are equipped with solar cells. As a consequence, SLIPT
is identified as a promising open research direction and further
investigation on the suitability of each policy in each one of
the aforementioned applications needs to be conducted.
D. PHY & MAC layer
By exploiting the capabilities of the optical-transceivers,
we report the new challenges and opportunities at the PHY
and MAC, including energy-efficient modulation-and-coding
(MC), energy-transfer and SLIPT policies. Consistently with
the demands of ultra-low power consumption, compact-size,
reliability and ultra-low complexity, new MC-schemes capable
of accommodating the different type of noise and channel
particularities need to be developed. Intensity-modulation and
direct-detection with power adaptation is a possible approach
to deal with channel’s stochastic behavior, whereas transmis-
sion and/or reception diversity through the exploitation of
multiple light-sources and photo-detectors can be employed
to deal with pointing errors, while offering diversity through
repetition coding schemes, which consequently can contribute
to a significant error-rate reduction. Moreover, channel particu-
larities and the transceiver limitations require the development
of novel channel-codes for in-body/transdermal OWCs. The
conventional capacity-approaching channel-codes, which are
designed to maximize the data-rate for a given transmit-power,
demand additional transmission-power and time for decoding
that may violate the application requirements.Thus, we need
to characterize the error-sources, i.e. both the noise and chan-
nel natures, and examine the trade-off between transmission-
power and decoding-time in order to design channel codes.
The use of low-complexity error-prevention coding schemes
might be a solution.However, the identification and design of
the optimal coding scheme is still an open issue.

7
From the MAC perspective, the directional nature of
OWCs combined with the high-data-rates increase the spatial-
synchronization requirements in nano-scale optical-networks
and call for novel MAC protocols that guarantee transceivers
alignment and enables mobility management. In this direc-
tion, schemes that allow devices to periodically operate in
sleeping, discovering, receiving, and transmitting modes may
be the solution. During discovering phase, each device sends
synchronization signals to all directions in order to announce
its availability as a RX or its intention to transmit a message
to a specific node. Meanwhile, its neighbors are capable of
discovering the aforementioned device. A distributed routing
protocol then needs to be designed that allows the TX to com-
municate with the intended-RX through intermediate-nodes.
The design of the optimal root should take into account not
only the minimization of the network’s energy consumption,
but also the power availability to each one of the intermediate
nodes. Another functionality that needs to be utilized is fast
optical-tracking and steering. Further research is needed to
address the development of appropriate field-of-view restric-
tions, effective steering and adaptive control systems for nano-
scale biomedical-applications. Finally, notice that due to the
directional nature of both THz and OWC networks, several
of the aforementioned challenges are the same. The main
difference between OWC and THz networks resides in the
different channels, each with distinct particularities, which
generate new PHY and MAC layer challenges for OWC nano-
scale networks. Therefore, solutions that have been applied in
THz nano-scale communications needs to be re-examined for
possible adoption in OWC nano-scale networks.
V. CONCLUSIONS
In this article, we presented the concept of optical-in-
body and transdermal-optical-communications. In more de-
tail, after presenting the possible biomedical-applications and
identifying their requirements, we described candidate system
architectures and features. Due to the fact that their designs
are in early stage, it is difficult to determine their final
form. However, their enablers and technology modules were
reported, namely channel modeling, the development of novel
PHY, MAC and SLIPT schemes, and the design of appropriate
stimulation units. Finally, future research directions were high-
lighted. Note that this article is not just a review or tutorial,
since it further aims at identifying the technology enables that
open the road to another promising application of OWCs.
ACKNOWLEDGMENT
This research has been supported by the EU and Greek
national funds through the Operational-Program Human-
Resource-Development, Education and Lifelong Learning.
REFERENCES
[1] S. E. Trevlakis, A.-A. A. Boulogeorgos, P. C. Sofotasios, S. Muhaidat,
and G. K. Karagiannidis, “Optical wireless cochlear implants,” Biomed.
Opt. Express, vol. 10, no. 2, p. 707, Jan. 2019.
[2] K. Agarwal, R. Jegadeesan, Y. X. Guo, and N. V. Thakor, “Wireless
power transfer strategies for implantable bioelectronics,” IEEE Rev.
Biomed. Eng., vol. 10, pp. 136–161, Mar. 2017.
[3] M. Uysal, C. Capsoni, Z. Ghassemlooy, A. Boucouvalas, and E. Udvary,
Optical Wireless Communications: An Emerging Technology, ser. Signals
and Communication Technology. Springer International Publishing,
2016.
[4] S. Trevlakis, A.-A. Boulogeorgos, and G. Karagiannidis, “Signal qual-
ity assessment for transdermal optical wireless communications under
pointing errors,” Technologies, vol. 6, no. 4, p. 109, Nov. 2018.
[5] F. Pisanello, L. Sileo, and M. D. Vittorio, “Micro- and nanotechnologies
for optical neural interfaces,” Frontiers in Neuroscience, vol. 10, Mar.
2016.
[6] S. A. Wirdatmadja, M. T. Barros, Y. Koucheryavy, J. M. Jornet,
and S. Balasubramaniam, “Wireless optogenetic nanonetworks for
brain stimulation: Device model and charging protocols,” IEEE Trans.
Nanobiosci., vol. 16, no. 8, pp. 859–872, Dec. 2017.
[7] S. Canovas-Carrasco, A.-J. Garcia-Sanchez, and J. Garcia-Haro, “A
nanoscale communication network scheme and energy model for a
human hand scenario,” Nano Communication Networks, vol. 15, pp.
17–27, mar 2018.
[8] E. A. Ferenczi, X. Tan, and C. L.-H. Huang, “Principles of optoge-
netic methods and their application to cardiac experimental systems,”
Frontiers in Physiology, vol. 10, Sep. 2019.
[9] M. Faria, L. N. Alves, and P. S. de Brito Andr´
e, Transdermal Optical
Communications. CRC Press, Jun. 2017, vol. 1, ch. 10, pp. 309–336.
[10] H. Guo, P. Johari, J. M. Jornet, and Z. Sun, “Intra-body optical channel
modeling for in vivo wireless nanosensor networks,” IEEE Transactions
on NanoBioscience, vol. 15, no. 1, pp. 41–52, jan 2016.
[11] J. M. Jornet, Y. Bae, C. R. Handelmann, B. Decker, A. Balcerak,
A. Sangwan, P. Miao, A. Desai, L. Feng, E. K. Stachowiak, and M. K.
Stachowiak, “Optogenomic interfaces: Bridging biological networks
with the electronic digital world,” Proceedings of the IEEE, vol. 107,
no. 7, pp. 1387–1401, Jul. 2019.
[12] A. Dieter, D. Keppeler, and T. Moser, “Towards the optical cochlear im-
plant: optogenetic approaches for hearing restoration,” EMBO Molecular
Medicine, vol. 12, no. 4, Mar. 2020.
[13] J. Selvarathinam and A. Anpalagan, “Energy harvesting from the human
body for biomedical applications,” IEEE Potentials, vol. 35, no. 6, pp.
6–12, Nov. 2016.
[14] Z. Wang, D. Tsonev, S. Videv, and H. Haas, “On the design of a solar-
panel receiver for optical wireless communications with simultaneous
energy harvesting,” IEEE J. Sel. Areas Commun., vol. 33, no. 8, pp.
1612–1623, Aug. 2015.
[15] L. M. Wangatia, S. Yang, F. Zabihi, M. Zhu, and S. Ramakrishna,
“Biomedical electronics powered by solar cells,” Current Opinion in
Biomedical Engineering, vol. 13, pp. 25–31, mar 2020.
Alexandros–Apostolos A. Boulogeorgos (S’11, M’16, SM’19) is a post-
doctoral researcher in the University of Piraeus and Aristotle University of
Thessaloniki (AUTh).
Stylianos E. Trevlakis (S’20) is a PhD candidate in AUTh.
Nestor D. Chatzidiamantis (S’08, M’14) is an Assistant Professor at AUTh.