Computer Networks 193 (2021) 108074

Contents lists available at ScienceDirect

Computer Networks
journal homepage: www.elsevier.com/locate/comnet

Survey paper

A Survey on Wearable Technology: History, State-of-the-Art and

Current Challenges
Aleksandr Ometov a ,∗, Viktoriia Shubina a,e , Lucie Klus a,d , Justyna Skibińska c,a , Salwa Saafi c,a ,
Pavel Pascacio d,a , Laura Flueratoru e,a , Darwin Quezada Gaibor d,a , Nadezhda Chukhno b,d ,
Olga Chukhno b,a , Asad Ali a,c , Asma Channa e,b , Ekaterina Svertoka e,c , Waleed Bin Qaim a,b ,
Raúl Casanova-Marqués c,d , Sylvia Holcer d,c , Joaquín Torres-Sospedra d , Sven Casteleyn d ,
Giuseppe Ruggeri b , Giuseppe Araniti b , Radim Burget c , Jiri Hosek c , Elena Simona Lohan a
a
Tampere University, Finland
b
University Mediterranea of Reggio Calabria, Italy
c
Brno University of Technology, Czech Republic
d Universitat Jaume I, Castellon, Spain
e University Politehnica of Bucharest, Bucharest, Romania

ARTICLE INFO ABSTRACT

Keywords: Technology is continually undergoing a constituent development caused by the appearance of billions new
Wearables interconnected ‘‘things’’ and their entrenchment in our daily lives. One of the underlying versatile technologies,
Communications namely wearables, is able to capture rich contextual information produced by such devices and use it to
Standardization
deliver a legitimately personalized experience. The main aim of this paper is to shed light on the history
Privacy
of wearable devices and provide a state-of-the-art review on the wearable market. Moreover, the paper
Security
Data processing
provides an extensive and diverse classification of wearables, based on various factors, a discussion on wireless
Interoperability communication technologies, architectures, data processing aspects, and market status, as well as a variety of
User adoption other actual information on wearable technology. Finally, the survey highlights the critical challenges and
Localization existing/future solutions.
Classification
Future perspective

1. Introduction new market segment – a separate niche requiring standalone efforts

has emerged as devices carried by humans but no necessarily gener-
Today, the rapid proliferation of the Information and Communica- ating the human-type data. Internet of Wearable Things (IoWT) has
tions Technology (ICT) niche is being pushed by an increasing number
emerged as part of a broader IoT, bringing new challenges from various
of new services and growing user demands. Generally, the number of
interconnected handheld devices has been growing tremendously from technological perspectives to the research community [2].
year to year, empowered by both consumers and broad penetration of The terms wearables, wearable devices, or also wearable technology
the Internet of Things (IoT) [1]. Small, affordable, and very different refer to small electronic and mobile devices, or computers with wireless
in shape, purpose, and application, the IoT devices had a tremendous communications capability that are incorporated into gadgets, acces-
impact on the development of the telecommunications field, not only
sories, or clothes, which can be worn on the human body, or even
bringing new long-range wireless technologies to the market, defining
invasive versions such as micro-chips or smart tattoos [3]. Compared to
new requirements in terms of reliability and availability but also push-
ing network operators and vendors to redesign the entire ecosystem, today’s smartphones and tablets, the primary added value is that wear-
switching from conventional human-generated traffics to more diverse ables can provide various monitoring and scanning features, including
IoT one. biofeedback or other sensory physiological functions such as biometry-
This tremendous impact on the ICT domain and proliferation of related ones [4]. Wearables can continuously measure such values —
the IoT allowed the developers to bring their attention to an entirely

∗ Correspondence to: Tampere University, P.O. Box 1001, FI-33014, Finland.

E-mail address: aleksandr.ometov@tuni.fi (A. Ometov).

https://doi.org/10.1016/j.comnet.2021.108074
Received 1 December 2020; Received in revised form 3 March 2021; Accepted 27 March 2021
Available online 8 April 2021
1389-1286/© 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
A. Ometov et al. Computer Networks 193 (2021) 108074

Fig. 2. Wearable market growth forecast [9].

Fig. 1. Perspective on the development of personal wearable ecosystems.

steadily increase over the following decades, with a predominant transi-

restricted by their battery constraints; they are convenient, seamless, tion from bracelets and sport trackers to smarter and more feature-rich
portable, and can offer hands-free access to electronics. wearables. Wearable technology has a tremendous impact upon ICT in-
The consumer-wearable devices of today are beyond their inception dustry, and smart wearables are expected to disrupt most personal and
but still very much in their infancy. Most people still use a combination business sectors, such as the industrial, healthcare, and sports domains.
of an activity tracker and a smartphone, see Fig. 1, limiting their Moreover, the global enterprise wearable market alone accounted
experience to the number of steps and heart rate. Indeed, the variety for over EUR 18 billion in 2017, and it is predicted to grow at a
of data collected and processed in a wearable ecosystem context can Compound Annual Growth Rate (CAGR) of 11.8% during the forecast
hand over an unprecedented user experience for humanity [5]. In ad- period, 2019 − 2026 [17]. The details on market changes as well as
dition to conventional sports trackers, smartwatches, on-body cameras, future perspective [9] are given in Fig. 2. Here, wearable devices could
heart rate meters, and eye-wear, the upcoming generation of wearables be used to improve employees’ physical activity, enable relaxation
will also involve augmented-, virtual-, mixed-, and enhanced-reality more effectively, and increase workers’ safety and work efficiency [18].
devices, various smart clothes, and industrial wearable equipment. Added value is also achieved in other sectors, such as smart city,
As foreseen by the ICT industry, almost 70% of early adopters have transportation, gamification, and infotainment. Wearable technology
shown their interest in correlating their lives with the next-generation relieves people from continuously having to hold their smartphones
wearables [6]. Predominately, a significant portion of IoWT available in their hands from the usability perspective. This convenience feature
on the market today already provides a smartphone-like experience by allows handling calls, emails, texts, and many other alerts without even
employing voice and gesture control together with well-designed input getting the bulky devices out [19,20]. Following Llamas’ work, ‘‘within
and output interfaces. the enterprise, wearables can help accelerate companies’ digital trans-
Nonetheless, the miniaturization trend, portability, wireless commu- formation by transmitting information back and forth while allowing
nication, energy-efficient computing, and advanced display technolo- workers to complete their tasks faster’’ [21].
gies have been combined to create state-of-the-art smart devices. The The mobile devices’ market growth brings new and usable devices,
patriarch of these devices, a smartphone, was released back in 1992 [7], numerous benefits, and new applications from the users’ perspective.
and intelligent media are now ready to lead the next great wave of One of the primary stimuli brought by wearable technology is the
innovation. Comparing traditional smartphones with wearables, both encouragement of proactive solutions to deal with healthcare, fitness,
have their pros and cons. A traditional smartphone’s main advantage aging, disabilities, education, transportation, enterprise, finance, en-
is its higher accuracy in various performance metrics due to less power trance systems, gaming, music, and many others. Since wearables, as
consumption limitations as a general trade-off to size. On the other known today, were historically planned as purely medical devices, let
hand, wearables are highly battery-constrained devices, yet have the us first consider an example from the healthcare field. Unfortunately,
potential to change the world as we know it — just as mobile devices people tend to deal with potential health issues reactively, e.g., when
did over the past 20 years. It is expected that they will improve the they feel sick or in pain, they tend to appoint a visit to their doctor.
technological and socio-cultural parts of our lives. Moreover, wearables Carrying a wearable device may potentially forecast the disease by
also have the strength to improve well-established sectors, such as continuous health tracking and even inform the doctor automatically
the smartphone industry and other hand-held devices. This trend is in order to take measures to prevent the incipient threat actively [22].
confirmed by many recent studies and will be discussed in this survey. Even the simple activity trackers are already capable of monitoring
From the monetary perspective, the wearable market is anticipated sleep patterns, heart rate, level of stress, or body temperature that could
to keep on growing exponentially in the coming years. The forecast is be utilized for improving the health habits of any individual [23].
at more than 20% growth rate annually, and the market is expected Significantly, the shift toward advanced devices, such as Augmented
to reach over 40 billion EUR per year in the next 5 years with more Reality (AR)/Virtual reality (VR)/Mixed Reality (MR)/eXtended reality
than 150 billion EUR by 2028 [8–10]. A recent number of wearable (XR) devices, low-end wearables, and other monitoring devices, to-
shipments, estimated at 113.2 million in 2017 with total market size gether with a transition to Beyond Fifth Generation (beyond-5G) mobile
of $70 billion in 2019, is forecast to reach 222.3 million deliveries networks would also bring several challenges for the device vendors,
yearly by 2021 [11,12]. Moreover, the outbreak of COVID-19 also network operators, and end-users. In particular, those challenges are
made a tremendous impact on the wearable devices evolution driven related to the paradigm shift from conventional Human-to-Human
by the implementation of various crowd-sensing and contact-tracing interaction (H2H) to more Machine Type Communications (MTC) in-
platforms [13–16]. The proliferation of wearables is thus expected to teractions [24]. This segment brings completely different requirements,

2
A. Ometov et al. Computer Networks 193 (2021) 108074

e.g., utterly different traffic patterns, higher reliability, lower latency, In line with the discussions so far, the main goals of this paper are:
highly mobile scenarios, stringent security and privacy needs, and
higher energy-efficiency demands than the H2H ones [25]. G1. To provide a detailed historical overview of wearable technol-
All of the conditions mentioned above lead to an indigenous in- ogy evolution;
crease in power consumption and the need to recharge the wear-
G2. To highlight the state-of-the-art in the field of wearable technol-
able/handheld devices daily, reducing their attractiveness and limiting
ogy, including mass-consumer perspective, main architectures,
the wearable applicability. Modern technologies, such as energy har-
market-available devices, and related technical classifications;
vesting [26] and wireless charging [27], may assist in solving the
energy bottleneck, but the corresponding impact on the user’s health G3. To identify open challenges and provide a vision of future devel-
has not yet been studied in detail [28]. Besides, energy harvesting opments related to wearable markets and wearable computing.
technologies are still far from mass adoption [29].
Overall, research in wearable technology is growing, which is The rest of the paper is structured as follows. Section 2 provides
demonstrated, among others, by the number of publications in the a detailed insight into the history and development of wearables from
IoWT domain. Nevertheless, only one unified review is available to the early 13th century till nowadays. Section 3 addresses the state-of-
the research community. In contrast, others have a particular focus on the-art of wearable-technology evolution and outlines the main classifi-
some specific area (eHealth, sensors, adoption, etc.), while this paper cation, architectures, communication possibilities, and data processing
attempts to synthesize a standalone executive summary of various aspects of modern wearable devices followed by the main trends in
technological aspects as well as related challenges. As of February active wearable-technology development from both industrial and aca-
2021, the authors identified the following papers related to our survey, demic perspectives. Section 4 contributes by providing the list of main
i.e., at a comparable level of abstraction and topics’ coverage. open challenges and future perspectives. Section 5 concludes the paper
Several works have already reviewed materials, technologies, and with a discussion summarizing the main findings of this work. The
applications involving Wearable devices. One of the broadest and de- appendix provides the results of the comprehensive market analysis
tailed surveys was shown to the IEEE community in 2017 [30]. Ac- reflecting various types and models of wearable devices being currently
cording to the authors, the paper has a product survey style and is available on the market and under research.
mainly focused on the market-available wearable accessories, eTex-
tile and ePatches classification, and computing-, energy specifics, and
2. Historical perspective on wearables
security aspects with a focus on present technologies. To note, close
to half of the paper deals only with the classification of wearable
devices available in 2017. A very recent review on wearable technology The wearables we know today are mostly treated as smart-by-
and consumer interaction [31] introduced five main themes, includ- definition devices. People tend to forget that ‘‘smartness’’ has not
ing decision-making, well-being; consumer behavior; utilities; and Big always been defined by processing the data on a chip, but rather by
Data analytics. Above all, they showed the lack of integration within delivering a better experience for actual users. The next subsections
wearable technologies, which is driving to fragmentation, disconnected give an overview of wearables’ evolution from the 13th century to 2015.
terminologies, and studies that are not based on appropriate results. The evolution is also graphically depicted in Fig. 3 for the ease of
The review performed by Xue et al. [32] introduced issues about perception.
technologies, users, and activities with intelligent wearables. Moreover,
it identified the main risks (privacy, safety, performance, social and 2.1. Before the 20th century
psychological) involving wearables, which were also mentioned as hot
topics/themes for wearables in other related reviews [33]. The analysis
The journey of wearables started with the invention of spectacles
introduced in [34] focused on the current research highlights based
around the 13th century by English friar Roger Bacon, who was based in
on nanomaterials and evaluated the electronics under the perspective
Paris and outlined the scientific principles behind the use of corrective
of actuators and sensors. The applications of Wearable Sensors were
lenses in his Opus Majus (c.1266) [40]. Before R. Bacon, there had
analyzed for healthcare and human movement monitoring. In contrast,
been mentions of presbyopic monks using segments of glass spheres
the review introduced in [35] targeted the evaluation/assessment of
that could be laid against reading material to magnify the letters (i.e., a
physical activity apps and wearables. They identified that 75 out of 111
magnifying glass called ‘‘reading stone’’), but those are more question-
analyzed works used in-device sensors to measure physical activity.
able in terms of actual wearability. Glasses by Bacon were the first
Numerous works specifically target the eHealth segment, e.g., the
wearables designed to be seamlessly carried and improving the vision,
work [36] dated 2015 provides an extensive overview of wearable
thus, becoming the pioneering smart glasses [41].
health-monitoring systems and related communication aspects sensors
and related implementation challenges. Another relevant work [37] The first pocket mechanical watch, possible to be carried around,
also covers the communication and architectural aspect of the eHealth dates back to the beginning of the 16th century, and it is believed
domain, extending [36]. Both are, however, strictly limited to this to be the Pomander (Bisamapfeluhr in German) watch [42]. Peter
particular segment. A cross-section of technologies for Sports and e- Henlein made it in 1505 as a portable but a not-very-precise clock.
Health was also assessed in [38]. The primary purpose was to evaluate This design started a hype in wearable watch development, followed
their reliability and suitability. They detected that most of the analyzed by more than ten different models in the next one hundred years [43].
technologies were not formally validated by an independent actor, and Later on, pocket watches have also been developed significantly with
only 5% provided a formal validation. The highlights provided in [39] the evolution of miniaturization [44], which led to the idea of strapping
included the rising of wearables and the future for monitoring human the device to the wrist in the 19th century. Military needs primarily
activity as well as biological signals. drove the developments at that time [45].
Different from the other works, this paper provides a synergy of From the smart ring’s perspective, the first known one is Abacus
various aspects of wearable devices from perspectives of communi- Ring from the early 17th century during the Qing Dynasty era [46].
cations, computing, localization, modern and future communication Back then, a standard abacus was combined of 10 parallel wires located
capabilities and identifies challenges for each of the directions. As such, between two boards on a frame with nine beads on each of them. It was
it provides a standalone dataset of devices up until 2020, discussing the specially designed to be a compact smart accessory used to help traders.
aforementioned aspects without losing itself in (unnecessary) technical It led the way towards modern wearable computers and, at the same
details or device classification, as in [30]. time, towards modern smart rings [47].

3
A. Ometov et al. Computer Networks 193 (2021) 108074

Fig. 3. Milestones of the wearable devices evolution.

2.2. Early 20th century Heilig, who patented ‘‘Stereophonic Television Head-Mounted Display’’
in 1960 [54]. It was soon followed by another patent of the ‘‘Sensorama
The first step from the portable cameras’ perspective is the Pid- Simulator’’ being an upgraded version of the initial device [55]. The
geon camera, developed by the German inventor Julius Neubronner device was indeed the first VR simulator with a binocular display,
in 1907 [48]. Unlike many other technological breakthroughs of those vibrating seat, stereophonic speakers, cold air blower, and an odors
times, the camera was designed to record Neubronner’s pigeon flights, generator [56].
while this invention is sometimes mistakenly referred to as based
on military demands. It is probably because pigeon photography was 2.3. Mid 20th century
widely used during the first world war for aerial surveillance together
In 1961, MIT researchers Edward O. Thorpe and Claude Shannon
with airplanes [49].
concealed a timing device in a shoe that could accurately predict the
Indeed, wearable development driven by the military during the
ball’s landing place on a roulette table. That became the first wearable
World War I-II period was enormous. In the first place, the first carried-
computer hidden in the shoe [57]. The real story behind the invention
on wireless systems were redesigned for field communications [50].
was published later, in 1998 [58].
Those were very bulky and, at first, used to be carried by cavalry horses.
Just a few years later, Hugo Gernsback invented the TV glasses [59].
A breakthrough in portable radio has been a ‘‘packset’’ system, which Those glasses weighed around 140 grams and were built around two
later became known as a ‘‘walkie-talkie’’, developed in 1937 by Donald battery-powered cathode-ray tubes allowing for the stereoscopic expe-
Hings [51]. rience [60], which was a breakthrough for 1963.
Wristwatches were necessary for the planning and coordination of Next, in 1968, Ivan Sutherland created the ‘‘Sword of Damocles’’
various operations, thus, enabling the mass adoption of wearables in – known to be the first VR Head-Mounted Display (HMD) system,
the prestigious military area [52] and, thus, allowing the marketing enabling users to immerse themselves in a 3-dimensional environ-
teams to adopt wristwatches globally. Simultaneously, the first wired ment [61]. The development has taken almost ten years while the
hands-free devices integrated with flight helmets were being developed prototype was partially see-through and allowing head tracking.
for navy and pilots [53]. Inspired by Edward O. Thorpe, Alan Lewis invented the digital
After World War II recovery, the next big step in the develop- camera-case computer to predict roulette wheels in 1972 [62]. Fol-
ment of wearable technology has been towards the VR by Morton lowing Thorpe’s approach, he used a radio connection between the

4
A. Ometov et al. Computer Networks 193 (2021) 108074

recipient of the information and the person. The recipient used a The first steps for personal and portable electronic assistants were
computer to predict the roulette wheel and whispered the prediction made already in 1994. Mik Lamming and Mike Flynn developed
via a radio link to the hearing aid radio receiver. ‘‘Forget-Me-Not’’, a continuous personal recording system [77]. It was
A significant breakthrough in the development of smartwatches was a technology that recorded the interaction with people and stored this
the appearance of the Pulsar Calculator Watch in 1975 [63]. The first- information in a database for future use.
ever market-produced calculator watch was retailed at a price as high In March 1996, Palm launched the first-ever made mass-produced
as $3, 950. personal digital assistant (PDA) – PalmPilot 1000 [78]. Being essentially
In 1977, Hewlett Packard released its first algebraic calculator a one-chip computer, it has 128 kB of Random-access memory (RAM)
watch [64]. The HP-01 was a genius of miniature and smart design and up to 12 MB of storage. These devices received a 160 × 160 pixel
with 28 keys on the clock display. Four keys are raised for ease of use screen plus a stylus-based text input.
(amount, alarm, memory, and time), and two were embedded, but one The year 1998 could be called the beginning of the wearable
could still use them using the fingers [65]. The rest of the keys were payment epoch, currently present on Apple Watch and Android Wear.
important to press with a pen that instantly snaps the bracelet into the The enabling device was the mBracelet [79]. It was a wrist-wearable
clasp. The cheaper versions, mainly produced by CASIO, still ensure the computer designed for financial transactions with Automated Teller
clock design of the calculator. Machine (ATM). It had three slots that could accept interchange-
During the same year, the first camera-to-tactile vest was designed able iButton buttons. The connection between the mBracelet and the
by the company ‘‘Smith-Kettlewell’’ for the blind in 1977. It took a host was through a three-color Light-emitting diode (LED) grid. The
decade of research. The device used a head-mounted camera to create mBracelet plug-in interface allowed users to exchange messages by
a tactile representation on 10 inches and 1024 points grid located on cross-shaking hands.
the person’s vest [66].
The mid-20th century period could be concluded by one more 2.5. Early 21st century
breakthrough in wearable devices – Portable Stereo Sony Walkman was
released in 1979 as the first commercially available portable personal Levi’s Industrial Clothing Division (ICD) Jacket led the beginning
stereo cassette player with earphones [67]. Although it had two jack of the 21st century, designed by Massimo Osti in collaboration with
outputs allowing for privacy, it was the first luxury wearable with a Philips [80]. The jacket was made of technological material with an
leather case and stylish design. internal network designed to interconnect electronic gadgets. The de-
velopment became revolutionary for its time and influenced the further
2.4. Late 20th century
development of brands such as Acronym and Ma.Strum.
The year 2001 is the most known for introducing the first Plantronics
The wearable development in the 80s passed relatively fast, mainly
M1000 Wireless Headset, followed by the launch of the lightweight
driven by improving existing technology from previous years and a
M1500 version [81]. It was a combination of an M1000 Bluetooth
new AR wave. In 1981, Steve Mann had formalized the EyeTap project
headset and an innovative Bluetooth mobile phone adapter that plugs
and developed the first backpack-like computer designed to process the
directly into the headset jack, giving all mobile users Bluetooth headset
data from a next-to-eye-mounted camera and showed it on the screen in
freedom.
front of an eye [68]. That was the first step toward modern AR glasses
Fossil Wrist PDA met the market in 2003. Its development began in
and the ancestor of Google Glasses [69].
1999 by Donald Brewer, who struggled to make the watch small enough
The next massive acceleration of wearable computing development
for the first year of development. He started the discussions with
was Seiko’s UC 200 WRIST PC introduction in 1981 [70]. It had 2 kB
Microsoft engineers looking for an over-the-wrist platform and con-
of storage and offered the possibility to spell the time and calculation.
centrated on developing ‘‘Smart Personal Objects Technology’’ (SPOT
Despite that, a separate dock station and keyboard were available for
purchase. watches) [82]. After the size has been reduced enough, the screen was
Nelsonic Space Attacker Watch pioneered a new niche of portable attached. The first device had 2MB of memory, which was expanded to
gaming in 1981 [71]. The watch was equipped with two buttons allow- 8MB for the commercial release. The price at debut was $249 US.
ing playing popular arcade games anywhere and anytime. After one In 2004, the first-ever GoPro camera met the world [83]. Nick
year, Nelsonic released portable Pac-Man and, later on, Super Mario Woodman founded the company after surfing in Australia at 2002.
Brothers. The first model was small, light, and waterproof while being AAA
The second big father of Google Glass is the Private Eye head- battery-powered.
mounted display developed and sold by Reflection Technology in Later in 2006, ‘‘Nike + iPod Sport Kit’’ was released [84]. It was a
1989 [72]. It had a monochrome monitor with a futuristic 720 × 280 device that measures and records distance traveled, pace, and more.
resolution for that time. Nike+ consists of a small accelerometer installed or already built into
The beginning of the 90s was lightened by creating The Active shoes, which connects to a Nike+ Sportband receiver connected to
Badge, the first portable indoor location tracker in 1990 [73,74]. It Apple products.
was made by Olivetti Research and was suitable to transmit unique Fitbit was founded in early 2007 by James Park and Eric Fried-
Identifiable Infrared (IR) signals to communicate a person’s location, man [85,86]. In 2008, Fitbit Classic was the first wireless activity
which could be treated as the birth of the Smart Room concept. tracker that could synchronize data with the Internet and have the same
Since the technological development pace was only increasing, Hip- data available on a mobile phone. He was also innovative in the sleek
PC debuted by Doug Platt as a shoebox-sized computer just one year form factor.
later [75]. It was based on Ampro ‘‘Little Board’’ Extended Technol- In 2009, Samsung S9110 Smart Watch was released. The company
ogy module and, together with Private Eye and keyboard, formed an continued Dick Tracy’s idea with a two-band wrist radio [87]. It was a
Agenda palmtop. It already had a floppy drive and many additional dual-band General Packet Radio Services (GPRS) phone with Exchange
extenders. email support. Samsung S9110 Smart Watch was the first smartwatches
Two years later, the development of widely known Knowledge- that included a full-color touchscreen, Bluetooth connectivity, music
based Augmented Reality for Maintenance Assistance (KARMA) has player, and voice recognition feature [88].
started at Columbia University [76]. The system was also utilizing the In early 2012, Eric Migikovsky thought about a device that could dis-
Private Eye for an overlay effect. This project’s main goal was to wire- play messages from selected smartphones (Android and Apple devices)
frame schematics and maintenance instructions on top of whatever was after creating the company Inpulse (Allerta) [89]. The initial version of
being maintained. the watch was attractive by the bold and original design. Moreover, it

5
A. Ometov et al. Computer Networks 193 (2021) 108074

was easy to read in daylight on the ‘‘smartwatch’’ [90]. The design was
highly accepted by the Kickstarter community [91]. After a few gen-
erations of watches and its acquisition by Fitbit, Pebble was removed
from the market, leaving the niche of e-paper display-based watches
abandoned.
In April 2012, information about Project Glass appeared on the
Google Plus social network. The first post of the account was about the
project’s goals: to build a portable computer that will help ‘‘explore and
share the world’’. A video called Project Glass was attached to the post
with the project [92]. Soon after Google showed the concept, people
saw the glasses in real life. Google co-founders Sergey Brin and Larry
Page put on glasses in late spring 2012. At the Google I/O event on
June 27, 2012, Google showed technology in action to a public audience
under the price of $1500 [93]. This milestone is the beginning of VR
and AR mass adoption.
The year 2014 could be described as the period of personal activity
trackers boom. As a continuation and improvement of Fitbit’s story, the
most advanced device of that time, Basis, differed from other fitness
ranges by collecting data such as heart rate, calorie consumption by
activity, several sleep stages, and sweating and skin temperature with
Body IQ technology [94]. The market faced numerous projects, and the
number of people wearing those went sky-high in just a few years [95].
A standalone invention of 2014 was Tommy Hilfiger’s solar power
jacket to charge the phone [96]. Solar batteries were sewn into the
jacket, connected, in turn, to the battery, which was located in one of
the front patch pockets. Two devices could be connected to the battery,
for example, a mobile phone or a tablet. Moreover, solar panels could
be easily detached.
The next breakthrough for end-users was later in 2014 by intro-
ducing Android Wear (currently known as Wear Operating System
(OS)) [97]. This marked the time when the first industrial giant of-
ficially stepped into wearables. Wear OS was the first operating sys-
tem specifically designed for wearable devices, particularly for smart-
watches, and was the beginning of Google’s move towards taking a vital
Fig. 4. Classification of wearable devices based on the on-body location.
position in the wearable market.
Pushed by the technological race, Apple released its first wearable,
Apple Watch, in 2015. While Google was mainly aiming at the market,
the Apple Watch story has a more tragic background [98]. Before 2011, 3.1. Variety of wearable devices
Steve Jobs had long and unsuccessfully fought pancreatic cancer, and
experienced the healthcare industry’s imperfection in the United States Generally, the classification of wearable devices could be outlined
firsthand. He saw how inconvenient it was for the nursing staff to com- from various perspectives based on various factors. Interestingly, de-
municate with patients, how difficult it was to monitor an outpatient’s vices worn and carried can have similar functionality but completely
condition and retrospectively collect the necessary information about different form-factors, technology levels, different on-body locations,
the time spent outside the hospital. It was then that Steve Jobs decided etc. Thus, the broadest classification is based on the application type,
that medical care could be improved with technology, and Apple should even though the other classification groups may significantly overlap.
solve the problem of collecting and structuring data. Existing activity One of the broadest classifications corresponds, but is not limited, to
trackers were not suitable for monitoring patient conditions daily. the following application/functionality types (discussed in more details
That is why Apple came up with the Watch. The integration with in Section 3.7), and sorted alphabetically in Table 1.
a powerful software and cloud platform was supposed to facilitate Another significant factor for classification is related to the de-
doctors’ work. Sadly, Steve Jobs passed away four years before the vice type (without relation to the application area). The types are
device was released. systematized in Table 2 and sorted alphabetically.
The evolution of wearable technology during the period from 2015
The variety of types could be broadened even further by decreasing
till now will be covered in the next section as a state-of-the-art and
the level of generalization. This subsection lists the main concepts
market overview.
present in the market and studied in the literature.
3. State-of-the-art and related work on wearable technologies From the broad adoption perspective, the most intuitive consumer
classification factor is related to the placement of the wearable on
This section overviews the current situation in the wearable technol- the human body. Here, the main groups (concerning device type) are,
ogy development domain. First, it provides the primary classification see Fig. 4:
types of wearable devices in Section 3.1 followed by the underlying
architectures in Section 3.2. Further, we list the data processing and • Head-mounted wearables: Those are mainly focused on percep-
computing techniques in Section 3.3. The discussion on currently avail- tion and control aspects. The group related to vision covers:
able and developing wearable communication technology is provided AR/VR/XR/MR glasses, relaxation masks as well as HMD and
in Section 3.4. The inseparability aspects from the wearable technology personal entertainment systems. Audio-related devices include
perspective are highlighted in Section 3.6. Section 3.7 shows an actual headsets, personal assistants, bass systems. A standalone group is
review of the leading wearable technology research directions from related to neural interfaces.
both academia and industry. Finally, this section is concluded with a • Body-worn devices: Those have much broader functionality and
list of the main directions driven by the mass market. could also be divided into the following subgroups:

6
A. Ometov et al. Computer Networks 193 (2021) 108074

Table 1
Classification based on the wearable application/functionality types (sorted alphabetically).
Type Brief description
Communication functionality (C) Provides the potential not to process the data locally but to exchange it with surrounding nodes and/or remote
cloud.
Control/input functionality (CI) A broad area of input devices ranging from smart buttons to sophisticated gesture recognition devices. This
group’s main task is to extend conventional Human-Computer Interaction (HCI) input focusing on the usability
of the devices keeping a small form-factor as a rule.
Education and professional sports (ES) Aim at improving the education and training by monitoring assistants.
Entertainment, gaming, and leisure functionality (E) The improvement of the perception experience include, e.g., audio systems, personal entertainment displays, etc.
Heads-up, Hands-free Information (H) Extend the conventional ways of the data delivery to the user utilizing personal assistants, AR, XR, Remote
Expert Devices, wearable cameras, etc.
Healthcare/medical functionality (HM) Separated from conventional sensing and monitoring ones due to the need to obtain medical device status that
requires significant effort in the device development and testing as well as providing a high level of the
obtained data trustability and the need for additional certification, however, covering similar devices, e.g.,
Electrocardiogram (ECG), Electroencephalogram (EEG) monitors, relaxation devices, neural interfaces,
exoskeletons, etc.
Location tracking functionality (LT) Requires to have either some Global Navigation Satellite System (GNSS) on board or, at least, a wireless
communication technology. On the one hand, the concept here corresponds to location awareness from the
node’s perspective and, on the other hand, to remote localization of the device if needed.
Notification functionality (N) Ranges from simple vibration notification to complex AR extensions. Similarly to sensing functionality, almost
any personal device connected to the cloud directly or via the gateway can carry this functionality.
Output functionality (O) Various visual, audio, or haptic-enabled devices to provide the user and/or people around with prompt
information from the personal ecosystem.
Safety and Security functionality (S) Personal safety devices, emergency assistants, etc.
Monitoring functionality (M) Extremely straightforward and cheap to implement this functionality. Generally, any device that has an
accelerometer on board can already provide some level of sensing. (Fitness and preventive healthcare —
Activity Trackers, ECG, EEG monitors, etc.)
Wearable devices for pets and animals (PF) Mainly covers smart collars, bark collars, smart clothes, etc.

– Near-body and Sport: A segment for the devices supplement- • Medium-power wearables include devices that may have a small
ing existing wearable ecosystem, such as e-patches, smart display with slightly higher capabilities than low-end wearables.
bands, supplementary activity tracking sensors, etc. They can also have multiple sensors on board with direct or
– On-body: EEG and ECG monitors, posture correcting de- indirect internet connectivity options demanding medium data
vices, safety devices, various smart clothes, etc., form this rates. These devices include smartwatches, fitness trackers, and
subcategory. other gadgets for activity/gesture recognition applications for
– In-body: The most significant niche from a medical perspec- individual, commercial, and industrial purposes [169].
tive includes implantables, smart tattoos, etc. • High-power wearables include devices that are more power-
hungry since they include heavy processing units demanding
• Lower-body devices: This group is still in the infancy but al-
high data rates and large displays capable of performing dif-
ready includes some wearables such as smart shoes, belts, insoles,
ferent compute-intensive tasks such as real-time image/video
etc. Most of them carry specific monitoring functionality for
processing, Machine Learning (ML), etc. Examples include various
professional sport or medical purposes.
headsets, glasses, head-mounted cameras for video crowdsensing,
• Wrist-worn and handheld wearables: Those are the most widely
and others [170–172].
adopted and market-filled niche covering smart rings,
wrist bands, smartwatches, gesture control devices beyond others.
Wearables could also be classified based on the type of battery they
use. Currently, there exist three different types of Lithium batteries
The wearable-placement classification is one of the most natural
available to be used in wearables thanks to the lightweight and high-
ones. Designers, researchers, and early integrators should carefully
consider that their device’s placement is selected appropriately to fulfill voltage characteristics of Lithium [173]. Lithium Coin, also known as
the application requirements listed in the previous subsection. button cell batteries, are one of the earliest batteries developed for
In addition to the classifications mentioned earlier, wearables can wearables such as watches, remote controls, etc. They are lightweight,
also be classified based on their energy-consumption profile. Usually, low cost, and compact in design, mostly similar in size and shape to
devices with displays are comparatively more power-consuming than a coin. However, these were non-rechargeable and disposable batteries
ones without a graphical output interface [2]. However, it also depends that needed to be replaced once depleted, which increases the e-waste
on the nature of the application and the extent of processing they per- problem [174]. Lithium-ion batteries are the most commonly used
form. Therefore, wearables can be broadly classified into low, medium, in wearables such as smartphones, smartwatches, fitness bands, etc.
and high-power wearables. They are rechargeable, lightweight batteries with high-power density.
However, the downside is that they are not very safe for wearables
• Low-power wearables are mostly devices involving low-power since these devices are in close contact with the human body. There
components with limited capabilities that need to operate for a have been instances where Lithium-ion batteries of smartphones have
longer time, mainly for data acquisition/sensing purposes. These exploded due to overcharging/heating. They require special circuitry to
may include different healthcare-related wearables requiring low ensure safe operating voltage and current. Moreover, their performance
data communication rates. For example, a smart ring for sensing degrades over time, even if not used [173]. Lithium Polymer, also
human physiological parameters can be classified as a low-power known as the Lithium-ion polymer, is rechargeable, lightweight, and
wearable since it is a compact device with a small battery, radio, comparatively safer than Lithium-ion. However, it is costly with a
and a few sensors onboard [168]. slightly lower power density than the Lithium-ion [175].

7
A. Ometov et al. Computer Networks 193 (2021) 108074

Table 2
Classification based on the wearable device type (sorted alphabetically and application types abbreviations are from Table 1).
Device Refs. Apps. Power Prof. Brief description
Activity [99,100] C, H, LT, N, S, L-M Simple and relatively cheap devices mainly focus on everyday activity monitoring, including the
trackers M, PP number of steps, basic heart rate, and/or body temperature data collection. The main goal is to
increase the overall physical activity participation of an average user.
AR devices [101–104] C, CI, ES, E, H, H Augmented reality applications can provide additional ‘‘seeing with more eyes’’ information that
LT, N, OP cannot be displayed and is usually hidden from the observer in a see-through manner. The most
attractive areas of the AR development are related to tourism, exhibitions, and manufacturing.
Audio [105,106] C, E, H, N, OP M-H Conventional wired and wireless headphones, bass systems, as well as hearing aids. Moreover,
systems high-quality wearable audio could be integrated as part of XR or MR system to improve the
immersion.
E-Skin (or [107–109] C, CI, M, PP L An artificial skin with mechanical properties of human skin, providing various sensing functions with
nano the main application area of artificial tactile systems. It is commonly located either right on the
patches) human skin or the arms of robotic systems to provide close-to-human perception abilities, e.g., to for
the operation of humanoids.
E-Textiles [110,111] C, CI, ES, N, OP, L-M It is very similar to the e-Skin concept but broadens the opportunities to any close-to-the-body textiles
(smart M, PP that incorporate electronic functionality. Here, the sensors, circuits, or input/output devices are
fabrics) directly integrated with the fabric, allowing for seamless integration of the technology into everyday
garments.
EEG and ECG [112–114] C, ES, M, PP L-M Allow monitoring the user’s health state from both fitness, medical, and professional sports domains,
belts potentially without the need for specialized medical equipment.
FPV, HMD [115–117] C, ES, E, H, N, H Devices for full immersion of, e.g., the Remote control (RC) of various robotic systems teleoperation,
OP human interaction, e.g., police or firefighters, and/or conventional movie watching.
Haptic suits [118–122] C, CI, ES, E, H, H Haptic feedback and capture both motion and biometrics features devices. Full or partial body haptic
N, M feedback systems are built into the suit and can be engaged in actions, on-demand, or in response to
motion capture comparison to provide deeper immersion in various reality applications.
Ingestible [123,124] C, M, PP L-M Objects that go in, through, and underneath the human body or may be a size of a medicine capsule
and and are packed with sensors, microprocessors, controllers, etc. Ingestibles are considered the next step
insertibles of wearable technology and used in healthcare for disease diagnostics and monitoring.
Location [73,125] C, ES, LT, N, S, L-M Functions of remote position estimation of the user. Those are of specific interest for pet owners and
tackers M, PP parents besides the historical crime-oriented market.
Neural [126–130] C, CI, H, N, OP, M Allow for a completely new experience in HCI for both complex medical states of the patients with
interfaces M, PP movement disability, treatment of tactile function, behavior monitoring, and gaming.
Personal [45,87,97,98] C, H, LT, N, OP, L-M Those could be considered as one of the earliest areas of mass wearable devices. When the first
notification S activity trackers received an embedded vibration motor and Bluetooth communications, it became
devices possible to send a simple sign to the user about the incoming call or received message. Today, we
cannot imagine almost any wrist-worn device without this function.
Portable [50–52] C, CI, H, LT, N, M-H Those devices were also taking place in the wearable devices evolution back in the first part of the
Radio OP, S 19th century. Starting with walkie-talkies, we have arrived at the era when surviving a day without
your smartphone could be problematic.
Relaxation [131,132] C, M L-M This group is an interesting set of devices that could be affiliated with luxury or medical purposes
masks but keep the same function of improving the sleeping experience. The devices could also be suitable
for people who travel a lot to improve the day-time adaptation period after, e.g., jet lag.
Safety [133,134] C, CI, LT, N, S, L This group corresponds to a specific set of notification devices but operating vise-versa, i.e., aiming to
buttons M notify either some special units, e.g., police or hospital or the user’s relatives, if something is
happening with the owner.
Smart Bands [135–138] C, ES, H, LT, N, L-M Carry the functionality of modern activity trackers but sometimes also provide gesture recognition,
S, M, PP stress/mood detection, or ECG monitoring functionality.
Smart clothes [96,139–141] C, CI, ES, E, H, M-H A broad segment coupling together various common-looking clothes, ranging from pants to scarfs, but
LT, N, OP, S, M with invisibly embedded features, such as heating, charging, displaying, etc.
Smart contact [142,143] OP L Devices to boost vision and monitor physiological parameters that help track blood glucose level from
lenses the body fluid, i.e., also tears intraocular pressure, with the help of the electronic device’s resistance
and capacitance.
Smart [144–146] C, CI, ES, E, LT, L-M Insoles, shoes, and socks are commonly used to monitor a person’s posture, gait, a number of steps,
footwear M beyond others, and are mainly utilized for training professional athletes from monitoring and
stimulation perspectives and monitoring of children.
Smart gloves [139,140,147–149] C, ES, E, H, M L-M Another hand-held type of wearables is commonly utilized for systems requiring either sophisticated
gesture recognition, rehabilitation, or providing better haptic feedback and other wearable devices.
Smart [150–154] C, LT, N, S, M L-M Luxury jewelry with activity tracking, health monitoring, posture correction, or safety functionality.
necklaces This group of devices did not find much attention due to the actual need for miniaturization and
keeping the appearance high.
Smart [108,155–157] C, M L Nodes consist of a peel-and-stick disposable part that is adhered to the skin and reusable sensor parts.
patches Smart patches are easy to attach, maintain, and remove, acting as an example of Wireless Body Area
Network (WBAN) system utilized in sports and healthcare monitoring.
Smart rings [46,47] C, CI, ES, E, H, L Similar functionality as activity trackers but in a smaller form-factor and without displays. Some smart
LT, N, OP, S, M rings also have a notification device functionality but are kept in a fashionable accessorize form.

(continued on next page)

8
A. Ometov et al. Computer Networks 193 (2021) 108074

Table 2 (continued).
Device Refs. Apps. Power Prof. Brief description
Smart tattoo [158,159] C, ES, M L A set of biosensors implanted under the skin that measure glucose levels and change color depending
on the result. Smart tattoos promise to make life easier for people with diabetes and become an
alternative to permanent blood collection from the finger.
Smart [45,97,160] C, CI, ES, E, H, L-M The most widely adopted wearables after the activity tracker. Generally, it provides almost the same
watches LT, N, OP, S, functionality as a smartphone. However, most smartwatches’ energy efficiency is still challenging
M without the gateway node due to the small form-factor.
VR/XR/ MR [161–166] C, ES, E, H, H Visual immersion in the virtual environment with well-progressing VR applications being
LT, N, OP, M entertainment, education, and healthcare. MR are a particular subset of VR that involve merging the
real world and the virtual world somewhere in a ‘‘continuity of virtuality’’ that augments completely
real environments to virtual ones, as defined in the basic work. Interestingly, the relatively new
concept XR is currently trying to unite all previously known reality-related paradigms.
Wearable [116,167] C, LT, OP, M, M-H Devices are utilized by special units and athletes to record the environment for future analysis or via
cameras PP real-time streaming.

Indeed, the modern wearable electronics market is extensive, and,

thus, the device classification approaches are also manifold. In this
subsection, we have listed the most significant ones and also briefly
highlighted less popular directions. The following sections provide
more details on the devices’ specifics and could also be considered an
extension of the traditional classifications.

3.2. Architectural aspects

Typically, wearable devices use the owner’s smartphone as a gate-

way to connect to the Internet. Therefore, operational efficiency may
be limited if the gateway battery is empty or the connection options are
insufficient. Currently, some wearable devices have a long-range wire-
less radio. However, continuous use is still not recommended due to
high power consumption [2]. The evolution of wearable architectures
is shown in Fig. 5.
Historically, legacy eHealth systems were based on a few specialized
devices, such as biomedical sensors, worn by the patient, and on some
storage and computation facility, either local or remotely located in
the Internet [176] – the first architecture of wearable devices is a
separate device. In such systems, wearables are allowed to monitor
the users’ status by collecting information ranging from physiological
signals (e.g., heart rate, skin temperature, and physical activities) and
their interaction with everyday objects.
At the next step of the architecture evolution, wearables became
connected to more powerful external devices via wired links for data
processing tasks that cannot be executed on the wearables due to
their form factor limitation (Fig. 5A). Such architectures, mainly in
healthcare-related applications, restrict user mobility, which is consid-
ered an obstacle to expanding wearable solutions. The key to smart
healthcare services is the ability to track patients’ health conditions
without the need for frequent visits to medical centers or hospitals
and provide real-time alerts in critical situations [177]. As a result,
traditional wired health monitoring systems have been replaced with
wireless wearable systems.
Due to the mass adoption of smartphones being a part of the
personal wearable cloud, manufacturers started to aim at other con-
sumer wearables as supplements in a personal ecosystem rather than
standalone devices. It resulted in a new architecture where personal
smartphones act as relays or gateways to transmit the data to the cloud
either directly or with pre-processing in the custom applications. Fur-
thermore, this architecture is supported by IoT technologies capable of
bridging in the proximity of the users to a multitude of general-purpose
smart devices. These IoT devices allow to gather further information
concerning the environment surrounding the user and adapt it to the
user’s needs.
According to the IoT architecture provided in [178], the gateways
are also called monitoring stations that are part of the perception layer
for people with disabilities. Logically, with the development of personal
wearable clouds, the logical star-like topology was supposed to join the
communications (see Fig. 5B). However, due to the broad adoption Fig. 5. The development of wearable architectures.

9
A. Ometov et al. Computer Networks 193 (2021) 108074

• Opportunistic (passive) sensing describes periodic action to gather

information, determined by elapsed time, distance, or other met-
rics from the last data gathering.
• Opportunistic mobile social networks refer to self-organizing, point-
to-point networks of devices sharing the information between
themselves. The devices create the networks based on the vicinity
to each other. Such networks can additionally assist other func-
tions of the wearable, including positioning or user protection.

3.3.2. Pre-processing
Even though the central part of data processing takes place in
different computing layers [181], sometimes minor filtering processes
are executed in wearable and IoT devices. The early pre-processing
stage is essential for wearable and IoT devices since they do not have
enough computational and storage resources to process ‘‘unnecessary’’
data.
Fig. 6. Wearable data processing.
It is essential to highlight that the collected data contain incomplete
and duplicate information, errors, inconsistencies, etc. It requires an
early preparation before the data processing (locally or in the Cloud).
of Bluetooth technology as a leading connectivity enabler, wearable Thus, once the data is collected, it should be filtered, structured,
devices started to communicate purely in a centralized way using the cleaned, and validated to improve the data’s quality. As a result, the
gateway devices as masters. data processing step will take less time, and the output information
Most fitness, sports, and healthcare wearable applications follow will deliver better decisions.
the fourth evolution step in Fig. 5A, while sharing their data with During the data preparation, some steps take place to enhance
applications installed on the smartphones. This option implies several the data and avoid unnecessary processing. For instance, gathering
situations where wearables cannot connect to the network, such as information is devoted to finding the most ‘‘representative’’ data from
the nonexistence of a smartphone nearby, non-installed applications the collected set. It will be used throughout all the data processing.
on the smartphones, or an outdated phone software [179]. Considering The cleaning and validation are then essential to detect the quality
such possible scenarios and the end-users ever-increasing expectations issues, including duplicate values, outliers, missing values, and incon-
and needs, device manufacturers consider equipping wearables with sistencies. Once these issues are detected, multiple techniques can be
short-range and long-range connectivity chipsets. As a result, these used to remove missing values, delete outlier samples, merge duplicate
standalone wearables can function independently from other devices. data, etc.
Although this standalone wearable-based architecture provides a
way towards separating wearables from other personal devices and an 3.3.3. Data transfer
opportunity to communicate directly with the cloud, it brings some The data transfer phase is an essential part of the wearable device
challenges in terms of network design and dimensioning. For instance, data chain, and it is further discussed in detail in Section 3.4. At this
additional loads on the wireless networks are expected due to the lack point, we only note that many wearable solutions limit the amount
of pre-processing, usually performed by the gateway devices. Another of power spent during the data transmission by pre-processing it to
challenge is related to the deployed communication technologies op- decrease the actual amount. For example, it is achieved by applying
timized for low-power operation, thus, long battery life and reduced compression schemes to the data before transmission [37]. Addition-
device complexity. ally, the power constraints usually limit the transmission’s security, as
Today, most wearable devices have at least two wireless interfaces proper data encryption is often computationally demanding.
on board. The trend of unifying the operation is expected to push
3.3.4. Computing paradigms
vendors to add more wireless modules, with the possibility of intelli-
Based on architectural development, it is reasonably straightforward
gent technology selection mechanisms based on improved situational
to understand the specifics of wearable data processing. Computation
awareness.
and operation on bulky and raw data was a challenging task. Thus,
during the firsts steps of the architectural evolution, wearables were
3.3. Data processing
mainly utilized as data collectors. For example, ECG monitors were
given to patients for a few days to collect the data on the heart’s
Wearable-based data are gathered and processed in large quantities rhythm and electrical activity. After that, the devices were connected
depending on the individual product, yet it is always similar at its core. to a personal computer with a wire for post-processing. This phase
Fig. 6 depicts the basic cycle of wearable data. The individual stages of could be treated as the introduction of Multi-Access Edge Computing
wearable data processing are characterized below before the details on (formely Mobile Edge Computing) (MEC) for wearables, i.e., moving
the data transmission aspects. heavy computations to the closes network device instead of local
computations.
3.3.1. Data collection With the smartphone’s introduction as a gateway, wearable data
Wearable devices primarily collect and process user-generated data, processing on it was adopted swiftly. Following the vendors’ foot-
and collective data collection is commonly referred to as crowdsourc- prints, most application developers aimed to save battery-constrained
ing, a powerful tool for collaboratively retrieving a large amount of wearable lifetime by moving the smartphones’ computations. The Edge
data. Crowdsensing is a way to enable the coordinated data gathering operation hardened. Simultaneously, some AR/VR devices still do not
from the devices. There are three standard sensing techniques [180]. have the required processing power or tend to overheat. Thus, the only
option is to utilize a powerful Edge device for computing in trade-off
• Participatory (active) sensing refers to obtaining information to transmission overheads.
through the user’s action, such as measuring one’s exercise on Today, wearable devices mainly follow Edge and Cloud computing
their smartwatch. paradigms, as well as Fog Computing [182,183]. The latter one is

10
A. Ometov et al. Computer Networks 193 (2021) 108074

coined by CISCO in [184,185] and bridges the gap between the cloud data, such as clustering, regression, classification, among others [190].
and end-devices, allowing the computation, storing, networking, and Additionally, these ML techniques can be run by using batch process-
data management near the target computationally-weak devices. In ing [191], real-time processing [192], and online processing [193].
contrast to Edge computing, which handles data on a nearest network Generally, the most significant part of the data collected by wear-
Edge node, Fog Computing processes data on a more powerful Fog node ables are time series [194]. This kind of data could be typically used
in a more general way, i.e., to equip the network Edge and network for classification purposes, anomaly detection, or forecasting. How-
devices with virtualized services in terms of processing and storage ever, there are few obstacles connected to the analysis of time series.
along with offering network services. To this end, the Fog found little The main one is the limited amount of data, which can be dealt
integration on wearables due to their proprietary communications/- with using data augmentation techniques. The other issues are limited
computing nature and lack of knowledge about the devices in proximity computation capacity together with limited power resources when the
and related communication aspects. computations need to be done directly on the hardware.
Cloud, Fog, and Edge computing are three levels that simplify Additionally, in the case of time series, other aspects should be con-
IoWT. Nevertheless, there are many more paradigms suitable for the sidered, such as the occurrence of class imbalance and the possibility of
integration with wearables in reality [186], and the whole cloud system multivariate time series. Unfortunately, the commonly applied method
is much broader and more complex [187]. The actual classification of augmentation could not be transferred from the image and speech
also covers Mist Computing, Mobile Computing, Mobile Cloud Com- domain into time series. This created demand to elaborate on the
puting, Mobile ad hoc Cloud Computing, Multi-access Edge Comput- augmentation method for time series data [195]. The categories of the
ing, Cloudlet Computing, and Transparent Computing [188]. In this time series’ augmentation techniques include six domains, i.e., time do-
section, we overview these paradigms and explain their peculiarities main, frequency domain, decomposition-based methods, model-based
from the wearable technology perspective, whereas the implementation methods, learning-based methods, and time–frequency domain.
challenges of computing paradigms will be discussed later in Section 4.
The recorded time series gathered by wearables contain a special
pattern, which indicates, for example, human behavior or physiological
• Mist Computing is operating at the Edge of the network, usually
consisting of energy-constrained microcontrollers and sensors. It disturbances, potential pathologies. For creating a predictive model,
is a network without data processing. In this case, the data is it is necessary to use special ML techniques. Nonetheless, the pre-
generated where it interacts with the physical world. processing steps are also needed to obtain the final expected results.
• Mobile Computing is a technology that allows transferring data, In terms of architectural choices, Convolutional Neural Network (CNN),
voice, and video through a computer or any other wireless de- Long Short-Term Memory (LSTM), Bidirectional Long Short-Term Mem-
vice (e.g., wearable, mobile phone, laptop, tablets). Mobile com- ory (BLSTM), Multilayer Perceptron (MLP), Gated Recurrent Units
puting is used to create context-sensitive applications (e.g., (GRU), and transfer learning are robust in obtaining the state-of-the-
location-based reminders). art results for wearable data [196–199]. Additionally, the usage of
• Mobile Cloud Computing is a platform that brings together mobile classifier XGBoost was registered for the scalars’ values [200]. How-
devices and cloud computing and creates an infrastructure where ever, it required earlier feature extraction steps. The LSTM and BLSTM
both data storage and processing take place outside of a wearable architectures are suitable for this purpose because they can learn the
device. It allows accessing the cloud remotely using a mobile long-term dependencies occurring in the signal [201,202]. 1D-CNN
device. allows the algorithm to extract features and learn in time series what
• Mobile ad hoc Cloud Computing refers to a group of mobile (wear- is expected for such purpose [203].
able) devices in the proximity willing to share their resources, After the data processing stage, the useful information is exposed to
taking some incentives and collectively solving the task. It is the user through reports, graphics, tables, etc. It is then available for
usually deployed over mobile ad hoc networks and performs further storage, applications, and development.
computation-intensive applications using other portable devices’
resources in specific scenarios, such as natural disasters, sports 3.3.6. Data storage
broadcasts, etc. In this last step of data preparation, the altered data is stored for
• Multi-access Edge Computing (formerly Mobile Edge Computing) further analysis, which is then used by recommendation systems for
pushes the boundaries of Edge Computing by providing com- marketing purposes and collective intelligence to improve the decision-
puting and storage resources next to low-power, low-resource making process.
mobile devices. It ensures connectivity and computing opportu-
nities through wireless networks. In contrast, Edge Computing
3.4. Communication technologies
works solely between the mobile device and the closest powerful
network node. This computing paradigm could be defined as an
The wide range of wearable devices and related technologies allows
enabler for future wearable systems.
for various supported connectivity solutions, defined by the wearable’s
• Cloudlet Computing is an infrastructure that uses small data cen-
requirements for range, data rate, power restrictions, mesh types, devel-
ters to move the functionality of the main data centers and
bring the cloud closer to service consumers (being, essentially, oper’s preferences, and numerous other aspects. Data transfer specifics,
wearables) [189]. Cloudlets represent virtual machine infrastruc- including encryption level, coding and transmission schemes, modu-
tures that operators place between devices and the cloud. This lation, and cyclic prefix, are also individually defined depending on
paradigm is mainly focused on real-time applications for resource- the utilized technology. The most commonly used data transmission
constrained devices. technologies in wearables include Near Field Communication (NFC),
Bluetooth Low Energy (BLE), Wireless Fidelity (Wi-Fi), ZigBee, Low-
Indeed, the variety of options to efficiently execute computing Power Wide Area Network (protocols) (LPWAN), and other cellular
between mobile nodes is vast. or non-cellular IoT transmission technologies [204]. As illustrated in
Fig. 5, both short-range and long-range communication technologies
3.3.5. Data processing are being deployed in wearable networks. This subsection outlines
During the data processing phase, different techniques and methods the related communication technologies already available on modern
are applied to the input data to obtain meaningful information. Nowa- wearables and promising candidates widely discussed to be integrated
days, ML techniques are broadly applied to analyze and process the into the devices.

11
A. Ometov et al. Computer Networks 193 (2021) 108074

3.4.1. Short-range communication technologies introduced in The 3rd Generation Partnership Project (3GPP) Release
Generally, wearable devices can communicate with each other in 13 specifications, both technologies are optimized to communicate
a Peer-to-Peer (P2P) manner and with the gateway nodes using short- small amounts of infrequent data with minimal power consumption and
range technologies in mobile wearable networks. maximal coverage. Taking into consideration the traffic requirements
Bluetooth (and its currently broadly adopted version BLE) is a of wearable applications that might be higher in comparison with smart
short-range communication protocol for Personal Area Network (PAN) city or other more common Low-Power Wide Area (LPWA) use cases,
working at a maximum transmission distance of 100 m and the 2.4 GHz several enhancements to the NB-IoT and LTE-M standards are suggested
unlicensed Industrial, Scientific, and Medical (ISM) band. A piconet in 3GPP Release 14 specifications to offer higher data rates while still
is the most straightforward pattern of a Bluetooth network, and it is consuming less power.
composed of the master of the connection (commonly, a gateway node), Nonetheless, presently deployed in cities, LPWA technologies are
and a maximum of seven served active slaves (clients) [205]. already utilized for cases of long-range communications and low energy
The second primary wireless short-range technology is defined by consumption requirements [215]. The first target of these technologies
the IEEE 802.11 standard, also known under the market name of was low-end IoT devices such as sensors. However, LPWA connectivity
Wi-Fi, which aims to provide connectivity to mobile devices within solutions have started attracting other application scenarios used in
the Wireless Local Area Network (WLAN). A direct connection between wearable systems. An example of non-lincensed LPWA technology is
devices is allowed by the ad-hoc mode foreseen by the standard, Long Range LPWAN protocol (LoRa). It is a long-range wireless tech-
nonetheless, the lack of efficient power saving and enhanced Quality of nology that operates in the license-free ISM radio band at 868 MHz and
Service (QoS) has pushed the Wi-Fi Alliance to make a move towards offers transmissions up to 25 km operational distances. This technology
the Wi-Fi Direct option, which handles the P2P communications more has been deployed mainly in e-Health applications, such as temperature
efficiently [206]. sensor-based wearable devices, to evaluate and predict heart failure
Moreover, the current development of AR, VR, and XR technologies events in high-risk patients [216]. Another alternative LPWA technol-
dictates the need for higher throughput and lower delays. Developers ogy used by proprietary wearable devices is Sigfox, which, similarly
are opting for technologies that operate in high and ultra-high fre- to LoRa, operates at the sub-GHz ISM band and provides extended
quencies, e.g., Institute of Electrical and Electronics Engineers (IEEE) coverage. While LoRaWAN is based on a spread spectrum technology,
802.11ad, IEEE 802.11ay, and IEEE 802.11ac that can provide Gigabit Sigfox is a narrowband technology owned by the Sigfox company in
data rates in millimeter-wave band (mmWave) [207]. Moreover, the charge of deploying their networks [217].
use of these high frequencies has its propagation limitations, which Table 3 provides additional technical details on the short-range and
is ultimately an advantage for dense packing scenarios, where human long-range communication technologies that are deployed in wireless
bodies act as a natural barrier and allow better utilization of spatial networks for the support of wearable applications. It lists most modern
reuse [208,209]. ISM-band technologies ranging from conventional Wi-Fi to Wireless
The Terahertz (THz) frequency band ranging from 100 GHz to 10 Gigabit Alliance (WiFi at 60 GHz) (WiGig) as well as paid-spectrum
THz offers substantial bandwidth, theoretically enabling capacity in the ones. In summary of this section, despite the conventional operation of
order of terabits per second at negligible latency. The THz frequency wearable devices via relays, the next-generation wireless networks aim
band has been experimentally utilized for communication purposes at to provide standalone wearables with the opportunity of establishing
short and medium ranges [210]. Currently, the THz frequency band direct communications with the cloud.
is a subject of exploration for point-to-point communication among
regulators, operators, and manufacturers. Visible Light Communica- 3.4.3. Device-to-Device (D2D) -based communication for wearables
tions (VLC) is among the connectivity solutions that operate in the THz Multiple works propose to support wearables through D2D commu-
frequency range, more precisely from 400 and 800 THz (780–375 nm). nication to increase the wireless networks’ capacity and coverage by
This technology can simultaneously achieve illumination and commu- enabling immediate communications between the users.
nication between two or more devices in proximity, thus improving
energy efficiency using existing lighting infrastructure [211]. 3.4.3.1. Task caching and offloading. The research community has ex-
The wearable communication technologies with the shortest range tensively discussed the rapid growth in the amount of data and the
correspond to Radio Frequency Identification (RFID)/NFC systems. number of computations on wearable devices. For example, the au-
RFID nodes can be worn by workers to provide them with hands- thors of [218,219] introduced a data-driven resource management
free operations and allow flexible and mobile asset tracking by being framework to perform service-aware resource allocation and improve
attached to gloves, hats, or helmets [212], and operating on the same resource ratio utilization with D2D communications. More specifically,
13.56 MHz frequency [213]. real-time communications between virtual reality and other wearable
More advanced opportunities are brought to wearables by long- devices are based on Edge caching. In this scenario, using D2D tech-
range technologies. In contrast to the short-range communication (with nology improves the cache hit ratio by considering mobile caching
short-range technologies enabling a longer battery life, cheaper and and small cell caching – Base Station (SBS). Besides, D2D communica-
less complex devices [214]), long-range ones eliminate the application tions increase wearable devices’ bandwidth in ultra-dense 5G and B5G
selection limit, and wearable devices can communicate directly with cellular networks.
an access point, a base station, or an edge node. As a result, some Moreover, the researchers have sufficiently covered task offloading
companies provide solutions for wearable devices with the opportunity strategies to minimize processing time and, therefore, save power for
for either physical or embedded subscriber identification module (SIM) wearable devices with low battery levels. The authors of [220] propose
cards to integrate long-range technologies. These solutions include using the D2D link to create mMTC and, thereby, overcome the limited
models from Apple, Samsung, LG, and Xiaomi. resources of mobile devices. A similar approach is used in [221],
where nearby wearable devices form a local mobile cloud and thus
3.4.2. Long-range communication technologies provide low latency. In [222], Dust (prototype code uploading D2D)
Motivated by the increased popularity of these solutions, 3GPP uses Google Glass as a prototyping device and offloads tasks using
introduced its Massive Machine Type Communications (mMTC) tech- D2D and the cloud. The authors of [223] provide computational of-
nologies for low-power operation and reduced-complexity devices, floading recommendations for AR applications running on wearable
namely Narrowband Internet of Things (NB-IoT) and Long Term Evolu- devices. In [224], the authors propose a data offload platform for
tion (LTE)-M [179]. The introduction of these technologies paved the other nearby applications and emphasize that Wi-Fi Direct D2D binds
way for the usage of cellular modems in standalone wearables. First to maintain high data rates for locally linked groups compared to

12
A. Ometov et al. Computer Networks 193 (2021) 108074

Table 3
Wireless technologies comparison for wearable communication.
Communication technology Topology Frequency bands Range Data rate Power profile
RFID P2P 125–134 kHz, Up to 100 m Varies with frequency Very Low
13.56 MHz,
860–960 MHz
NFC P2P 13.56 MHz <0.2 m Up to 424 kbps Very Low
Short-range
BLE (IEEE 802.15.1) P2P, mesh, star 2.4–2.48 GHz Up to 100 m Up to 24 Mbps Low
Zigbee (IEEE 802.15.4) P2P, mesh, star 868–868.6 MHz, Up to 100 m Varies with frequency Very Low
902–928 MHz,
2.4–2.49 GHz
Wi-Fi (IEEE D2D, mesh, star 2.4–2.48 GHz, 20-250 m 2-600 Mbps Medium
802.11a/b/g/n) 4.9–5.8 GHz
Wi-Fi 5 (IEEE 802.11ac) D2D, mesh, star 4.9–5.8 GHz Up to 70 m Up to 3.5 Gbps High
Wi-Fi 6 (IEEE 802.11ax) D2D, mesh, star 1–6 GHz Up to 120 m Up to 9.6 Gbps High
WiGig (IEEE 802.11ad/ay) D2D, mesh, star 57–70 GHz 10–100 m Up to 20 Gbps Very High
VLC (IEEE 802.15.7) D2D, mesh, star 400–800 THz Up to 100 m Varies with frequency Very High
NB-IoT Star-of-star LTE frequency bands Up to 15 km Up to 250 kbps Low

Long-range LTE-M D2D, Star-of-star LTE frequency bands Up to 10 km Up to 1 Mbps Low

LoRa Star 867–869 MHz Up to 50 km 50 kbps Very Low
Sigfox Star 868–878.6 MHz Up to 50 km 100 bps Very Low

Bluetooth and BLE. The context-sensitive upload structure (to cloud or 3.4.3.4. Social-aware communication. Beyond the conventional connec-
smartphone) was developed in [225] to achieve a trade-off between tivity and interaction between heterogeneous IoT and IoWT devices,
the usability of wearable devices and the energy savings on the smart- the concept of Social Internet of Things (SIoT) supports a plethora of
phone. As a result, the power consumption of unimportant tasks and socially-driven collaborations among objects [235,236]. The synergy of
the waiting time for urgent tasks are significantly reduced. social networking and IoWT paradigms offers some benefits and allows
devices to socialize, collaborate, and establish group-communications
3.4.3.2. Software-defined D2D communication. To combat a problem
according to the relationships. For instance, Social object Relationship in-
of computational constraints and other critical issues faced by wear-
cludes conventional human social relationships (e.g., owners of devices
able communications (see Section 4.7), the authors in [214] propose
multiple-layer communication architecture, which incorporates D2D are friends on Facebook, Linkedin, etc.), Ownership object Relationship
cloud/edge technologies, Software-Defined Networks (SDN), brings together devices held by an owner, Co-Work Object Relation-
Heterogeneous Cloud Radio Access Network (HC-RAN), Multiple Input ship integrates objects that work together to provide service for a
Multiple Output (MIMO), and mmWave technology, among others. common IoT application, Parental Object Relationship clusters objects
In [226], SDN architecture for supporting D2D communication between belonging to the same production batch, whereas Co-Location Object
nodes of Wireless Sensor Network (WSN) is proposed to maintain Relationship combines devices in the same place. The concept of using
connectivity even in the case of the cellular infrastructure failure or elements of social networks in IoT has attracted an unprecedented
when it becomes partially unavailable. The authors particularly exploit amount of attention from the research community [237,238]. Not
D2D technology to build a mobile cloud and perform computational surprisingly, most of the research contributions in this field focus
tasks on nearby devices. In the event of a disaster, the SDN controller on devices and their owners, their particular aims and needs, and
selects a multi-hop routing path. Here, the Cluster Head (CH) on each trustworthiness.
mobile cloud associates with other CHs using inband D2D to convey the An overview of the integration of social-awareness and D2D com-
information of out of coverage devices to the other part of the network. munication is presented in [239]. More precisely, the authors offer
Further, SDN based multi-standard cognitive radio is studied in [227] state-of-the-art, socially-driven D2D communications by analyzing re-
for heterogeneous wearable wireless networks. It can control, config- lay discovery and peer selection, communication mode selection, spec-
ure, select, and decide on switching between multiple technologies trum resource allocation, and management when social concepts are
based on the specific requirements in D2D communications. applied to them. In [240], the authors focus on social-aware D2D
communications in 5G cellular networks using matching theory for
3.4.3.3. Millimeter wave-based D2D communications. Using the mmWave
caching problems. The authors state that the combination of D2D
link is a tempting approach to achieve high data rates in wearable net-
works [228,229]. Moreover, the industry recognizes the importance of and content caching can considerably improve system performance.
mmWave technology for high-quality wearable devices [230]. For ex- The work [241] presents a trust-based solution, where reliability and
ample, the IEEE 802.11 working group is looking at high-performance, reputation of devices based on SIoT concept are used for effective
mmWave-based wearables in public places as a possible use case. D2D enhanced cooperative content uploading in NB-IoT. In [242], D2D
As pointed out in [231], mmWave is an ideal candidate for D2D in communication is applied to acquire the context-aware information and
very dense scenarios because it requires directional transmission to integrate it with networking information elaborated at BS for infor-
overcome high path losses, which in turn reduces interference between mation diffusion time estimation in the The Fifth Generation Cellular
adjacent lines [232]. Meanwhile, D2D is also beneficial for mmWave Network Technology (5G) network. In [243], the authors combine the
picocells due to its high attenuation, especially for users at the cell’s Diffie-Hellman Key Exchange (DHKE) algorithm with the SIoT concept
edge. Open areas of research in mmWave D2D communications, such to efficiently allocate network resources to reliable nodes. In [244],
as gaps in ML, channel modeling, D2D clustering techniques, green con- a D2D-based group technique to operate in sensor network environ-
nectivity, and user mobility in mmWave D2D networks, are discussed ments, which contains a vast amount of devices, is proposed. This
in [233]. Another interesting topic is covered in [234], which discusses approach allows reducing data congestion and ensuring outstanding
relay selection in mmWave D2D with dynamic obstacles. power efficiency.

13
A. Ometov et al. Computer Networks 193 (2021) 108074

3.5. Localization To summarize, choosing a suitable localization technology depends

on the application requirements. Table 4 compares some of the most
In recent years, the importance of wearable devices in our daily popular wireless technologies that can be used in the localization of
lives to provide location and tracking of users within an environment wearables. Some technologies are already present in wearables (GPS,
and their interaction with the plethora of Location-Based Service (LBS) Wi-Fi, or BLE) but have accuracy in the range of dm to m. UWB and
has increased considerably. However, the wearables that perform these RFID-based localization technologies are relatively accurate but are
features present performance limitations in some scenarios and appli- infrastructure-based.
cations. Therefore, researchers are developing new positioning systems
considering the latest wearable devices’ strengths and limitations to
counteract these limitations and increase the positioning performance. 3.6. Interoperability
These new solutions include the implementation of new technologies,
techniques, and methods.
Given the fast growth of wearable devices that use positioning The integration of multiple subsystems is the basis for modern IoT
and localization services, it is common to find them equipped with and IoWT systems. However, it generates new interoperability prob-
GNSS chips. Currently, GNSS-based navigation systems have about 158 lems and requires additional attention from the industry and research
satellites in orbit for positioning purposes. communities to create seamless programmability of the various things
Hence, some wearable devices can estimate the user’s position by to connect, collaborate, and effectively exchange the data among
using GPS or other constellations. Some companies are implementing them [277]. Moreover, interoperability faces diverse challenges under
new technologies to provide highly reliable positioning services in different scenarios in various fields and domains.
wearable devices. For instance, in order to reduce the power consump- Different factors can trigger interoperability issue [278]. For in-
tion, Sony Corporation will incorporate a GNSS receiver in low profile stance, integration between different manufacturers, limited communi-
devices when dual-frequency is used to compute the user or device cation between different transport protocols, and no established rules
position [245]. or standards at the application level can prevent seamless integration.
We can distinguish two main classes of positioning systems based on
In this context, it is essential to manage the intrinsic heterogeneity
the main wearable architectures: infrastructure-based and
of IoT systems and provide effective solutions for seamless interoper-
infrastructure-less [246]. On the one hand, infrastructure-based sys-
ability among smart devices, sensors, etc. Accordingly, interoperability
tems are those systems that depend on a specific infrastructure to
is investigated from various perspectives, including device, technology,
estimate the position. The infrastructure can be deployed exclusively
network, syntactic, semantic, and platform [279].
for the system or be part of another system designed for a different
purpose [247,248]. On the other hand, infrastructure-less systems work
without requiring any surrounding infrastructure [249,250].
3.6.1. Device interoperability
With respect to the application environment, wearable positioning
systems are classified as indoor and outdoor positioning systems. In IoWT is composed of a wide variety of smart objects or things
indoor environments, applications demand high accuracy position com- ranging from high-end to low-end devices. The former have sufficient
pared to those used outdoors [251]. Additionally, each environment resources and computational capabilities, whereas the latter type of
presents particular challenges for positioning systems, for example, devices is generally resource-constrained. However, the devices, which
the occlusion of view of the skies due to ceiling and signal multi- are different in terms of computational capabilities and use differ-
path due to the complex geometry of indoor environments. Therefore, ent communication technologies, may need to exchange information.
depending on the type of environment, the positioning technologies This challenges interoperability among various types of heterogeneous
and techniques implemented are diverse to face the environmental devices of the IoT ecosystem. Device interoperability is related to
conditions [252]. the data exchange between heterogeneous devices and communication
In outdoor wearable positioning, the Global Positioning System protocols and the possibility of integrating new devices into any IoWT
(GPS) and the cellular networks are the most used technologies [252, platform [279].
253]. Regarding cellular networks, in each of their generations, the po-
sitioning accuracy is improved. Since the second to the fifth generation,
the actions carried out to improve the positioning accuracy are: use 3.6.2. Network interoperability
the cell-ID localization technique, timing via synchronization signal, The integration of the IoT heterogeneous systems requires interop-
specific reference signal designed for localization, and using all the erability at various protocol levels to work together [278] seamlessly.
layers of protocol stack [254]. The technologies most used for indoor This interoperability type deals with mechanisms that enable seam-
positioning are: BLE, Wi-Fi, RFID, NFC, Ultra-Wide Band (UWB), Light,
less, ubiquitous connectivity. To truly ensure network interoperability,
Computer Vision, Magnetic Field, Sound, Dead Reckoning, and Tactile
each system must be able to provide the data exchange with other
Odometer [255].
systems over different networks by focusing on such network lay-
In addition to traditional positioning systems, more complex sys-
ers functions as addressing, routing, mobility, security, and resource
tems have been proposed to enhance positioning accuracy and perfor-
optimization [280].
mance. For example, Collaborative Positioning Systems utilize neigh-
boring wearables/users in the positioning system method, and systems
based on sensor fusion combine diverse sensor information to reduce
3.6.3. Syntactic interoperability
the estimation error of position [256].
Localization can significantly enhance the capabilities of wearables. As mentioned above, interoperability reflects the ability to exchange
For instance, it can help users navigate and provide clues to devices information transmitted between systems in the form of messages. In
about interacting with users based on their location. Besides, location this context, a message can be studied on the syntactic level, where
services play a crucial role in emerging fields such as assisted living, its internal properties are considered. Syntactic interoperability refers
smart homes, and smart cities. Thus, the recent Location-Based Service to the interaction of format and data structure used in any exchange
(LBS) must be compliant with the current legislation in terms of privacy between the IoT heterogeneous objects and can be reached through
and security [257]. defined data and interface formats and encodings [277].

14
A. Ometov et al. Computer Networks 193 (2021) 108074

Table 4
Comparison of localization technologies for wearables.
Technology Refs. Accuracy Maximum Power Localization Advantages Disadvantages
range consumption approach used
GNSS (GPS) [258,259] 10 cm Global, 0.5–2.5 W Time-based Global, widely-available. Only outdoors, low accuracy,
outdoors high power consumption.
Wi-Fi [246,260–262] dm- or 250 m outdoors, <1 W Fingerprinting, Widely-available, low-cost, often Good-accuracy localization
m-level 50 m indoors AOA, TOA does not need dedicated methods based on fingerprinting
infrastructure. require extensive training,
relatively low accuracy.
UWB [263–265] Order of cm 300 m 100–500 mW TOA, AOA High accuracy and precision, Not widely-available (yet), needs
moderate costs. dedicated infrastructure.
BLE [266,267] Order of m 100 m <50 mW Fingerprinting, Widely-available, ultra-low-power, Low accuracy.
TOA protocol stack suitable for the
IoT.
Non-visible [268] Sub-mm 10 m N/A TOA Very high accuracy. Requires LOS, expensive.
light
Visible light [269,270] cm- or 1.4 km N/A TOA Can use ambient light sensors Requires Line-of-Sight (LOS),
dm-level commonly available in high-accuracy solution requires
smartphones. dedicated hardware.
LoRa [271,272] m-level 20 km <100 mW AOA, RSS Ultra-low-power, low-cost, Low accuracy.
long-range, designed for the IoT
and sensor networks.
RFID (passive) [273,274] cm-level 15 m – AOA, POA Ultra-low-power, low-cost, Expensive infrastructure, small
high-accuracy. coverage area.
RFID (active) [125,275,276] m-level 150 m <100 mW RSS Low-cost, medium range. Moderate accuracy, need to be
powered.

3.6.4. Semantic interoperability 3.7. Mass consumer directions

The need for semantic interoperability emerges when heteroge-
neous devices use different syntax to encode and decode their in- In this section, we first overview the main wearable development
formation [281]. For instance, reliable communication between the activities happening in the industry and proceed with the literature
source and gateway devices can be interrupted if a sink device fails to review on recent academic works in the area.
decode a source data format into an understandable form. In addition to Generally, the wearable devices’ market is flooded with new so-
the fact that the data generated by wearables may have defined data lutions developed at both the academic and industrial levels. This
formats, the data models and schemes exploited by different sources subsection provides a vision of a variety of prototypes and research
are usually dissimilar and not always compatible [279]. Further, the directions present in the industry. The next subsection would give an
data may consist of different information and be represented in other overview of the leading academic projects and activities. Note that this
measurement units. This semantic incompatibility between the data subsection particularly focuses on projects that are not yet present on
and information models leads to IoWT systems’ failure to dynamically the market. A detailed overview of ready-for-purchase devices (together
and automatically inter-operate (as they have different descriptions with prototypes) is given in Appendix.
or understandings of resources and operational procedures [282]). To Moreover, today’s mass-market of wearables is driven by corpo-
this aim, semantic interoperability enables different agents, services, rations aiming to push the adoption of high-end wearables to cus-
and applications to exchange information, data, and knowledge in a tomers. For example, Samsung is the most prominent owner of wear-
meaningful way, on and off the Web.
able patents, counting 3421 [284] active ones. At the same time,
Samsung was the global leader in the number of granted patents in the
3.6.5. Platform interoperability
years 2010–2018, counting 796 [285] granted patents.
Platform interoperability issues arise in IoWT due to different oper-
While focusing on other industrial giants together with Samsung (Mi-
ating systems, programming languages, data structures, architectures,
crosoft, LG Electronics, Sony, Alphabet, Epson, Philips, BOSE, Apple,
and mechanisms for things and data. The lack of platform interoperabil-
and OPPO Electronics), their intellectual property concerns more than
ity can lead to the inability to connect IoWT devices to heterogeneous
50% wearables. Therefore, more applications of market-available wear-
platforms, developing applications that use multiple platforms, and
ables include monitoring of disabilities, education, enterprise, finance,
delaying the implementation of IoWT technologies on a large scale. In
contrast, cross-platform interoperability will ensure broad collaboration gaming, and entertainment. In general, the Asia–Pacific market is the
between platforms to offer the best solutions in different areas [283]. most active in the world in this area.
To this end, one may need to acquire comprehensive knowledge about
Application Programming Interface (API)s and information models of 3.7.1. Entertainment and AR/VR/XR
platforms to adjust applications from different platforms [279]. Concerning entertainment, another fast pacing wearable market
Although IoT interoperability issues are addressed by several aca- niche is related to VR devices. The adoption of said technology was
demic and industry studies, there is still no proper foundation that can previously limited by hardware and communication issues, i.e., low
cover some related research issues discussed later in Section 4. The computational power and high bandwidth that were not reachable with
lack of standards and the absence of cutting-edge technologies slow IoT conventional wireless systems. Today, several VR headsets could be
development. Therefore, interoperability is a challenge that needs to be purchased with reliable wireless connectivity while keeping the user
extensively studied to provide consolidated, reliable, secured, scalable, experience on a high level [208].
and adaptable solutions for IoT networks. Seamless IoT device inter- Companies are also integrating positioning and heading into their
operability is believed to push and open the massive opportunity in state-of-the-art devices. A clear example is Sony, which is developing
the IoT market, accelerate industry innovation, and help developers and a HMD. The contents on the HMD are adjusted depending on the user’s
companies produce solutions quickly. pose and position to provide a full augmented reality experience for

15
A. Ometov et al. Computer Networks 193 (2021) 108074

gaming on other applications [286]. Thus, the position and heading are through the on-board headphones changes accordingly. A new wave
essential, but the head tilt is needed to re-allocate the elements on the of head-mounted wearables is related to VR. During Mobile WOrld
screen. Although the positioning techniques are well-known and most Congress (MWC), Nreal Light smartglasses provide mixed reality for the
of their problems are already solved, they face the restriction of energy masses [296]
consumption imposed by the gadget size. Therefore, their efforts are Indeed, enabling an enhanced immersion in other realities could
dedicated to the efficient combination of all the available sensors, as not be possible without easy-to-use assistance devices. One of those
demonstrated in [287]. is a haptic VR suit for industrial training. Brian Heater developed
Another area of interest includes Industrial Internet of Things (IIoT) by Valkyrie Industries [297]. Similarly to the Iron Man suite aims to
field. In particular, RealWear company focused on products specifically provide the ‘‘haptics is the sense of touch, particularly relating to the
designed for connected enterprise employees to increase job satisfac- perception and manipulation of objects using the senses of touch and
tion, productivity, and, most importantly, safety [288]. The company proprioception’’. Interestingly, the suite is designed not for gaming but
has been developing an industrial hands-free portable computer for for professional applications, the list of which is kept untold.
many years, aiming to improve the employee’s situational awareness,
providing vital information on demand in the harshest conditions. 3.7.2. Medical and close-to-medical devices
Today’s technology combines a small Android computer and a camera, A massive sector is related to close-to-medial wearable devices,
speaker, and microphone in a tiny screen project information in front i.e., ones that are not certified as medical ones but still aiming to detect
of the user’s eyes. This rugged hard hat is an intelligent headset that some health-related issues. For example, a French piece of technology,
works similarly to the original Google Glass but uses the underlying Crhonolife smart vest, presented at the Consumer Electronics Show
technology in a completely different way. (CES) 2019 event, aims to predict the heart-attack before it hap-
Another project called Kopin Golden-i Infinity is also designed to pens [298]. It is specifically designed to prevent medical emergencies in
improve workers’ life with AR technology [289]. The device is used to patients with congestive heart failure (CHF). Some devices are designed
understand better the repair guidelines for field maintenance, access to monitor the owner’s posture, including Upright and Lumo Lift [299].
to production plans, view utility, and energy information, collaborate Many companies, including POLAR, Garmin, Fitbit, and others [300],
with remote experts, and more. Besides, the device boasts a nine-axis allow for heart rate monitoring but often do not focus on Food and
stabilization head, triaxial accelerometer, gyroscope, and magnetome- Drug Administration (FDA) approval.
ter. This sophisticated tracking system evaluates the effect of the Earth’s On the other hand, some medical wearable devices have also suc-
magnetic field on the user’s geographic position and head movement, ceeded in receiving FDA approval, including the well-known Apple
providing unprecedented tracking accuracy for applications requiring Watch [301] for the detection of falls and irregular heart rhythm,
incredible accuracy. Kopin also develops medical AR glasses for surgery thus, achieving Steve Jobs’ goal. Nonetheless, SYNCHRONY and SYN-
purposes [290]. CHRONY 2 cochlear implants have also received the approval for peo-
Nonetheless, Nsoft and Trimble made a hard hat with Microsoft ple with single-sided deafness (SSD) and those with Asymmetric Hear-
HoloLens XR headset built-in specifically for industrial workers [291]. ing Loss (AHL) [302]. Next, Neofect’s powered glove for people with
The XR10 is a customizable helmet with an integrated rotatable paralysis is shipping this summer [148]. It provides the users with mo-
HoloLens 2 lens that allows construction work to be visualized on site. bility in a paralyzed hand (potentially present due to stroke, Multiple
This first partnership is a clear sign of where Microsoft hopes to go Sclerosis (MS), and Amyotrophic Lateral Sclerosis (ALS). It allows
with its second-generation headset, taking the technology outside of performing simple everyday tasks such as brushing their teeth, opening
the office to real-world sites. doors, or drinking from a cup.
Keeping AR and VR headsets as a standalone solution is a trend for Another active niche of wearable industrial research leads to an ar-
heavy machinery, crate, drivers, and maintenance employees that do tificial nervous system [303]. Researchers from NUS Materials Science
not require constant mobility. However, complex computation on head- and Engineering developed the touch equivalent to human skin with
mounted devices is still challenging due to battery and computation the Asynchronous Coded Electronic Skin (ACES). It can assist future
limitations and overheating issues. Company Zebra produces portable robotics in terms of providing sensual feedback for tactile feedback in
computers aiming to overcome those issues [292] and potentially of- prosthetics.
fload the computations to powerful yet portable wearable cloud nodes. A lot of research activity is currently executed related to Parkinson’s
Complete Zebra solutions are based on Android enterprise devices disease. Based in Austria, EVER Pharma recently received CE Mark
delivering unrivaled scan performance and manageability. Zebra’s fully approval and is releasing its D-mine Pump in Europe [304].
wearable solutions combine mobile with wearable technology and Patents have also shown the trends for medical wearables. In [305],
ease of pairing features providing the operational flexibility needed to the authors introduce a wearable device integrated into a partial or full
exchange complexity for productivity. body soldier’s garment, able to deliver wound sealant to an injury and
Started with Google Glass, many companies aim to develop personal activate selective small balloon inflation to add pressure to a wound.
AR-based assistants. One of those is the well-known Vuzix [293], An alternative use-case considers its usage to stabilize the injured while
which aims to develop next-generation AR smartglasses using Mi- being transported.
croLED Technology with the primary goal of miniaturization of the Honda Motors patented a method to support the vehicle con-
integrated screen. trol [306]. The wearable devices can capture diverse physiological data
Interestingly, Tesla files a patent for Google Glass-like AR system to determine one or more vehicle occupants’ health state. These data,
for faster, accurate vehicle production [294]. According to the patent such as the blood rate pattern, can also be used to identify a vehicle
application, the technology uses computer vision to recognize objects occupant. Their method includes the total or partial control of the
based on the object’s colors displayed on the camera view and the vehicle based on its occupants’ health state.
device’s location. The recognized object is then compared with the
3D model library’s corresponding model and overlaid on top of the 3.7.3. Fashion and comfort
digital data. A massive niche is allocated to the mass consumer by means of
In contrast, Bose AR with audio provided an entirely new di- improving the comfort of everyday activity and fashionable look. The
mension of audio immersion by merging Bose AR technology with number of devices ranges a lot, from luxury smart rings to invisible
sunglasses [295]. The device detects where the user’s location and personal assistants.
which direction the owner is facing via a nine-axis head motion sensor An exciting segment of wearables is novel wearable air conditioning
and the GPS on tethered Android or iOS devices. The audio pumped systems. To name a few, Sony Reon Rocket is a body-cooling wearable

16
A. Ometov et al. Computer Networks 193 (2021) 108074

that promises to chill the owner in the heat [307]. Sony is not the only designed to induce sexual pleasure. Besides practical products, some
company attempting to capitalize on the rising heat. Embr Labs is a design and technological efforts have been made to separate the drive
startup founded by three MIT graduate students who were frustrated means from the stimulation means, enabling a fully autonomous oper-
that it was always freezing in their lab in the summer, even when there ation [322], or, even, to allow remote controlling [323]. Other bodies
are only a few people in the building [308]. and muscle stimulation applications are also explored, for instance, to
tackle incontinence [324].
3.7.4. Consumer sports
Besides the conventional market segment of activity trackers and
3.7.7. Active (Inter-)national projects
smartwatches, few more vendors focus on the development for a spe-
While industrial giants aim to occupy more significant shares of
cific group of users — athletes aiming to improve their results utilizing
better monitoring and improved training [309]. Catapult [310] and the market and increase their revenues, the academic sector is mainly
STATSports [311] are trendy suppliers of wearables to the top sports focused on promising technologies and evaluating existing devices’
teams in soccer, football, hockey, rugby, and baseball. These producers performance.
offer small devices worn in the special vest pockets on the back and A number of the international EU H2020 research projects are
can perform indoor and outdoor localization and track the heart rate. focused on wearable solutions both from the Hardware and the Soft-
Specially developed software performs analysis based on collected data ware domains. Some of the ongoing projects are as following: Wear-
and allows to manage workloads, monitor athlete progress, and avoid able Electroactive Fabrics Integrated in Garments (WEAFING) [325],
situations that could lead to injury. Smart and Flexible Energy Supply Platform for Wearable Electron-
The authors in [312] introduced a wearable device used to record ics (Smart2Go) [326], Personalized Body Sensor Networks with Built-In
fishing data. The wearable can receive motion data from one or more Intelligence for Real-Time Risk Assessment and Coaching of Ageing
motion sensors. Collected motion data is analyzed to determine if it workers, in all types of working and living environments (BIONIC) [327],
corresponds to a casting motion while fishing. In such a case, the cast Real-time Analytics for the Internet of Sports (RAIS) [328]. Moreover,
occurrence is recorded and timestamped. more than 400 projects related to wearables have received support only
Xsens [313] developed a real-time motion capture system named from the EU since 2016.
MVN Analyze, which allows accurate tracking and analysis of the From non-EU perspective, exceptional sense of touch for robots is
athletes’ movements even during high-intensity activity. The system described in research [303,329]. As the device is called, ACES has a
is available in two versions: 17 wireless sensor straps (MVN Awinda) somatosensory perception, which identifies a touch 1,000 times faster
or a Lycra suit (MWN Link) — both comfortable for the user and not than humans’ nervous system. A potential application for this technol-
restricting his movement. ogy will allow performing critical surgery with better responses and
Another promising niche is related to the growing vehicular mar- taking relevant actions in limited time slots.
ket [314,315]. In particular, wearable devices may be used to assist in Another project, described in [330], refers to the approach where
continuous passenger biometrics monitoring during the autonomous ve-
scientists present ultrasoft electronics as a concept for Body Area Net-
hicle operation, thus, improving the overall safety of the system [316].
work (BAN). The researchers in [331] have developed a sample of heart
These include authentication while accessing the vehicle, passenger
cells with a soft sensor and changed the way of embedded medical
presence, driver’s biometric status, sleep monitoring, etc.
devices.
The work described in [332] claims to create Magnetic Resonance
3.7.5. Security
Imaging (MRI) detectors that fit like a glove. The developed solution
A sector with great research potential for wearables is related to
the information security domain. Already today, wearable devices may covers the situations where the hand moves in the magnetic resonance
enable safe transactions/payments on a mobile device based on a scanner without degrading the signal quality, thanks to the MRI de-
wearable security token. It not only separates the security management tectors that seamlessly follow the contours of the hand using separate
from the device performing the transaction but also reduces the risk flexible high impedance elements.
of impersonation or relay attacks [317]. Similarly, other wearable The activity in [333] presents a novel approach to using the NFC
devices have been patented to, for instance, safely store encrypted technology to extract valuable knowledge from sweat analysis. This
information, which is only sent to a pre-specified mobile terminal after study proves to cover a gap as a non-invasive solution working on
verification [318]. Some wearables combine biometrics and advanced advancing the deployed sensors’ form factor. Another study in [334]
encryption to allow specific secured transactions with a counterpart focuses on sweat analysis and claims wearables could analyze the data
communication device/system [319]. to get more insightful information regarding the user’s physical and
However, wearables are not only used for secure monetary or infor- mental state for improving the quality of training.
mation storage/transactions. The authentication possibilities through Overall, it can be stated that financial agencies are interested in
biometrically secured access may also be used to, e.g., automatically investing research funding in wearable technology research activities.
manage other smart devices [320] or send alerts in response to critical Interestingly, those are not anymore limited to eHealth and medical
situations according to the wearer’s vital signs [321]. domains, thus, allowing to cover activities in other fields.

3.7.6. Other industrial activities

3.8. Section summary
Some smaller companies are involved in developing novel niche
solutions and creating new applications. The examples are wearables
for well-being. Relaxation is a part of wellness that corresponds to the This section outlined the classification of wearable devices based
fashion and comfort segment of wearables. The Silentmode’s Power- on, e.g., location on the body, application type, functionality, etc.
Mask is a relaxation mask that keeps records of heart rate variability, It highlighted the communication technologies, architectures, and re-
as stated in [132]. The Dreamlight Zen meditation & sleep mask is a lated paradigms, including technical details, elaborated on various
product that uses lights and music to help users calming down and data processing-related aspects, and involved computing paradigms,
restoring their vitality. outlined localization strategies for indoor and outdoor operation, and
The possibility of wearable devices to sense and act has also found provided insight into the interoperability aspects, and, finally, delved
applications in stimulating sensitive parts of the body and/or measure into the mass consumer market status. Additional information on the
signals. In the adult entertainment market, various wearables were devices is detailed in Appendix.

17
A. Ometov et al. Computer Networks 193 (2021) 108074

4. Open challenges and future perspective method [344] leading to the understanding that it may be a promising
solution for classification tasks based on wearable data. Additionally,
This section outlines the majority of modern problems related to combining many modalities for detecting diseases allows decision sup-
wearable data processing, privacy, security, and transmission aspects. port methodologies for achieving higher quality results in prediction.
Next, localization and communication-related challenges are drawn, The limited number of works leaves a huge research gap, leaving ample
followed by the adoption, hardware constraints, interoperability, and opportunity for future directions. The multimodal approaches were
scalability aspects. used mainly hitherto for the human activity recognition [39], Parkinson
detection [345], difficulty in cardiorespiration [156].
4.1. Data acquisition and processing A new trend of wearable technology for data gathering becomes
possible with the development of smart tattoos constructed from thin-
A large part of the data acquisition challenges is in the first com- film polymers, metals, ceramics, 2D materials, and their integration
ponent of the data processing chain, namely the wearable itself. The with rigid ICT [346]. Part of the smart tattoos will serve for the biomed-
quality, quantity, resolution, and other parameters of the data are ical tasks, i.e., gathering physiological parameters in real-time [347,
defined there. As crowdsourcing gathers data from multiple users, 348], e.g., the genetically programmed cells could be transferred into
multiple devices, all in different scenarios and often using relatively wearable devices.
cheap and energy-efficient sensors, the quality of the data itself can Interestingly, smart tattoos were designed by Nokia as Vibrating
enormously vary across all individual data sources [335]. The quality Tattoos, which are placed on the skin. They contain ferromagnetic ink
and quantity can be related to spatial resolution, temporal resolution, and vibrate when someone is texting or the mobile phone is calling.
or data resolution itself [160]. Spatial resolution determines the phys- Princeton and Tufts University developed the next unique smart tattoo
ical distance between two consecutive measurements from a single to measure bacteria levels in the mouth based on saliva with electronic
source and can be equidistant or strongly varying. Temporal resolution wireless tattoo [346].
describes the time intervals between the two measurements from a Additionally, recent micro- and nanotechnology developments sup-
single user and can be uniformly distributed or strongly randomized. port non-invasive biomedical measurements and wearable in-body and
The data resolution relates to the accuracy of the measurement sensors
on-body sensing, processing, and communication. Examples of existing
equipped in the user wearable, device-specific processing inside that
solutions are insertable cardiac monitors, continuous glucose monitors,
wearable, and outside factors affecting the measurement itself.
Insertable Drug Deliverables Systems (IDDS). Shortly, the development
The subsequent challenge lies in the large-scale data gathering pro-
of the analyte monitors is being expected (for example, for monitoring
cess, i.e., finding the optimal method of gathering the users’ measure-
the oxygen concentration in interstitial fluid), insertable EEG moni-
ments for further utilization [336] and its energy-efficient transmission.
tors (dedicated mainly for epileptic patients), eye health monitors (for
The question of privacy, security, trustworthiness, or user selection is
glaucoma patients). The most promising solutions are controlling the
nowadays still an open challenge [337].
infant, therapeutic drug monitoring, and delivery [124]. Furthermore,
Data processing challenges also describe the difficulties in data uti-
the microneedle sensors’ usage will allow clinicians to monitor the
lization after the acquisition phase. As mentioned above, the data have
electrochemical parameters, for instance, measuring the level of lev-
different spatial and temporal resolutions and varying data quality.
odopa for Parkinson’s patients [349]. However, the main challenges
Converting raw, gathered data into useful data is the main task of data
of the insertables belong, among others, the foreign body response,
pre-processing. The challenge lies in unifying the measured quantities’
which impedes the biosensors’ functioning, likewise the transmission
parameters and removing errors and statistical outliers from the data.
of the data or unexpected migration of the device [124]. One of the
Further data analytics require a clean input dataset and represent a con-
approaches for reliable wearable platforms is to address the optimal
secutive challenge itself. The efficiency of the analytics determines the
choices of sensors and user interfaces and conduct the validation of
amount of useful information extrapolated from the set of individual
methods based on the target application (e.g., clinical, social well-
measurements. Nowadays, wearable-related data processing solutions
utilize machine learning techniques more often [338,339]. being, etc.), data security and confidentiality, decision support, and
One of the main challenges of wearable data processing is the user acceptance.
lack of a sufficient amount of data. To overcome this issue, a few Significantly, the eHealth segment is a strongly regulated sector
categories of the time series data augmentation were proposed [195]. with high accuracy and reliability requirements. Mass-market wear-
One of them is the augmentation in the time–frequency domain with ables’ accuracy cannot compete with often expensive hospital devices,
wavelet transforms as one of the potential successors in augmentation and currently, available consumer wearables are usually not acceptable
of non-stationary time series and non-Gaussian noises [195]. More- for medicine purposes. The reason is that wearable sensors often have
over, the combination of augmentation methods dedicated to imbal- insufficient accuracy compared to laboratory devices. Another issue is
anced datasets from the used architectures weighting is also proposed that devices from different production series can also suddenly measure
in [195]. However, using data augmentation methodology brings po- different values due to different electrical components, which in the
tential problems, such as not reflecting the real data and possible case of medicine can have serious consequences [350].
overfitting. Crowdsensing data collection provides a significant advantage for
Next, the lack of etiquette of the data states an extra problem wearables over large and accurate laboratory sensors. It becomes an ef-
because of the limited number of available datasets. Additionally, fective and low-cost approach for automatic multi-purpose data gather-
obtaining high-quality labels is time-consuming and often requires ing and processing [351]. Nowadays, it is commonly used, for instance,
specialists for this purpose or engagement of the users [340,341]. One to automatically generate radio maps or collect health parameters
solution is to use a semi-supervised methodology when the number of for eHealth monitoring and activity recognition. Crowdsensing over
the annotated labels is small. It could primarily be applied for the data wearables has many unresolved challenges, which include data redun-
gathered by sensors and wearable devices with long records, e.g., the dancy and outliers’ identification and elimination, reducing the data
Human Activity Recognition (HAR) task when not all the labels are processing, transfer, and storage costs, leveraging the temporal and
correctly reported by the users or labeled by the specialists [341]. These spatial correlations of information to reduce the amounts of stored and
same semi-supervised methods minimize the obstacles connected to processed data, and dealing with heterogeneous devices and technolo-
training the data lacking in labels. gies [351]. Wearables can collect a variety of user-centric data, such
HAR was successfully used in temporal assembling of deep as location-related knowledge, information on the user’s well-being or
LSTM [342], LabelForest [343] or bi-view semi-supervised learning health [352], and radio communication-specific data.

18
A. Ometov et al. Computer Networks 193 (2021) 108074

The development of the eHealth/mHealth field by applying wear- significantly fewer samples than required by the Nyquist criteria. Com-
able sensors is also of particular importance due to the rapid growth of pressive sensing has proven to bring several benefits such as efficient
the elderly population in Europe. For example, monitoring the elderly bandwidth and energy consumption, which are the two most vital
suffering from civilization or neurodegenerative diseases could allow network resources in wearable networks [369,370].
to detect of Parkinson’s or Alzheimer’s disease early and predict the
illnesses’ development [353]. 4.3. Security and privacy aspects
Approximate computing is another technique that allows to trade
precision in exchange for increased energy and performance [354]. In the previous subsection, we highlighted that wearable tech-
Many wearable applications, including ML, signal processing, image nology’s main features are sensing and collecting data continuously.
processing, Big Data analysis, and more, may not require very accu- Therefore, with the advancement of the IoWT in next-generation cel-
rate results. Instead, results that are ‘‘good enough’’ can serve this lular networks, security and privacy are vital characteristics for con-
purpose [355]. Thus, approximation has become an effective method sideration in wireless communication scenarios [371]. Regarding the
for improving the performance and energy efficiency of devices with widespread of modern data collection techniques and sensor develop-
limited resources, such as [356] wearables. However, this comes with ment, user concerns about reliability and trust in online platforms and
many issues, such as determining the minimum required accuracy, location-based services are continually growing. The authors of [372]
defining approximate problems, and monitoring the results of the appli- state that most modern devices include built-in sensors, which provide
cation [357]. Hence, efficiently implemented approximation methods accurate location data and information about physical activity level and
can achieve optimal performance gains in latency, execution speed/- mental health. The study shows that the data is easily accessible and
time, and battery consumption. does not correlate with users’ awareness. There is no unified solution to
cover all threats in wearable technology security. Thus, further research
4.2. Data transmission aspects and development are required.

As we move towards a more connected ecosystem, data generation 4.3.1. Data-related aspects
will continue to increase, primarily due to 5G and beyond technologies In [392], the authors show concern about the actual owner of the
that ensure faster connectivity. However, not only technologies are data collected via smart connected devices. Due to the proximity to
developing, but also the problems that surround them. a human body, hereinafter, we claim wearable technology is one of
Although a centralized cloud has traditionally been used for data the most exposed wireless solutions to privacy threats. Privacy issues
management, processing, and storage, it has two major problems: (i) are relevant to data integrity, and an attacker might physically affect
the latency to process data may be critical; (ii) all this data creates the owner via malicious hardware or software. As long as embedded
a significant load on the overall network performance. In contrast, sensors in smart wearable devices crowdsource personal information
Edge computing offers a solution to latency by moving critical data about technology users, data processing should be privacy policy com-
processing to the network’s Edge [358]. Edge devices can collect and pliant. There are specific laws in countries worldwide that protect
process data in real-time, allowing them to respond faster and more the data privacy of their citizens. Currently, the primary document in
efficiently. However, there are issues related to the devices’ capabilities, charge of data privacy and security protection in Europe is the General
including developing software and hardware to handle the cloud’s Data Protection Regulation (GDPR) law. This document implements
computing load. newer focus areas, such as privacy rights, data security, data control,
Different computing paradigms demonstrate great opportunities and and governance. Accordingly, all data collected via wearables can be
introduce new techniques to improve system performance. In practice, considered sensitive (e.g., genetic, biometric, and other health-related
computing paradigms are sometimes discussed as mutually exclusive data). Consequently, our survey provides a taxonomy for multiple
approaches to network infrastructure. Although they may function types of data that wearable devices are gathering via built-in sensors
differently, utilizing one of them does not preclude using the others. in Table 5. As this survey investigates, the most popular solution to
In this context, some approaches have been proposed to solve this exchange data on wearables is BLE, and many research experiments
issue (e.g., a cognitive IoT Gateways [359], a platform that automat- report on its vulnerability to security attacks.
ically determines the best environment to execute a task [360], and Evidently, in most of the security and privacy preservation scenar-
a framework for scheduling applications over a hybrid public–private ios, it is necessary to perform complex cryptographic operations, which
cloud [361]). However, the automatic switching between computing are limited in wearables due to their low computing power and limited
paradigms is a challenge for future research, in any case. resources.
It brings the researchers to the problem related to the migration
of services or a Digital Twin (DT) of the wearable device, for in- 4.3.2. Impact on energy consumption
stance, from one Edge server to another [362,363]. When a user moves The execution of the related complex operations implies, in most
across different geographical locations [364], e.g., from home to the cases, extra consumption of energy and resources, which considerably
workplace, the service/DT may follow the user’s device. However, reduces the duration of the battery. Despite the mitigation strategies
the determination of an optimal migration decision is challenging already discussed, these drawbacks mean that it would be an excellent
because of the trade-off between the migration and data transmission starting point to reduce the potential risk of future attacks. Their real
costs [365]. On the one side, migrating a service/DT may incur net- implementation on wearable devices becomes quite challenging. Today,
work overhead or even interrupt the connection. On the other side, developers and researchers explore wearable technology’s security as-
not relocating a service/DT may increase the data transmission delay pects, especially in the medical and industrial segments. The authors
between the user and the edge server. In the literature, there are several of [393] investigated the executability of various primitives used in
approaches to reduce the transmission delay in traditional wireless symmetric and asymmetric cryptography, block ciphers, elliptic curve
communication networks concerning the migration procedure [366– cryptography, and conventional hash functions. The set of measure-
368]. However, ensuring seamless migration among Edge servers is still ments showed that the use of widely used methods leads to a significant
an open problem that needs to be investigated. computational load on wearable devices compared to, for example,
Next, compressive sensing is gaining increasing popularity from the smartphones. As one of the solutions, this work highlights the need to
lower layers perspective, especially for IoT applications that involve develop specific lightweight primitives, taking into account the trade-
sparse data signals. It is a signal acquisition and reconstruction tech- off between energy consumption and safety as one factor in creating
nique that enables the receivers to reconstruct the actual signal using efficient devices with limited resources.

19
A. Ometov et al. Computer Networks 193 (2021) 108074

Table 5
Taxonomy of sensors and data collected via wearable devices.
Sensor Biometrics Location Audio/Video Altitude Inclination Health data Other References
Gyroscope ✓ ✓ [373]
Accelerometer ✓ ✓ [373,374]
Altimeter ✓ ✓ [375]
Proximity sensor ✓ [376,377]
Temperature sensor ✓ [378]
Magnetometer ✓ ✓ [379]
IMU ✓ ✓ [380]
Camera ✓ ✓ ✓ [167,381]
Optical sensors ✓ ✓ [382,383]
Microphone ✓ [381]
Beacons ✓ [384]
Chemical sensors ✓ [385]
Antenna module ✓ [386,387]
Ambient light sensor ✓ ✓ [388]
NFC sensor ✓ ✓ [389]
Touch sensor ✓ [390,391]

4.3.3. Privacy aspects being present due to the complexity and continually changing of indoor
Nonetheless, from Radio Access Network (RAN) perspective, several environment [396]. Factors alter the signal propagation as propaga-
entities can be considered to have access to the most sensitive data col- tion losses (e.g., reflection, diffraction, scattering), Non-Line-of-Sight
lected by wearable devices. The actual device might be an adversary to (NLOS), multipath propagation [396,397].
data integrity and security if not enough safety measures are considered NLOS propagation radio signal is the primary source of the inaccu-
or no encryption has been used. rate position, so mitigating its effects is one of the main challenges in
Since most of the wearable communications utilize D2D solutions, positioning [398–400]. A vast amount of methods for mitigating the
i.e., when multiple devices are communicating in proximity (e.g., exe-
effects of NLOS on position accuracy has been proposed. For example,
cute any collaborative tasks), securing those links became an essential
the authors of [398] classify them into two categories, methods that
field of research [394]. D2D nodes in the proximity can share sensitive
mitigate the NLOS into the positioning algorithm without previously
information about the user identity and other personal data. As a
consequence, this personal information can be exploited for illegal identifying NLOS and methods that implement a NLOS identification
purposes by eavesdroppers [206]. Network security protocols in D2D phase first and then, in a second phase, mitigate the NLOS error from
networks were developed to allow users to avoid any information leak- the position estimation. In [401], the authors consider a cooperative po-
ages. The major issue primary in secure D2D networks is the dynamic sitioning approach to address the NLOS estimation error issue. Despite
readjustment required due to user mobility. Therefore, the need for an the great variety of methods proposed in the literature, the effects of
adaptive mechanism for D2D network security emerges when the users NLOS have not been completely mitigated. Crowded environments and
join or leave the coalition [395]. complex geometries scenarios still represent a challenge to overcome.
To conclude, although there has been a plethora of work related to Finding a good trade-off between the accuracy and the energy
wearable architectures for efficient dynamic D2D communications, the consumption of a localization system is an on-going challenge. In
following topics are mostly missing in the literature: multi-connectivity particular, energy efficiency is a critical constraint of wearables due
heterogeneous Radio Access Technology (RAT)s and mobility in the to their small form factor and portability requirements, which prevent
IoWT, optimized link selection for highly dynamic multi-tenant net- them from using large batteries. While outdoors, the de facto local-
works, need for efficient privacy-preserving techniques, security issues, ization system for many wearables is GNSS, indoors there are many
the enhancements offered by a set of innovative 5G technologies in
competing technologies with varying degrees of localization accuracy
practical IoWT contexts, among others. There are still no appropriate
and energy efficiency. Widely-available technologies like Wi-Fi and
communication architectures to converge wearable devices and net-
Bluetooth often rely on fluctuating Received Signal Strength (RSS)
works with external Internet via the edge-computing infrastructures al-
measurements or imprecise Time-of-Arrival (TOA) estimates, which
lowing seamless and well-orchestrated integration and interoperability.
Therefore, multiple design choices and criteria need to be thoughtfully decreases their localization accuracy. UWB-based localization systems
studied in this field. For instance, the Inside-the-body (ITB) component can achieve cm- or dm-level accuracy and have a low energy consump-
of the network could be connected to the user’s equipment (e.g., a tion. However, they are not readily available in most of the wearables
smartphone or high-end wearable device). These require real-time already on the market. Future efforts are required to develop accurate
access to the proximate edge cloud ecosystem, which offloads the localization techniques that can be implemented with low costs and low
battery-constrained devices and processes their demanding computing overhead on small devices.
tasks remotely. This construction became even more challenging when Location privacy is a major concern of wearable localization, as
a user is on the move, and numerous technical challenges need to be location information in the hands of non-trustable parties can lead
resolved in high dynamic scenarios. to a disclosure of unwanted private information, identity theft, or
decreased safety. Communication privacy and authenticity validation
4.4. Localization aspects in wearables are also essential, as privacy suggests that an individual
keeps personal data as much as possible under control and to share
Modern wearable technology heavily relies on the precision of
them only with ‘‘trusted’’ devices [402].
positioning techniques embedded in the devices. This subsection lists
Building large-data collection platforms constitutes a Software (SW)
the main challenges related to the localization of resource-constrained
wearable devices. engineering challenge, which requires innovation in data analysis and
To start with, the geometry of indoor scenarios plays an essential visualization techniques to gain insights into individual and aggregated
role in the robustness, precision, and accuracy of the positioning sys- user patterns, i.e., Spatio-temporal clustering, visual analytics, etc.
tems used on wearable devices. Mainly because the dominant position- Standardization is the most powerful framework to provide fast,
ing technology used by wearable devices relies on wireless technolo- secure, and straightforward integration between different devices, com-
gies, e.g., BLE, Wi-Fi, UWB, with various signal propagation problems ponents, and systems. Various standards have not been fully adopted

20
A. Ometov et al. Computer Networks 193 (2021) 108074

in existing localization systems, specifically in indoor localization plat- and reputation assessment require intelligent management. To address
forms. For instance, the authors in [403] evaluated indoor localiza- these challenges, the development of novel Edge/Fog architectures for
tion systems and provided general comments on the International D2D to improve wearables’ energy efficiency and support massive,
Organization for Standardization (ISO)/International Electrotechnical heterogeneous, and multi-connectivity devices is required.
Commission (IEC) 18305 standard. The authors emphasized the lack
of standard evaluation procedures for the customized ILS and Indoor
4.5.1. In-band/out-band D2D communication
Positioning System (IPS), impeding to be embraced in other environ-
Regarding spectrum usage, D2D communication is primarily classi-
ments. However, standardization is not only limited to the evaluation
fied into in-band and out-band. In in-band D2D communication, both
procedures. It is also applicable to positioning technologies, indoor
cellular and D2D communications utilize the same spectrum licensed to
maps, software communication protocols, devices, discovery protocols,
among many others. the network operator, whereas out-band D2D communication operates
The ISO/IEC 18305 [404] is an ISO standard devoted to identify and in an unlicensed frequency band. As pairs of devices are supposed to
define the most appropriate metrics and scenarios to assess localization be in close proximity to establish D2D communication, less transmit
and tracking systems. Indeed, many researchers have reported difficul- power is required than while using longer-range transmissions, which
ties in comparing the methods with the published results because the can prolong the devices’ battery life [406]. Besides, D2D communi-
evaluation metrics reported differing in the literature (e.g., averaged cation can also result in low-transmission delay and high throughput
positioning error, mean squared error, median error, among others). compared to Base Station (BS)-assisted communication. On the one
The experimental part and data collection in the literature present hand, the use of out-band D2D eliminates the interference between
significant differences with various evaluation area sizes and strategies D2D communication and network users, while at the same time, devices
to evaluate the method, procedures, and protocols for data collection. have to maintain two active wireless interfaces, which leads to higher
Thus, the ISO/IEC 18305 standards can be considered as a guide energy consumption. On the other hand, in the case of in-band D2D,
on how evaluation results should be reported, including aspects on where devices need only one active interface, the interference poses the
the evaluation area, with special interest indoors. Additionally, some main issue. To handle the interference, efficient and easy-to-implement
initiatives standardize the evaluation of localization systems through methods are required, which pose a challenging task to the academy.
extensive sets of pre-collected data (datasets), providing an evalua-
tion framework that can be reproduced. Localization competitions are
4.5.2. Packet aggregation
also becoming popular in the evaluation as they allow the compar-
Wearable devices are generally physically close, thus, they perform
ison of real-time localization solutions under the same conditions in
local end-to-end interactions with traffic that follows similar patterns.
challenging areas.
In such a scenario, an aggregator node can collect data packets from
Although some of these standards are available in the literature,
neighboring devices and then forward them to the final destination em-
most application platforms do not adapt to these standards. It results
ploying D2D technology. The advantages of adopting this forwarding
in non-standards localization platforms that are difficult to integrate
with other systems, which is the major challenge for developing indoor approach can be to lessen energy consumption and improve communi-
localization technology nowadays. As stated in [405], we need a com- cation performance thanks to a top-notch selection of the aggregator
mon taxonomy with the related services and protocols, as it is the most node and a proper setting of its transmission. As shown in [407],
relevant missing piece in indoor localization now. The next steps should using a more robust Modulation and Coding Scheme (MCS) in the
focus on defining a taxonomy of ILS/IPS, outlining the framework to uplink can improve the energy-efficiency of communications with small
allow the collaboration of multiple ILS/IPS, and drawing the required data transmissions by reducing data rate and lowering the transmission
standard services and applications in the near future. power. In [408], the authors propose a strategy where a single device
aggregates locally generated machine-type data packets, supplements
4.5. Communication and architectural aspects them with its data, and then transmits them to the BS in order to
alleviate the impact of the massive number of Machine-to-Machine
Smart wearables need to be interconnected in a heterogeneous (M2M) devices on a cellular system.
manner (i.e., different technologies and devices communicating with
each other), battery-operated. They may harvest energy, e.g., from the 4.5.3. Higher cellular system capacity
sun or human motion. They enable various smart functions, which co-
A dense communication environment poses a challenge from the
operate in a decentralized manner and offer Internet access. Designing
connection establishment point of view as interference from the devices
efficient D2D-based capabilities is one of the vital unsolved challenges
of neighboring users can make establishing connection difficult [230].
in wearable communications due to the need for heterogeneous, de-
The level flow dynamics of connectivity in wearable devices also add
centralized, low-cost, and low-power architectures. Despite advances in
further complexity to this problem. Since interference footprint from a
semiconductor technology and energy-efficient system design, overall
device can be managed by controlling the transmit power, power man-
energy consumption by communications systems is still proliferating.
agement [409] approaches can be utilized to minimize the interference
As users’ expectations for the networking devices’ features and battery
and potentially provide more efficient and reliable connectivity.
life continue to grow, it is imperative to optimize performance and
power consumption further. Therefore, the collection, transmission, Generally, D2D communications allow extending the system ca-
computation, and storage of data must be addressed consistently in pacity thanks to data offloading and reuse gain. Freeing up licensed
terms of power and/or time. bandwidth through unlicensed D2D communications improves the scal-
In dynamic Large Scale Environments (LSEs), where wearables ability of the system, thus allowing better management of the resources
move around while exchanging data, issues related to scalability and available for the wearable environment [410]. Furthermore, due to
interoperability emerge to be carefully addressed. These LSEs rely advances in research on methods for the interference management
on Edge/Fog computing architectures to implement the intelligence between local D2D communications and the BS, resource allocation
needed by the proximate devices and improve the quality of service in the cell and model selection, it is possible to implement frequency
they offer. Edge/Fog architectures require adequate models with a reuse techniques that spur an increase in the network capacity. Finally,
high degree of device autonomy and management. Since such sys- the joint utilization of D2D communications and MEC technologies im-
tems typically span over a wide area and include a large number proves cellular networks’ computation capacity by the tasks offloading
of interacting devices, the aspects of service and object discovery, to a nearby D2D device and an Edge node [411].

21
A. Ometov et al. Computer Networks 193 (2021) 108074

4.5.4. Group communications dynamically and in a ‘‘plug-and-play’’ manner to hide differences in IoT
Group communications consist of one-to-many transmissions suit- technology, configuration, and data format.
able for scenarios in which the same data must be sent to multiple de- Another issue related to human involvement is the biological safety
vices. In wearable technology, performing group communications (also under radio-frequency radiation. Health concerns attract exploding
known as multicast) could help tackle energy savings, scalability, and interests from the research community due to extreme proximity and
network traffic reduction. Furthermore, to improve the QoS offered to direct physical contact with the human skin. Moreover, wearable de-
all gadgets, multicast D2D communications can be established towards vices are recognized to cause a higher level of Specific Absorption Rate
devices experiencing adverse channel conditions [412]. In this case, the (SAR) [416]. The SAR value is determined when the device is operating
forwarding mechanism requires that a device, chosen as a relay node, at maximum power. The human body absorbs electromagnetic radi-
after receiving the data directly from the BS, forwards it to a cluster ation, which causes thermal or non-thermal radiation in the affected
of devices set up using a proper approach. Researchers have proposed tissues. In this regard, it is vital to direct the research towards regula-
many cluster-based multicast transmission methods for D2D communi- tion aspects based on new device types, materials, operated frequencies,
cations [413]. Moreover, in the dynamic multi-hop multicast schemes, and transmit power [214]. For instance, in [417], the exposure amounts
the optimal number of relays must be appropriately selected by taking in different spectrum bands are investigated. Furthermore, industry
into account a trade-off between multicast gain and multi-channel bodies should be introduced with the latest regulations to address users’
diversity. concerns better and promote this new technology.

4.5.5. Computational offloading 4.7. Hardware constraints

In scenarios with low connectivity opportunities, mobile devices
may use local resources on the other devices in close proximity to Several advanced functionalities are being added to wearable de-
compute a common task when offloading to remote clouds fails [358]. vices to enable new services and target new use cases. However, they
Using D2D communication, users can improve mobile cloud comput- are still to be executed on tiny and resource-constrained devices. As
ing performance in terms of computational acceleration. A significant a result, more features may result in increased energy consumption,
problem in studying D2D connectivity as an alternative to Mobile Cloud which often compromises the quality of the final wearable applica-
Computing (MCC) is how users can discover and use other mobile tions. Hence, energy consumption is considered as one of the most
devices’ computing resources. Since D2D connection is usually intermit- critical challenges in wearable computing [2]. The power alimentation
tent due to user mobility, access schemes need to be carefully designed. techniques of wearable devices improve with the evolution and the
The users can make the most of nearby mobile devices’ computing increase of the user demands. For instance, the difficulty in plugging
resources without spending too much power on device discovery. wearable devices into power sources due to their unavailability led
to wireless charging techniques. The first commercial devices with
4.6. User adoption aspects integrated wireless charging were cell phones. However, device manu-
facturers continued deploying these techniques in other devices, such as
The issue of user involvement is especially acute in two areas: med- laptop computers, electric vehicles, and wearables. Although it acts as
ical and industrial. In all other situations, the use of wearable devices is an alternate solution to regular wired charging and helps remove the
a person’s choice. It is often a matter of necessity in the case of medicine charging port from the physical design, the disadvantage of wireless
and industry. There is a severe risk that pervasive devices, due to charging is that it needs long charging times [418].
their complexity and excessive ‘‘intrusiveness’’, ingenerate unnecessary When selecting a wearable device, users consider the energy con-
discomfort feelings and psychological stress in users. Some patients sumption metric as one of the top selection criteria. This fact motivated
humble themselves in the case of medical examinations since they see the device manufacturers and the research community to explore tech-
a direct connection between the need for monitoring and recovery (or niques of extending the battery life of wearable devices, i.e., increasing
maintaining their condition). Moreover, workers usually do not fully energy efficiency. The research community identified three main ways
understand the purpose of monitoring in the case of industries [414], for achieving energy efficiency, namely minimizing power consump-
and attempts to implement wearable devices in industries often meet tion, advancing the battery techniques, and energy harvesting. First,
with strong resistance from the intended users. the energy consumption of a sensing process in several sensor-based
Nonetheless, the information ownership issue generates user distrust applications may be comparable to, or even greater than, the energy
of wearable technologies. The privacy reasons suggest an individual to consumption of the radio communication task [419]. Therefore, se-
keep personal data as much as possible under control and to share them lecting low-power sensors and using compressed sensing techniques
only with ‘‘trusted’’ devices. The lack of attention to privacy control can help minimize the sensing energy consumption, thus, extending
would imply, for the user, a constant intolerable fear that his sphere is the wearable devices’ battery lifetime. Second, the choice of com-
violated and could ultimately lead to a state of depression. In the case of pact and long-lasting batteries is a significant design factor in wear-
industries, workers are often afraid that the information collected from ables. Li-ion batteries are becoming the most common type in many
wearable devices could be used against them as grounds for dismissal wearable devices, thanks to their flexible and efficient energy-storage
or transferred to third parties [415]. features [420].
Another point directly related to the degree of user involvement is While the aforementioned energy efficiency measures can extend
their technical skills. Young people are quick to learn new technologies, wearable devices’ battery life, they cannot eliminate the need to
while older people, who most often need medical attention, are usually recharge and replace the batteries. Consequently, new energy harvest-
not so good. As a result, they may prefer more traditional technologies ing techniques have been explored to enable the use of self-powered
and resist new ones, such as wearable devices. wearable devices. Achieving autonomy by harvesting energy from the
An essential step in overcoming the resistance of users of new tech- environment is especially attractive for wearable systems [421]. The
nologies is creating detailed and, at the same time, simple instructions environment’s collected energy can be in the form of motion, temper-
that would explain how the device works, what data it collects, and ature gradients, light, electromagnetic radiation, etc. These methods
how this data is protected. include microkinetic energy harvesting systems that use frequencies
Finally, users want to take care of themselves. Sometimes, this generated by a human movement to harvest energy [422], powering
means quickly adding sensing, measurement, computation, and com- wearable devices by harvesting solar energy [423], self-powered smart
munication devices to the platforms already in use. It should be done tissue [424], and wireless power transmission for implants [425,426].

22
A. Ometov et al. Computer Networks 193 (2021) 108074

However, the development of energy harvesting systems for wear- diffused in Consumer Electronic devices [435]. The work [434] also
able devices has several problems compared to other traditional de- proposes to transport ISO/IEEE 11073-2060 messages over the
vices. These problems are mainly related to wearable devices’ internal CoAP [445] with Representational state transfer (REST) API, which has
limitations regarding physical size, weight, and body movement. As been designed by the Internet Engineering Task Force (IETF) Constrained
a rule, modern energy harvesting efficiency is not high enough to RESTful Environments (CORE) Working Group [446] also approach to
autonomously power wearable devices. Also, the availability of envi- constrained nodes.
ronmental energy is not always guaranteed. Finally, energy harvest- CoAP is considered the reference Web transfer protocol to en-
ing requires integrating several additional hardware components on able the inclusion of the simplest object in the Web. The transport
board, such as environmental energy collectors, additional batteries of ISO/IEEE 11073-20601 over CoAP is not straightforward. While
to store the collected energy, etc., which becomes a challenge when communications in CoAP are initiated from a client towards a server,
considering a small form factor. Therefore, a significant amount of both a manager or a client can initiate a communication in ISO/IEEE
research is required in this direction so that future wearable devices 11073-20601. This discrepancy requires the deployment of a CoAP
can continuously generate energy from external sources [427]. client and a CoAP server on the manager and agents, thus increasing
Self-powered devices are not considered the last step towards the computational load of devices.
energy-efficient operations of wearables. The circuit technology ad- Similarly to [435], the work [447] proposes to transport ISO/IEEE
vances aim to reduce the device power consumption up to keeping 11073-20601 messages over the UPnP protocol, a technology intro-
them perpetually alive. As a result, a future breakthrough for achieving duced by the Digital Living Network Alliance (DLNA). This open net-
always-on things in IoT is the concept of zero-energy devices [428]. work architecture enables discovery and control of networked devices
This concept is based on zero-power sensing, where all the energy and services and is widely deployed in modern Consumer Electronic
needed for the measurement data is generated by the specific phys- (CE) devices, such as media servers and smart TV. Besides basic com-
ical phenomenon being measured. Therefore, no other power supply munication features, UPnP also provides management features as a de-
sources need to be present in the system. The generation of the energy vice discovery mechanism that allows applications to opportunistically
from the measured phenomenon is performed using two fundamental use the nodes near a patient with no prior knowledge.
enablers for zero-energy devices, namely Wireless Energy Transfer In [434,435], the Application Hosting Device (AHD) was chosen as
(WET) and backscatter technologies. the conjunction point between the Continua architecture and IoT. In
other words, this is the device where ISO/IEEE 11073-20601 is mapped
onto CoAP/UPnP AHD plays this role because it is the device closest to
4.8. Interoperability aspects the patient and powerful enough to run multiple protocols and perform
complex operations.
Other proposals in the literature retrofit biometric sensors with
While academia, industry, and standardization bodies have shown
CoAP, but are either non-compliant with the continua guidelines or
a substantial interest in developing solutions for the IoWT ecosystem,
with the ISO/IEEE 11073 standard set [448]. It cannot be considered a
there are still multiple aspects requiring additional attention, including
real integration with the legacy eHealth system, which will continue
scalability, mobility, dynamicity, heterogeneity and interoperability,
to represent and exchange information according to their standards.
adaptability, security, standards, etc. Among these challenges, interop-
Therefore, most of the benefits coming from the integration are lost.
erability is one of the main issues since various components, objects,
communication technologies, applications, services, etc., need to seam- Abstracting from the integration of the IoWT and eHealth systems,
the work [279] presents the open challenges of IoT interoperability
lessly cooperate and interact with each other to realize the full potential
needed to be solved. One of the open challenges is ensuring the in-
of the automated IoWT systems [429].
teroperability of IoWT applications and solutions from all perspectives
Despite the impressive potentialities of such systems, the full inte-
such as devices, networks, and others (for more details, see Section 3).
gration between IoWT and, e.g., eHealth systems are still far from an
In contrast, the current solutions face interoperability from a specific
actual implementation. Two main factors hinder such an integration.
perspective only. This finding highlights the direction for future IoWT
The first issue is focused on the interoperability between eHealth
adoption. Another area of the research work identified in [279] is the
devices and wearables. Reference architectures and protocols devised
implementation of scalable interoperability solutions able to ensure
so far for eHealth and IoWT exhibit many differences. The second one
reliable D2D connectivity [449] between devices with different com-
involves managing the potentially numerous and heterogeneous node
putational capabilities (e.g., high-end and low-end devices). Moreover,
candidates integrated into the eHealth platforms.
with the enormous current interest in digitalization, the transition
Moreover, personalization and interoperability represent two corner- toward distributed cloud computing is an inherent part of future 5G and
stones of eHealth systems [176]. Each patient has their peculiarities: beyond platforms. It is crucial to develop interoperability solutions for
one can be computer literate or not, mostly healthy or seriously ill, different platforms and applications with a high level of flexibility in
more prone or reluctant to accept a given treatment (e.g., due to cul- this context. Finally, it is vital to address interoperability between more
tural biases). On the other hand, pathologies have their characteristics, than a couple of platforms. Solutions must be realistic and scalable
and each one requires specific treatment and monitoring. As a result, across multiple platforms regardless of the area and domain, scenarios,
it is impossible to devise a single one-fits-all eHealth platform; instead, and even utilized technologies.
patients need to rely on various personalized solutions. The only way To sum it up, seamless and ubiquitous communications are expected
to economically achieve the required high degree of personalization to accelerate technological innovation, increase productivity and relia-
is to construct eHealth systems based on basic devices with a high bility, and open up new opportunities for IoWT technologies. Therefore,
technological interoperability level. end-to-end solutions for wearable things that offer seamless integration
Most of the attempts to integrate IoWT and eHealth systems carried into existing systems and processes are of great concern. Moreover,
out so far aim at achieving full interoperability between eHealth de- it is essential to note that weight, size, power, durability, reliability,
vices and wearables. The main idea is to transport part of the ISO/IEEE and ease of use are key considerations in finding solutions for the
11073-20601 over protocols typically used in the IoT [434] or widely end-users [450].

23
A. Ometov et al. Computer Networks 193 (2021) 108074

Table 6
Summary of the main challenges related to modern wearable technology (sorted alphabetically).
Challenge Groups Refs. Observed existing approach
Appropriate service/DT placement and A, DP [430] Utilization of an algorithm for replica placement
migration [366] An algorithm for the optimal migration strategy of services based on dynamic programming
[367] A model for resource management based on the regularization technique
[368] A framework to design optimal service migration policies based on Markov decision process
Data collection issues DP, SW [336,337] Development and application of a unified, trusted crowdsourcing platforms
Datasets’ imbalance DP [431] The application of data augmentation methods for imbalanced datasets and weighting architecture
Energy overheads related to sensing DP, HW, SW [419] Following of the sensing energy consumption minimization strategy
A feeling of constant surveillance UR [414,432] A detailed explanation of the purpose of the monitoring and what is going to be done with the
collected data
[433] Providing the user with privacy and security of collected data
Inefficient computing resources A, N [358] Carefully developed access shames allowing the users to utilize nearby mobile devices’ computing
detection resources as much as possible while not wasting too much energy on other devices’ discovery
Inefficient data analytics DP, SW [338,339] Application of techniques from the field of statistics and ML
Inefficient switching among resources A, N, HW [359–361] Selection and utilization of improved task offloading schemes
in hybrid/heterogeneous networks
Insufficient computing capabilities A, N, HW [358] The use of D2D communication to improve mobile cloud computing’s performance
Lack of appropriate data labeling DP [340,341] Application of semi-supervised methods
Lack of direct interoperability between A, DP [434] A system architecture based on IEEE 11073, Constraint Application Protocol (CoAP), and Universal
eHealth devices Plug and Play (UPnP)
[435] A reference implementation based on Transmission Control Protocol/Internet Protocol (TCP/IP),
CoAP
Lack of modern energy harvesting HW [422] The use of micro-kinetic energy harvesting systems
opportunities [423] Powering wearables with solar energy harvesting
[424]] The application of self-powering smart fabric
[425,426] The integration of wireless power transfer option for implantables
[427] Continuous power generation from ambient sources
Lack of network resources A, N [410] The application of D2D to extend the system capacity thanks to data offloading and reuse gain
[411] The use of D2D-MEC system to improve the computation capacity of the whole system
Problems related to broadcast A, N [412] Group communications (multicast transmission)
communications
Limited amount of the available data DP [195] Extension of the available data by applying the data augmentation techniques
Low data quality DP [160] The application of post-processing techniques to improve the quality
Low data resolution DP [335] The utilization of active and passive data collection approaches to increase the resolution
Low QoS indicator A, N [413] Cluster-based multicast transmission methods for D2D communications
Low technical skills UR [415] A detailed guidance and training including seminars and interactive lessons for every age group
[436,437] Friendly and responsive customer support
Non-optimized security and privacy HW, DP, SW [438] Content agnostic privacy and encryption protocol eliminating the need for asymmetric encryption
enablers
[393,439] Integration of lightweight cryptography solutions including more appropriate elliptic curve types
or algorithm implementations
[440] More efficient utilization of manufacturer-provide System on Chip (SOC)s accelerated for
cryptographic primitives execution
[441] Finding trade-offs between the primitive and required level of the provided security
Problems of classification, anomaly DP, SW [442–444] The application of ML approaches
detection, forecasting problems

A – Architecture; N – Networking; DP – Data Processing; HW – Hardware specific; SW – Software specific; UR – User-related.

4.9. Management/scalability aspects The scientific community is only scarcely working on devising
proper ways to manage billions of devices, which will populate the
The IoWT paradigm is rooted in the simple consideration that the forthcoming IoWT [452]. The issue is still open, and very few proposals
space around us is saturated by a multitude of devices connected to are focused explicitly on integrating the IoWT in today’s networks.
the Internet and unequivocally addressable [451]. The devices usually
have limited capabilities when considered singularly, but they become 4.10. Section summary
capable of very complex operations when a few cooperate, to say it in
a motto: Their number is their strength. This section outlined the main challenges present in modern and
However, the efficient management of the considerable IoWT de- future wearable devices and underlying technologies. The summary of
vices’ number is not straightforward to achieve, and maybe the most the most significant ones is presented in Table 6.
prominent factor which hinders the full accomplishment of the IoWT These challenges are still unresolved because they are extremely
paradigm [452]. Wearables are usually utilized by single users/com- complex and limited by various factors locked in small form-factor
panies with high mobility patterns. Therefore, it becomes close to devices. Current research efforts are fragmentary and very narrow in
impossible to define the number of available nodes in a given area, their scope. They focus either on the machine-centric perspective or on the
capabilities, and the way of their integration to solve a given problem. human-centric perspective, which results in a lack of synergy. Solving

24
A. Ometov et al. Computer Networks 193 (2021) 108074

the challenges described above requires a unique, multidisciplinary, Acknowledgments

and inter-sectoral approach and a concerted effort of the best experts
from all over the world to make progress beyond the state-of-the-art. This work was supported by European Union’s Horizon 2020 Re-
search and Innovation programme under the Marie Skłodowska Curie
5. Review summary grant agreement No. 813278 (A-WEAR: A network for dynamic WEar-
able Applications with pRivacy constraints), http://www.a-wear.eu/
The evolution of modern electronics towards miniaturization paves
List of Acronyms
the way for a relatively young segment of IoT devices – wearables,
the ones we carry and wear on us daily. Indeed, the main building
3G The Third Generation Cellular Network Technology
block for mass adoption and broad integration of modern wearables
3GPP The 3rd Generation Partnership Project
is technology, including computation, communication, battery, chip
size aspects, among many others. The history of wearables, provided 5G The Fifth Generation Cellular Network Technology
in detail in this paper, is enormous and dates hundreds of years ago, A-WEAR A Network for Dynamic Wearable Applications with Privacy
leading to the devices currently surrounding people and soon joining Constraints
our ecosystems. AAFT Adjusted Fourier Transform
As for modern devices, evolving from the first healthcare devices ACES Asynchronous Coded Electronic Skin
and conventional activity trackers, different wearables are presently AHD Application Hosting Device
found on various parts of our body, mainly depending on the appli- AHL Asymmetric Hearing Loss
cation scenario and data collection/output needs. Wearables of today AOA Angle of Arrival
preliminary communicate via short-range wireless technology, with AI Artificial Intelligence
few exceptions based on infrastructure connectivity. This tendency is ALS Amyotrophic Lateral Sclerosis
mainly due to present battery limitations and the overheads brought
API Application Programming Interface
by higher power consumption while using longer-range communication
AR Augmented Reality
technology. Nonetheless, most wearables are still utilized for data col-
lection purposes requiring sophisticated techniques to achieve higher ATM Automated Teller Machine
efficiency of the entire data processing life cycle. However, the tight BAN Body Area Network
coupling of various systems provided by different vendors is still one BLE Bluetooth Low Energy
of the most significant challenges of wearables due to the lack of good- BS Base Station
practices on interoperability and the appropriate standardization in the BLSTM Bidirectional Long Short-Term Memory
young IoWT niche. CAGR Compound Annual Growth Rate
To summarize, wearable technology is an essential building block in CE Consumer Electronic
the future ICT systems. It is still in its infancy, and several critical chal- CES Consumer Electronics Show
lenges from data acquiring and processing, communications, security, CSI Channel State Information
privacy aspects, hardware limitations, and user adoption are still to be CHF congestive heart failure
addressed. This paper highlights those and provides the readers with CoAP Constraint Application Protocol
an excessive summary of potential solutions to overcome the present
CORE Constrained RESTful Environments
literature.
CNN Convolutional Neural Network
CH Cluster Head
CRediT authorship contribution statement
D2D Device-to-Device
DLNA Digital Living Network Alliance
Aleksandr Ometov: Project administration, Data curation, Visual-
DHKE Diffie-Hellman Key Exchange
ization, Writing – original draft. Viktoriia Shubina: Writing – original
DT Digital Twin
draft, Writing – review & editing. Lucie Klus: Visualization, Writing –
review & editing. Justyna Skibińska: Writing – review & editing. DTW Dynamic Time Warping
Salwa Saafi: Writing – review & editing. Pavel Pascacio: Writing – re- ECG Electrocardiogram
view & editing. Laura Flueratoru: Writing – review & editing. Darwin EEG Electroencephalogram
Quezada Gaibor: Writing – review & editing. Nadezhda Chukhno: FPV First-person view
Writing – review & editing. Olga Chukhno: Writing – review & eMTC enhanced Machine Type Communication
editing. Asad Ali: Writing – review & editing. Asma Channa: Writing – GNSS Global Navigation Satellite System
review & editing. Ekaterina Svertoka: Writing – review & editing. GLONASS Global Navigation Satellite System
Waleed Bin Qaim: Writing – review & editing. Raúl Casanova- GPRS General Packet Radio Services
Marqués: Writing – review & editing. Sylvia Holcer: Writing – re- GDPR General Data Protection Regulation
view & editing. Joaquín Torres-Sospedra: Writing – review & editing.
GPS Global Positioning System
Sven Casteleyn: Writing – review & editing. Giuseppe Ruggeri:
GRU Gated Recurrent Units
Writing – review & editing. Giuseppe Araniti: Writing – review &
H2H Human-to-Human interaction
editing. Radim Burget: Writing – review & editing. Jiri Hosek: Writ-
ing – review & editing. Elena Simona Lohan: Project administration, HAR Human Activity Recognition
Writing – review & editing. HMD Head-Mounted Display
HRD Health Recording Devices
Declaration of competing interest HW Hardware
HC-RAN Heterogeneous Cloud Radio Access Network
The authors declare that they have no known competing finan- HCI Human-Computer Interaction
cial interests or personal relationships that could have appeared to IAAFT Iterated Adjusted Fourier Transform
influence the work reported in this paper. ICD Industrial Clothing Division

25
A. Ometov et al. Computer Networks 193 (2021) 108074

ICT Information and Communications Technology RF Radio Frequency

IDDS Insertable Drug Deliverables Systems RC Remote control
IEEE Institute of Electrical and Electronics Engineers RFID Radio Frequency Identification
IETF Internet Engineering Task Force RNSS Regional Navigation Satellite Systems
IF Interface RSS Received Signal Strength
IIoT Industrial Internet of Things SAR Specific Absorption Rate
ILS Indoor Location Systems SDN Software-Defined Networks
IoT Internet of Things SIM subscriber identification module
IoWT Internet of Wearable Things SIoT Social Internet of Things
ITB Inside-the-body SOC System on Chip
IPS Indoor Positioning System SSD single-sided deafness
IEC International Electrotechnical Commission SW Software
IP Ingress Protection Code SW Software
IR Identifiable Infrared TCP/IP Transmission Control Protocol/Internet Protocol
ISM Industrial, Scientific, and Medical THz Terahertz
ISO International Organization for Standardization
TOA Time-of-Arrival
KARMA Knowledge-based Augmented Reality for Maintenance
UPnP Universal Plug and Play
Assistance
UWB Ultra-Wide Band
LAN Local Area Network
VLC Visible Light Communications
LBS Location-Based Service
VR Virtual reality
LSI Large-Scale Integration
WAN Wide Area Network
LED Light-emitting diode
LPWA Low-Power Wide Area WBAN Wireless Body Area Network
LPWAN Low-Power Wide Area Network (protocols) WPAN Wireless Personal Area Network
LSE Large Scale Environment WSN Wireless Sensor Network
LSTM Long Short-Term Memory WET Wireless Energy Transfer
LTE Long Term Evolution Wi-Fi Wireless Fidelity
LoRa Long Range LPWAN protocol WiGig Wireless Gigabit Alliance (WiFi at 60 GHz)
LOS Line-of-Sight WLAN Wireless Local Area Network
MAC Medium Access Control XR eXtended reality
M2M Machine-to-Machine
MCC Mobile Cloud Computing Appendix. Description of the wearable devices’ dataset
MCS Modulation and Coding Scheme
During the preparation of this survey, our team has analyzed the
MEC Multi-Access Edge Computing (formely Mobile Edge
market and exiting research projects, which resulted in the dataset
Computing)
allowing for easy analysis of the available wearable devices. The most
mMTC Massive Machine Type Communications
recent version of the dataset is available via Zenodo repository https:
mmWave Millimeter Wave //zenodo.org/record/4575153. As of the paper submission date, it
MR Mixed Reality provides the data about 224 wearable devices.
MRI Magnetic Resonance Imaging In particular, it consists of the following fields (some fields could be
MS Multiple Sclerosis empty due to the unavailability of the module):
MTC Machine Type Communications
• Device – name of the wearable on the market or project title;
MLP Multilayer Perceptron
• Location type – data on the wearablity, based on the proposed
ML Machine Learning classification;
MIMO Multiple Input Multiple Output • Application type based on the proposed classification;
MWC Mobile WOrld Congress • Device type based on the proposed classification;
NB-IoT Narrowband Internet of Things • Hi/low end – type of the device according to its’ functionality;
NFC Near Field Communication • Prototype or availability on the market;
NLOS Non-Line-of-Sight • Energy-related information – information either on power sup-
OGC Open Geospatial Consortium ply or on the battery type;
OS Operating System • CPU – information about the processor;
P2P Peer-to-Peer • GPU – information about the graphical card;
PAN Personal Area Network • Camera – information about the camera(s);
• Cellular – information about the cellular module and technology;
POA Phase of Arrival
• WiFi – information about the WiFi module and technology;
PoW Proof of Work
• Bluetooth – information about Bluetooth module and, optionally,
PDA personal digital assistant
version;
QoS Quality of Service • RFID – information about RFID module;
RAM Random-access memory • Other connectivity – the list of other connectivity options
RAN Radio Access Network present on the device;
RAT Radio Access Technology • GNSS – information about conventional positioning systems;
REST Representational state transfer • RAM – information about integrated RAM;

26
A. Ometov et al. Computer Networks 193 (2021) 108074

• Storage – information about available internal storage or slots [4] S. Khan, S. Parkinson, L. Grant, N. Liu, S. Mcguire, Biometric Systems Utilising
available for external memory stick; Health Data from Wearable Devices: Applications and Future Challenges in
Computer Security, ACM Comput. Surv. 53 (4) (2020) 1–29.
• Display – information about the graphical output interface;
[5] J. Zhou, Z. Cao, X. Dong, N. Xiong, A.V. Vasilakos, 4S: A Secure and Privacy-
• Audio output – information about audio output devices; Preserving Key Management Scheme for Cloud-Assisted Wireless Body Area
• Mic – presence of audio input interface; Network in m-Healthcare Social Networks, Inform. Sci. 314 (2015) 255–276.
• Release – first release date or year; [6] A. Ometov, Social, Private, and Trusted Wearable Technology under Cloud-
Aided Intermittent Wireless Connectivity (Ph.D. thesis), Tampere University of
• Price $ – market price as of 2020;
Technology, 2018.
• Luxury – information if the device could be considered as a [7] S. Tweedie, The World’s First Smartphone, Simon, Was Created 15 Years Before
luxury accessories; the iPhone, Bus. Insider (2019) [Online] https://www.businessinsider.com/
• Activity tracking – presence of the activity tracking feature with worlds-first-smartphone-simon-launched-before-iphone-2015-6?r=US&IR=T
(Accessed March 3, 2021).
details;
[8] Business Wire Inc., The Wearable Technology Ecosystem: 2016-2030 –
• Sensors – list of the sensors and actuators integrated in the Opportunities, Challenges, Strategies, Industry Verticals and Forecasts
device; for the $40 Billion Market – Research and Markets, 2016, [Online]
• Description (optional) – short description of the device; https://www.businesswire.com/news/home/20160504005813/en/Wearable-
Technology-Ecosystem-2016-2030---Opportunities-Challenges (Accessed March
• Other comments – additional notes on the device;
3, 2021).
• Link – example link with the device information. [9] I-SCOOP, Wearables Market Outlook 2020: Drivers and New Markets, 2016,
[Online] https://www.i-scoop.eu/wearables-market-outlook-2020-drivers-new-
An example of one entry could be found in the following listing: markets/ (Accessed March 3, 2021).
[10] J. Hayward, Wearable Sensors 2018-2028: Technologies, Markets & Players,
IDTechEx Ltd, 2018.
[ 1 [11] J. Smith, Wearables Shipments Will Double By 2021, Insider Inc, 2017,
{ 2 [Online] https://www.businessinsider.com/wearables-shipments-will-double-
" Device " : " F o c a l s by North " , 3 by-2021-2017-12?r=US&IR=T (Accessed March 3, 2021).
" L o c a t i o n type " : " Head_mounted " , 4 [12] J. Hayward, Wearable Technology Forecasts: 2020-2030, IDTechEx, 2020,
[Online] https://www.idtechex.com/en/research-report/wearable-technology-
" A p p l i c a t i o n type " : " Eyewear " , 5
forecasts-2020-2030/747 (Accessed March 3, 2021).
" Device type " : "AR" , 6
[13] M.T. Rashid, D. Wang, CovidSens: A Vision on Reliable Social Sensing for
" Hi or low end " : " high " , 7 COVID-19, Artif. Intell. Rev. (2020) 1–25.
" P r o t o t y p e " : " yes " , 8 [14] E. Hernández-Orallo, P. Manzoni, C.T. Calafate, J.-C. Cano, Evaluating How
" Energy−r e l a t e d i n f o r m a t i o n " : " 700 mAh/18 hours " , 9 Smartphone Contact Tracing Technology Can Reduce the Spread of Infectious
" CPU " : " Qualcomm APQ8009w" , 10 Diseases: The Case of COVID-19, 2020, IEEE Access.
"GPU" : " n/a " , 11 [15] A.K. Tripathy, A.G. Mohapatra, S.P. Mohanty, E. Kougianos, A.M. Joshi, G. Das,
" Camera " : " yes " , 12 EasyBand: A Wearable for Safety-Aware Mobility during Pandemic Outbreak,
" C e l l u l a r " : " no " , 13
IEEE Consum. Electron. Mag..
[16] V. Shubina, A. Ometov, E.S. Lohan, Technical Perspectives of Contact-Tracing
" WiFi " : " no " , 14
Applications on Wearables for COVID-19 Control, in: Proc. of 12th International
" B l u e t o o t h " : " yes " , 15
Congress on Ultra Modern Telecommunications and Control Systems and
" RFID " : " no " , 16 Workshops (ICUMT), IEEE, 2020, pp. 229–235.
" Other c o n n e c t i v i t y " : " no " , 17 [17] R.T. Llamas, Enterprise Wearable Market: Information By Technology (IoT,
" GNSS " : " no " , 18 Bluetooth, BLE), Product (Wristwear, Footwear, Eyewear), Application (Info-
"RAM" : " n/ a " , 19 tainment, IT), and Regional Outlook – Forecast Till 2026, Tech. Rep. SR1195,
" S t o r a g e " : " n/ a " , 20 Straits Research, 2019.
" D i s p l a y " : " yes " , 21 [18] DVV Media HR Group Ltd, The Benefits of Wearable Technology in the Work-
place, 2018, [Online] https://www.employeebenefits.co.uk/benefits-wearable-
" Audio output " : " yes " , 22
technology-workplace/ (Accessed March 3, 2021).
" Mic " : " yes " , 23
[19] Guardian24 Ltd, 4 Benefits of Wearable Technology, 2017, [Online] https:
" R e l e a s e " : " n/a " , 24
//guardian24.co.uk/4-benefits-of-wearable-technology/ (Accessed March 3,
" P r i c e $ " : " n/ a " , 25 2021).
" Lyxury " : " yes " , 26 [20] N. Costa, 6 Ways Wearable Tech Makes Everything Better (in Work and
" A c i t i v i t y t r a c k i n g " : " yes " , 27 Life), Koombea, Inc, 2018, [Online] https://www.koombea.com/blog/6-ways-
" S e n s o r s " : " 9− a x i s IMU , Ambient L i g h t Sensor , 28 wearable-tech-makes-everything-better/ (Accessed March 3, 2021).
Proximity sensor " , 29 [21] R.T. Llamas, IDC Forecasts Steady Double-Digit Growth for Wearables As
" D e s c r i p t i o n " : " Smart AR g l a s s e s . Could be 30 New Capabilities and Use Cases Expand the Market Opportunities, Inter-
national Data Group (IDG), 2019, [Online] https://www.idc.com/getdoc.jsp?
used as h a n d s f r e e and with a s s i s t a n t s . " , 31
containerId=prUS44930019 (Accessed March 3, 2021).
" Other comments " : " IP 55 " , 32
[22] Welbi, 5 Health Benefits of Wearable Tech, 2019, [Online] https://www.welbi.
" Link " : " h t t p s : //www. bynorth . com/ t e c h " 33
co/single-post/5-health-benefits-of-wearable-tech (Accessed March 3, 2021).
}, 34 [23] C. Kresser, The benefits of using wearable technology for health track-
<... > 35 ing, 2019, [Online] https://chriskresser.com/the-benefits-of-using-wearable-
] 36 technology-for-health-tracking/ (Accessed March 3, 2021).
[24] A. Ometov, D. Kozyrev, V. Rykov, S. Andreev, Y. Gaidamaka, Y. Koucheryavy,
Reliability-centric Analysis of Offloaded Computation in Cooperative Wearable
Applications, Wirel. Commun. Mob. Comput., 2017.
References [25] M. Pedram, M. Rofouei, F. Fraternali, Z.E. Ashari, H. Ghasemzadeh,
Resource-Efficient Computing in Wearable Systems, arXiv preprint arXiv:1907.
[1] Cisco Visual Networking Index, Global Mobile Data Traffic Forecast Update, 03247.
2017–2022 White Paper, 2019, [Online] https://www.cisco.com/c/en/us/ [26] O. Galinina, H. Tabassum, K. Mikhaylov, S. Andreev, E. Hossain, Y. Kouch-
solutions/collateral/service-provider/visual-networking-index-vni/white-paper- eryavy, On Feasibility of 5G-grade Dedicated RF Charging Technology for
c11-738429.html (Accessed March 3, 2021). Wireless-Powered Wearables, IEEE Wirel. Commun. 23 (2) (2016) 28–37.
[2] W.B. Qaim, A. Ometov, A. Molinaro, I. Lener, C. Campolo, E.S. Lohan, J. Nurmi, [27] Y. Li, Y. Chen, C.S. Chen, Z. Wang, Y.-h. Zhu, Charging while Moving:
Towards Energy Efficiency in the Internet of Wearable Things: A Systematic Deploying Wireless Chargers for Powering Wearable Devices, IEEE Trans. Veh.
Review, IEEE Access. Technol. 67 (12) (2018) 11575–11586.
[3] T. Luczak, R. Burch, E. Lewis, H. Chander, J. Ball, State-of-the-Art Review of [28] M. Cheikh, Y. Vassilieff, A. Fortes, R. Benbouhout, H. Foligné, Human body
Athletic Wearable Technology: What 113 Strength and Conditioning Coaches exposure to low frequency wireless charging: Direct coupling mechanisms and
and Athletic Trainers from the USA Said about Technology in Sports, Int. J. interferences with medical devices, in: Proc. of Progress in Electromagnetics
Sports Sci. Coach. 15 (1) (2020) 26–40. Research Symposium-Fall (PIERS-FALL), IEEE, 2017, pp. 1048–1054.

27
A. Ometov et al. Computer Networks 193 (2021) 108074

[29] B. Garland, Future of Wireless Charging, TechNative, 2018, [Online] https: [52] G. Myre, From Wristwatches To Radio, How World War I Ushered in
//www.mymemory.co.uk/blog/future-of-wireless-charging/ (Accessed March 3, the Modern World, NPR, 2017, [Online] https://www.npr.org/sections/
2021). parallels/2017/04/02/521792062/from-wristwatches-to-radio-how-world-war-
[30] S. Seneviratne, Y. Hu, T. Nguyen, G. Lan, S. Khalifa, K. Thilakarathna, M. i-ushered-in-the-modern-world?t=1568188507348 (Accessed March 3, 2021).
Hassan, A. Seneviratne, A Survey of Wearable Devices and Challenges, IEEE [53] J. Stamp, A Partial History of Headphones, 2013, [Online] https:
Commun. Surv. Tutor. 19 (4) (2017) 2573–2620. //web.archive.org/web/20130409012634/http://blogs.smithsonianmag.com/
[31] J.J. Ferreira, C.I. Fernandes, H.G. Rammal, P.M. Veiga, Wearable Technology design/2013/03/a-partial-history-of-headphones/ (Accessed March 3, 2021).
and Consumer Interaction: A Systematic Review and Research Agenda, [54] B. Ticknor, Virtual Reality and the Criminal Justice System: Exploring the
Comput. Hum. Behav. 118 (2021) 106710, http://dx.doi.org/10.1016/j. Possibilities for Correctional Rehabilitation, Lexington Books, 2018.
chb.2021.106710, URLhttps://www.sciencedirect.com/science/article/pii/ [55] M.L. Heilig, Sensorama simulator, 1962.
S0747563221000327. [56] B. Rhodes, A Brief History of Wearable Computing, MIT, 2019, [Online]
https://www.media.mit.edu/wearables/lizzy/timeline.html (Accessed March 3,
[32] Y. Xue, A Review on Intelligent Wearables: Uses and Risks, Hum. Behav.
2021).
Emerg. Technol. 1 (4) (2019) 287–294, http://dx.doi.org/10.1002/hbe2.173,
[57] E.O. Thorp, Beat the Dealer: A Winning Strategy for the Game of Twenty One,
arXiv:https://onlinelibrary.wiley.com/doi/pdf/10.1002/hbe2.173 URL https://
Vol. 310, Vintage, 1966.
onlinelibrary.wiley.com/doi/abs/10.1002/hbe2.173.
[58] E.O. Thorp, The Invention of the First Wearable Computer, in: Digest of Papers.
[33] N. Niknejad, W.B. Ismail, A. Mardani, H. Liao, I. Ghani, A Comprehensive
Second International Symposium on Wearable Computers (Cat. No. 98EX215),
Overview of Smart Wearables: The State of the Art Literature, Recent Advances,
IEEE, 1998, pp. 4–8.
and Future Challenges, Eng. Appl. Artif. Intell. 90 (2020) 103529, http://dx.
[59] E. Kurland, History of VR, in: Virtual Reality Filmmaking, Routledge, 2017, pp.
doi.org/10.1016/j.engappai.2020.103529, URL https://www.sciencedirect.com/
7–17.
science/article/pii/S0952197620300348.
[60] E. Ackerman, The Man Who Invented VR Goggles 50 years Too Soon, IEEE
[34] W.A.D.M. Jayathilaka, K. Qi, Y. Qin, A. Chinnappan, W. Serrano-Garcí a, C. Spectr. (2016).
Baskar, H. Wang, J. He, S. Cui, S.W. Thomas, S. Ramakrishna, Significance of [61] T. Emerson, Mastering the Art of VR: On Becoming the HIT Lab Cybrarian,
Nanomaterials in Wearables: A Review on Wearable Actuators and Sensors, Adv. Electron. Libr. 11 (6) (1993) 385–391.
Mater. 31 (7) (2019) 1805921, http://dx.doi.org/10.1002/adma.201805921, [62] S.D. Guler, M. Gannon, K. Sicchio, A Brief History of Wearables, in: Crafting
arXiv:https://onlinelibrary.wiley.com/doi/pdf/10.1002/adma.201805921 URL Wearables, Springer, 2016, pp. 3–10.
https://onlinelibrary.wiley.com/doi/abs/10.1002/adma.201805921. [63] Unique Watch Guide, Calculator watches, 2015, [Online] http://www.
[35] C. McCallum, J. Rooksby, C.M. Gray, Evaluating the Impact of Physical Activity uniquewatchguide.com/calculator-watches.html (Accessed March 3, 2021).
Apps and Wearables: Interdisciplinary Review, JMIR Mhealth Uhealth 6 (3) [64] R. Bensene, Hewlett Packard HP-01 Calculator Wristwatch, 2019, [Online]
(2018) e58, http://dx.doi.org/10.2196/mhealth.9054, URL http://mhealth.jmir. https://www.oldcalculatormuseum.com/hp-01.html (Accessed March 3, 2021).
org/2018/3/e58/. [65] MoHPC, The Museum of HP Calculators: HP-01, 2019, [Online] https://www.
[36] P.J. Soh, G.A. Vandenbosch, M. Mercuri, D.M.-P. Schreurs, Wearable Wireless hpmuseum.org/hp01.htm (Accessed March 3, 2021).
Health Monitoring: Current Developments, IEEE Microwave Mag. 16 (4) (2015) [66] K.A. Popat, P. Sharma, Wearable Computer Applications a Future Perspective,
55–70. Int. J. Eng. Innov. Technol. 3 (1) (2013) 213–217.
[37] A. Mosenia, S. Sur-Kolay, A. Raghunathan, N.K. Jha, Wearable Medical Sensor- [67] B. Harrison, Design Moment: Sony Walkman, 1979: The Silver and Blue Portable
based System Design: A Survey, IEEE Trans. Multi-Scale Comput. Syst. 3 (2) Personal Stereo Sold beyond All Expectations, The Irish Times, 1979, [On-
(2017) 124–138. line] https://www.irishtimes.com/life-and-style/homes-and-property/interiors/
[38] J.M. Peake, G. Kerr, J.P. Sullivan, A Critical Review of Consumer Wearables, design-moment-sony-walkman-1979-1.3465629 (Accessed March 3, 2021).
Mobile Applications, and Equipment for Providing Biofeedback, Monitor- [68] S. Mann, J. Fung, C. Aimone, A. Sehgal, D. Chen, Designing EyeTap Digital
ing Stress, and Sleep in Physically Active Populations, Front. Physiol. 9 Eyeglasses for Continuous Lifelong Capture and Sharing of Personal Experiences,
(2018) 743, http://dx.doi.org/10.3389/fphys.2018.00743, URL https://www. 2005.
frontiersin.org/article/10.3389/fphys.2018.00743. [69] O. Peltola, Introduction to Wearable Healthcare Technology (Master’s thesis),
Juväskylän Yliopisto, Faculty of Information Technology, 2017.
[39] P. Kumari, L. Mathew, P. Syal, Increasing Trend of Wearables and Multimodal
[70] Tech Blog, Before Smartwatches There was the Seiko UC-2000 Wrist Computer,
Interface for Human Activity Monitoring: A Review, Biosens. Bioelectron. 90
2015, [Online] https://www.dailymail.co.uk/sciencetech/article-2584437/Is-
(2017) 298–307.
wearable-computer-300-year-old-Chinese-abacus-ring-used-Qing-Dynasty-
[40] The College of Optometrists, The Invention of Spectacles: How and
help-traders.html (Accessed March 3, 2021).
Where Glasses May Have Begun, 2019, [Online] https://www.college-
[71] Polygon, A Visual Guide to Watch Gaming’s 40-year History, 2019, [On-
optometrists.org/the-college/museum/online-exhibitions/virtual-spectacles-
line] https://www.polygon.com/a/smartwatch-history-guide-evolution/watch-
gallery/the-invention-of-spectacles.html (Accessed March 3, 2021).
timeline (Accessed March 3, 2021).
[41] T. Cassidy, A Short History of Spectacles, [72] J. Peddie, Augmented Reality: Where We Will All Live, Springer, 2017.
[42] G. Oestmann, The Origins and Diffusion of Watches in the Renaissance: [73] R. Want, A. Hopper, V. Falcão, J. Gibbons, The Active Badge Location System,
Germany, Comune di Cremona, 2016, pp. 141–143. ACM Trans. Inf. Syst. 10 (1) (1992) 91–102, http://dx.doi.org/10.1145/128756.
[43] Oldest.org, 10 Oldest Watches in the World, 2019, [Online] http://www.oldest. 128759.
org/technology/watches/(Accessed March 3, 2021). [74] D. Greaves, Olivetti research active badge, 2010, [Online] http://koo.corpus.
[44] D. Boettcher, The Invention of the Wristwatch, 2019, [Online] https://www. cam.ac.uk/projects/badges/index.html (Accessed March 3, 2021).
vintagewatchstraps.com/wristwatchinvention.php (Accessed March 3, 2021). [75] O. Amft, P. Lukowicz, From Backpacks to Smartphones: Past, Present, and
[45] U. Friedman, A Brief History of the Wristwatch: The Military Ori- Future of Wearable Computers, IEEE Pervas. Comput. 8 (3) (2009) 8–13.
gins of Wearable Tech, a Century Before the Apple Watch, 2015, [76] S. Feiner, B. Macintyre, D. Seligmann, Knowledge-Based Augmented Reality,
[Online] https://www.theatlantic.com/international/archive/2015/05/history- Commun. ACM 36 (7) (1993) 53–62.
wristwatch-apple-watch/391424/ (Accessed March 3, 2021). [77] M. Lamming, M. Flynn, Forget-me-Not: Intimate Computing in Support of
[46] E. Zolfagharifard, Is This the First Wearable Computer? 300-year-Old Chinese Human Memory, in: Proc. of International Symposyum on Next Generation
Abacus Ring was Used During the Qing Dynasty to Help Traders, Daily Mail Human Interface, 1994, p. 4.
(2014) [Online] https://www.dailymail.co.uk/sciencetech/article-2584437/Is- [78] D. Knight, A history of palm part 2: Palm PDAs phones 1996 to 2003,
wearable-computer-300-year-old-Chinese-abacus-ring-used-Qing-Dynasty- 2017, [Online] https://lowendmac.com/2016/a-history-of-palm-part-2-palm-
help-traders.html(Accessed March 3, 2021). pdas-and-phones-1996-to-2003/ (Accessed March 3, 2021).
[79] D. Papadopoulos, MBracelet (1999): New York, New York, USA & Lon-
[47] T. Berman, The Smart Ring: From the 17th Century Wearable Abacus
don, UK with the Knowledge Lab, NCR, 2019, [Online] http://www.
to Today, 2017, [Online] https://interestingengineering.com/smart-ring-17th-
fashionabletechnology.org/press/photosbook/hi-res/ft-book-p117.pdf (Accessed
century-wearable-abacus-today (Accessed March 3, 2021).
March 3, 2021).
[48] The Public Domain Review, Dr Julius Neubronner’s Miniature Pigeon
[80] J.-H. Kim, A Study on the Characteristics of Modern Fashion Design for Digital
Camera, 2019, [Online] https://publicdomainreview.org/collections/dr-julius-
Nomadic Culture, Fash. Text. Res. J. 9 (1) (2007) 6–14.
neubronners-miniature-pigeon-camera/ (Accessed March 3, 2021). [81] Plantronics, An Industry Pioneer Reflects on a Decade’S Worth of Achievements
[49] A. DenHoed, The Turn-of-the-Century Pigeons That Photographed Earth from To Honour the 10th Anniversary of Bluetooth Headsets, 2010, [Online]
Above, 2018, [Online] https://publicdomainreview.org/collections/dr-julius- https://newsroom.poly.com/press-release/consumer/plantronics-celebrates-10-
neubronners-miniature-pigeon-camera/(Accessed March 3, 2021), years-bluetooth-headset-innovation (Accessed March 3, 2021).
[50] The National Archives, Fighting talk: First World War Telecommuni- [82] J. McHugh, Wrist-Top Revolution, 2001, [Online] https://www.wired.com/
cations, 2019, [Online] http://www.nationalarchives.gov.uk/first-world-war/ 2003/04/fossil/ (Accessed March 3, 2021).
telecommunications-in-war/ (Accessed March 3, 2021). [83] C. Evers, Researching Action Sport with a GoPro Camera: An Embodied and
[51] D. Hinhs, Walkie Talkie, 2019, [Online] http://www.dlhings.ca/walkietalkie. Emotional Mobile Video Tale of the Sea, Masculinity and Men-Who-Surf, in:
html (Accessed March 3, 2021). Researching Embodied Sport, Routledge, 2015, pp. 145–162.

28
A. Ometov et al. Computer Networks 193 (2021) 108074

[84] S. Shoul, Nike and Apple Team Up To Launch Nike+iPod, 2006, [On- [111] L.M. Castano, a.B. Flatau, Smart Fabric Sensors and E-Textile Technologies: A
line] https://www.apple.com/newsroom/2006/05/23nike-and-apple-team-up- Review, Smart Mater. Struct. 23 (5) (2014) 053001.
to-launch-nike-ipod/ (Accessed March 3, 2021). [112] S. Imani, a.J. BandOdkar, a.V. Mohan, R. Kumar, S. Yu, J. Wang, P.P.
[85] G. Marshall, the Story of Fitbit: How a Wooden Box Became a $4 Billion Mercier, A Wearable Chemical–Electrophysiological Hybrid Biosensing System
Company, 2016, [Online] https://www.wareable.com/fitbit/youre-fitbit-and- for Real-Time Health and Fitness Monitoring, Nature Commun. 7 (1) (2016)
you-know-it-how-a-wooden-box-became-a-dollar-4-billion-company (Accessed 1–7.
March 3, 2021). [113] C.-T. Lin, C.-H. Chuang, Z. Cao, a.K. Singh, C.-S. Hung, Y.-H. Yu, M. Nascimben,
[86] OKTRACKME, Fitbit Wireless Activity and Sleep Tracker, 2013, [Online] http: Y.-T. Liu, J.-T. King, T.-P. Su, Others, Forehead EEG in Support of Future Fea-
//oktrackme.com/fitness-tracking/fitbit/ (Accessed March 3, 2021). sible Personal Healthcare Solutions: Sleep Management, Headache Prevention,
[87] Inc. Media Genesis, Watch Out: A Timely History of the Smartwatch, 2019, [On- and Depression Treatment, IEEE Access 5 (2017) 10612–10621.
line] https://mediag.com/blog/watch-out-a-timely-history-of-the-smartwatch/ [114] J.L. Park, M.M. Fairweather, D.I. Donaldson, Making the Case for Mobile
(Accessed March 3, 2021). Cognition: EEG and Sports Performance, Neurosci. Biobehav. Rev. 52 (2015)
[88] F. Graham, Samsung S9110 Watch Phone Lets Our Wrists Do the Talking, 117–130.
2009, [Online] https://www.cnet.com/news/samsung-s9110-watch-phone-lets- [115] C. Pittman, J.J. LaViola Jr, Exploring Head Tracked Head Mounted Displays for
our-wrists-do-the-talking/ (Accessed March 3, 2021). First Person Robot Teleoperation, in: Proc. of the 19th International Conference
[89] T. Jowitt, Tales in Tech History: Pebble Smartwatch, 2017, [Online] https:// on Intelligent User Interfaces, 2014, pp. 323–328.
www.silicon.co.uk/mobility/tales-tech-history-pebble-smartwatch-220973 (Ac- [116] E. Taylor, Lights, Camera, Redaction... Police Body-Worn Cameras; Autonomy,
cessed March 3, 2021). Discretion and Accountability, Surveill. Soc. 14 (1) (2016) 128–132.
[90] C. Steele, A Look Back At Pebble’S Rise and Fall, 2016, [Online] [117] R. Pirmagomedov, D. Moltchanov, a. Ometov, K. Muhammad, S. Andreev,
https://web.archive.org/web/20170718150503/https://www.pcmag.com/ Y. Koucheryavy, Facilitating MmWave Mesh Reliability in PPDR Scenarios
article/350161/a-look-back-at-pebbles-rise-and-fall (Accessed March 3, 2021). Utilizing Artificial Intelligence, IEEE Access 7 (2019) 180700–180712.
[91] PBC Kickstarter, Pebble: E-Paper Watch for iPhone and Android, 2012, [118] Teslasuit, Full Body Haptic VR Suite, 2019, [Online] https://teslasuit.io
[Online] https://www.kickstarter.com/projects/getpebble/pebble-e-paper- (Accessed March 3, 2021).
watch-for-iphone-and-android (Accessed March 3, 2021). [119] R.W. Lindeman, R. Page, Y. Yanagida, J.L. Sibert, Towards Full-Body Haptic
[92] E. Betters, Google Glass: A Brief History, 2015, [Online] https://www.pocket- Feedback: The Design and Deployment of a Satialized Vibrotactile Fedback Sys-
lint.com/ar-vr/news/google/132399-google-glass-a-brief-history (Accessed tem, in: Proc. of ACM Symposium on Virtual Reality Software and Technology,
March 3, 2021). 2004, pp. 146–149.
[93] E. Mack, Confirmed: Google Glass Arrives in 2013, and under $1, 500, 2013,
[120] A. Al Maimani, A. Roudaut, Frozen Suit: Designing a Changeable Stiffness Sit
[Online] https://www.cnet.com/news/confirmed-google-glass-arrives-in-2013-
and Its Application to haptic Games, in: Proc. of CHI Conference on Human
and-under-1500/ (Accessed March 3, 2021).
Factors in Computing Systems, 2017, pp. 2440–2448.
[94] 6 Pack Fitness, Best Fitness Trackers: Top 5 Fitness BandS of 2014, 2014,
[121] K. Antonakoglou, X. Xu, E. Steinbach, T. Mahmoodi, M. Dohler, Toward Haptic
[Online] https://www.sixpackbags.com/blogs/news/fitness-b{and}s (Accessed
Communications over the 5G Tactile Internet, IEEE Commun. Surv. Tutor. 20
March 3, 2021).
(4) (2018) 3034–3059.
[95] M. Nguyen, The Most Successful Wearables for Consumers, 2016, [Online]
[122] A.M. Okamura, Haptic Feedback in Robot-Assisted Minimally Invasive Surgery,
https://www.wearable-technologies.com/2016/01/the-most-successful-
Curr. Opin. Urol. 19 (1) (2009) 102.
wearables-for-consumers/ (Accessed March 3, 2021).
[123] S. Pradhan, The Present and Future of Ingestible Sensors – the New Taste
[96] M. Starr, Tommy Hilfiger Launches Solar Power Jackets To Charge Your
of Science, 2019, [Online] https://www.prescouter.com/2019/01/ingestible-
Phone, 2014, [Online] https://www.cnet.com/news/tommy-hilfiger-launches-
sensors-innovations/ (Accessed March 3, 2021).
solar-power-jackets-to-charge-your-phone/ (Accessed March 3, 2021).
[124] A.S. Barbone, M. Meftah, K. Markiewicz, K. Dellimore, Beyond Wearables and
[97] D. D’Orazio, Google Reveals Android Wear, an Operating System for
Implantables: A Scoping Review of Insertable Medical Devices, Biomed. Phys.
Smartwatches, 2014, [Online] https://www.theverge.com/2014/3/18/
Eng. Express 5 (6) (2019) 062002.
5522226/google-reveals-android-wear-an-operating-system-designed-for
[125] L.M. Ni, Y. Liu, Y.C. Lau, A.P. Patil, LANDMARC: Indoor Location Sensing using
(Accessed March 3, 2021).
Active RFID, in: Proc. of 1st International Conference on Pervasive Computing
[98] T. Bajarin, The Real Reason Apple Made the Apple Watch, TIME, 2016, [Online]
and Communications (PERCOM), IEEE, 2003, pp. 407–415.
https://time.com/4323318/apple-watch-steve-jobs-health/ (Accessed March 3,
2021). [126] J. Ting, A. Del Vecchio, D. Friedenberg, M. Liu, C. Schoenewald, D. Sarma,
[99] K.-J. Brickwood, G. Watson, J. O’Brien, a.D. Williams, Consumer-Based Wear- J. Collinger, S. Colachis, G. Sharma, D. Farina, et al., A Wearable Neural
able Activity Trackers Increase Physical Activity Participation: Systematic Interface for Detecting and Decoding Attempted Hand Movements in a Person
Review and Meta-Analysis, JMIR MHealth and UHealth 7 (4) (2019) e11819. with Tetraplegia, in: Proc. of 41st Annual International Conference of the
IEEE Engineering in Medicine and Biology Society (EMBC), IEEE, 2019, pp.
[100] C. Attig, T. Franke, I. Track, I Track Therefore I Walk–Exploring the Motiva-
1930–1933.
tional Costs of Wearing Activity Trackers in Actual Users, Int. J. Hum.-Comput.
Stud. 127 (2019) 211–224. [127] A. Cutrone, S. Micera, Implantable Neural Interfaces and Wearable Tactile
[101] E. Olshannikova, A. Ometov, Y. Koucheryavy, Towards Big Data Visualiza- Systems for Bidirectional Neuroprosthetics Systems, Adv. Healthc. Mater. 8 (24)
tion for Augmented Reality, in: Proc. of IEEE 16th Conference on Business (2019) 1801345.
Informatics, Vol. 2, IEEE, 2014, pp. 33–37. [128] M. Chae, K. Chen, W. Liu, J. Kim, M. Sivaprakasam, A 4-Channel Wearable
[102] M.C. Tom Dieck, T.H. Jung, D. Tom Dieck, Enhancing Art Gallery Visitors’ Wireless Neural Recording System, in: Proc. of IEEE International Symposium
Learning Experience using Wearable Augmented Reality: Generic Learning on Circuits and Systems, IEEE, 2008, pp. 1760–1763.
Outcomes Perspective, Curr. Issues Tour. 21 (17) (2018) 2014–2034. [129] L.-D. Liao, C.-Y. Chen, I.-J. Wang, S.-F. Chen, S.-Y. Li, B.-W. Chen, J.-Y. Chang,
[103] I.P. Tussyadiah, T.H. Jung, M.C. Tom Dieck, Embodiment of Wearable Aug- C.-T. Lin, Gaming Control using a Wearable and Wireless EEG-Based Brain-
mented Reality Technology in Tourism Experiences, J. Travel Res. 57 (5) (2018) Computer Interface Device with Novel Dry Foam-Based Sensors, J. Neuroeng.
597–611. Rehabil. 9 (1) (2012) 5.
[104] S.K. Ong, A.Y.C. Nee, Virtual and Augmented Reality Applications in [130] G. Cattan, C. Mendoza, a. Andreev, M. Congedo, Recommendations for Integrat-
Manufacturing, Springer Science & Business Media, 2013. ing a P300-Based Brain Computer Interface in Virtual Reality Environments for
[105] S. Alkhalifa, M. Al-Razgan, Enssat: Wearable Technology Application for Gaming, Computers 7 (2) (2018) 34.
the Deaf and Hard of Hearing, Multimedia Tools Appl. 77 (17) (2018) [131] S. Draper, 4 Smart Sleep Masks To Help You Get Some Sleep While
22007–22031. on the Go, Wearable Technologies International, 2019, [Online]
[106] S.-N. Yao, Headphone-Based Immersive Audio for Virtual Reality Headsets, IEEE https://www.wearable-technologies.com/2019/07/the-4-best-smart-sleep-
Trans. Consum. Electron. 63 (3) (2017) 300–308. masks-to-help-you-get-some-sleep-while-on-the-go/ (Accessed March 3, 2021).
[107] M.L. Hammock, a. Chortos, B.C.-K. Tee, J.B.-H. Tok, Z. Bao, 25th Anniversary [132] Silentmode’S PowerMask Is a $200 Connected Relaxation Mask, 2021,
Article: The Evolution of Electronic Skin (E-Skin): A Brief History, Design [Online] https://techcrunch.com/2018/11/23/silentmodes-powermask-is-a-
Considerations, and Recent Progress, Adv. Mater. 25 (42) (2013) 5997–6038. 200-connected-relaxation-mask/ (Accessed March 3, 2021).
[108] S. Gong, D.T. Lai, B. Su, K.J. Si, Z. Ma, L.W. Yap, P. Guo, W. Cheng, [133] Ripple Safety, The Revolutionary Safety Button, 2019, [Online] https://
Highly Stretchy Black Gold E-Skin Nanopatches As Highly Sensitive Wearable ripplesafety.com(Accessed March 3, 2021).
Biomedical Sensors, Adv. Electron. Mater. 1 (4) (2015) 1400063. [134] J. Patel, R. Hasan, Smart Bracelets: Towards Automating Personal Safety
[109] R. Dahiya, N. Yogeswaran, F. Liu, L. Manjakkal, E. Burdet, V. Hayward, H. using Wearable Smart Jewelry, in: Proc. of 15th IEEE Annual Consumer
JÖRntell, Large-Area Soft E-Skin: The Challenges beyond Sensor Designs, Proc. Communications & Networking Conference (CCNC), IEEE, 2018, pp. 1–2.
IEEE 107 (10) (2019) 2016–2033. [135] A. Ferrone, F. Maita, L. Maiolo, M. Arquilla, A. Castiello, A. Pecora, X. Jiang,
[110] a. Komolafe, R. Torah, Y. Wei, H. Nunes-Matos, M. Li, D. Hardy, T. Dias, M. Tu- C. Menon, L. Colace, Wearable Band for Hand Gesture Recognition Based on
dor, S. Beeby, Integrating Flexible Filament Circuits for E-Textile Applications, Strain Sensors, in: Proc. of 6th IEEE International Conference on Biomedical
Adv. Mater. Technol. 4 (7) (2019) 1900176. Robotics and Biomechatronics (BioRob), IEEE, 2016, pp. 1319–1322.

29
A. Ometov et al. Computer Networks 193 (2021) 108074

[136] H. Liu, L. Wang, Robot Collaboration, Gesture Recognition for Human-: A [161] E. Olshannikova, A. Ometov, Y. Koucheryavy, T. Olsson, Visualizing Big Data
Review, Int. J. Ind. Ergon. 68 (2018) 355–367. with Augmented and Virtual Reality: Challenges and Research Agenda, J Big
[137] Z. Yang, Q. Zhou, L. Lei, K. Zheng, W. Xiang, An IoT-Cloud Based Wearable Data 2 (1) (2015) 22.
ECG Monitoring System for Smart Healthcare, J. Med. Syst. 40 (12) (2016) [162] C. Fertleman, P. Aubugeau-Williams, C. Sher, A.-N. Lim, S. Lumley, S. Delacroix,
286. X. Pan, A Discussion of Virtual Reality as a New Tool for Training Healthcare
[138] M. Zubair, C. Yoon, H. Kim, J. Kim, J. Kim, Smart Wearable Band for Stress Professionals, Front. Public Health 6 (2018) 44.
Detection, in: Proc. of 5th International Conference on IT Convergence and [163] P. Milgram, F. Kishino, A Taxonomy of Mixed Reality Visual Displays, IEICE
Security (ICITCS), IEEE, 2015, pp. 1–4. Trans. Inf. Syst. 77 (12) (1994) 1321–1329.
[139] NEOFECT, Smart Glove for Hand Rehabilitation, 2016, [Online] https://www. [164] Å. Fast-Berglund, L. Gong, D. Li, Testing and Validating Extended Reality (XR)
neofect.com/en/ (Accessed March 3, 2021). Technologies in Manufacturing, Proc. Manuf. 25 (2018) 31–38.
[140] M. Zhu, Z. Sun, Z. Zhang, Q. Shi, T. He, H. Liu, T. Chen, C. Lee, Haptic- [165] C. Andrews, M.K. Southworth, J.N. Silva, J.R. Silva, Extended Reality in Medical
Feedback Smart Glove As a Creative Human-Machine Interface (HMI) for Practice, Curr. Treat. Options Cardiovasc. Med. 21 (4) (2019) 18.
Virtual/Augmented Reality Applications, Sci. Adv. 6 (19) (2020) eaaz8693. [166] W. Hoenig, C. Milanes, L. Scaria, T. Phan, M. Bolas, N. Ayanian, Mixed Reality
[141] GadgetFlow, The World’S First Heated Jacket Managed By Smartphone, for Robotics, in: Proc. of IEEE/RSJ International Conference on Intelligent
2016, [Online] https://thegadgetflow.com/portfolio/worlds-first-heated-jacket- Robots and Systems (IROS), IEEE, 2015, pp. 5382–5387.
managed-smartphone/ (Accessed March 3, 2021). [167] N. Kolokas, S. Krinidis, A. Drosou, D. Ioannidis, D. Tzovaras, Gait Matching
[142] J. Park, J. Kim, S.-Y. Kim, W.H. Cheong, J. Jang, Y.-G. Park, K. Na, Y.-T. by Mapping Wearable to Camera Privacy-Preserving Recordings: Experimental
Kim, J.H. Heo, C.Y. Lee, et al., Soft, Smart Contact Lenses with Integrations of Comparison of Multiple Settings, in: Proc. of 6th International Conference
Wireless Circuits, Glucose Sensors, and Displays, Sci. Adv. 4 (1) (2018) 9841. on Control, Decision and Information Technologies (CoDIT), IEEE, 2019, pp.
[143] J. Kim, M. Kim, M.-S. Lee, K. Kim, S. Ji, Y.-T. Kim, J. Park, K. Na, K.-H. Bae, 338–343.
H.K. Kim, Others, Wearable Smart Sensor Systems Integrated on Soft Contact [168] M.S. Mahmud, H. Fang, H. Wang, An Integrated Wearable Sensor for Unob-
Lenses for Wireless Ocular Diagnostics, 2017. trusive Continuous Measurement of Autonomic Nervous System, IEEE Internet
[144] B. Pellegrini, C. Zoppirolli, L. Bortolan, P. Zamparo, F. Schena, Gait Models Things J. 6 (1) (2018) 1104–1113.
and Mechanical Energy in Three Cross-Country Skiing Techniques, J. Exp. Biol. [169] M. Xu, F. Qian, M. Zhu, F. Huang, S. Pushp, X. Liu, Deepwear: Adaptive Local
217 (21) (2014) 3910–3918. Offloading for on-Wearable Deep Learning, IEEE Trans. Mob. Comput. 19 (2)
[145] F. Laamarti, F. Arafsha, B. Hafidh, A. El Saddik, Automated Athlete Hap- (2019) 314–330.
tic Training System for Soccer Sprinting, in: Proc. of IEEE Conference on [170] A.D. Cheok, W. Weihua, X. Yang, S. Prince, F.S. Wan, M. Billinghurst, H. Kato,
Multimedia Information Processing and Retrieval (MIPR), IEEE, 2019, pp. Interactive Theatre Experience in Embodied+ Wearable Mixed Reality Space,
303–309. in: Proc. of International Symposium on Mixed and Augmented Reality, IEEE,
[146] Cheeky Rascals, Know Your Baby Is Okay: The Owlet Smart Sock Tracks Your 2002, pp. 59–317.
[171] M. Golkarifard, J. Yang, Z. Huang, A. Movaghar, P. Hui, Dandelion: A Unified
Baby’s Heart Rate and Oxygen Levels, 2019, [Online] https://owletbabycare.
Code Offloading System for Wearable Computing, IEEE Trans. Mob. Comput.
co.uk (Accessed March 3, 2021).
18 (3) (2018) 546–559.
[147] H. Witt, T. Nicolai, H. Kenn, Designing a Wearable User Interface for
[172] H.-J. Hong, C.-L. Fan, Y.-C. Lin, C.-H. Hsu, Optimizing Cloud-based Video
HandS-Free Interaction in Maintenance Applications, in: Proc. of IEEE Inter-
Crowdsensing, IEEE Internet Things J. 3 (3) (2016) 299–313.
national Conference on Pervasive Computing and Communications Workshops
[173] Is Lithium-Ion the Ideal Battery?, 2021, [Online] Available: https://
(PERCOMW’06), IEEE, 2006, pp. 4–pp.
batteryuniversity.com/learn/archive/is_lithium_ion_the_ideal_battery (Accessed
[148] B. Heater, Neofect’s Powered Glove for People with Paralysis is Shipping
March 3, 2021).
this Summer, 2019, [Online] https://techcrunch.com/2019/01/07/neofects-
[174] Lithium coin cells, 2021, [Online] Available: https://www.microbattery.com/
powered-glove-for-people-with-paralysis-is-shipping-this-summer/ (Accessed
coin-cells (Accessed March 3, 2021).
March 3, 2021).
[175] Lithium-ion vs lithium-polymer: what’s the difference?, 2021, [Online]
[149] J. Blake, H.B. Gurocak, Haptic Glove with MR Brakes for Virtual Reality,
Available: https://www.androidauthority.com/lithium-ion-vs-lithium-polymer-
IEEE/ASME Trans. Mechatronics 14 (5) (2009) 606–615.
whats-the-difference-27608/ (Accessed March 3, 2021).
[150] F. El-Amrawy, M.I. Nounou, Are Currently Available Wearable Devices for
[176] R. Carroll, R. Cnossen, M. Schnell, D. Simons, Continua: An Interoperable
Activity Tracking and Heart Rate Monitoring Accurate, Precise, and Medically
Personal Healthcare Ecosystem, IEEE Pervas. Comput. 6 (4) (2007) 90–94.
Beneficial?, Healthc. Inform. Res. 21 (4) (2015) 315–320.
[177] M. Magno, L. Benini, C. Spagnol, E. Popovici, Wearable Low Power Dry Surface
[151] H. Kalantarian, N. Alshurafa, T. Le, M. Sarrafzadeh, Non-invasive Detection of
Wireless Sensor Node for Healthcare Monitoring Application, in: Proc. of IEEE
Medication Adherence Using a Digital Smart Necklace, in: Proc. of IEEE Inter-
9th International Conference on Wireless and Mobile Computing, Networking
national Conference on Pervasive Computing and Communication Workshops
and Communications (WiMob), IEEE, 2013, pp. 189–195.
(PerCom Workshops), IEEE, 2015, pp. 348–353. [178] M.C. Domingo, An Overview of the Internet of Things for People with
[152] N. Alshurafa, H. Kalantarian, M. Pourhomayoun, J.J. Liu, S. Sarin, B. Shahbazi, Disabilities, J. Netw. Comput. Appl. 35 (2) (2012) 584–596.
M. Sarrafzadeh, Recognition of Nutrition Intake Using Time-Frequency Decom- [179] R. Djapic, G. Vivier, B. Zhen, J. Wang, J. Lee, W. Haiming, Wearables White
position in a Wearable Necklace Using a Piezoelectric Sensor, IEEE Sens. J. 15 Paper, White paper, TNO etc, 2018.
(7) (2015) 3909–3916. [180] F. Delmastro, V. Arnaboldi, M. Conti, People-Centric Computing and
[153] Opter Life Inc., The Smart Posture Necklace: Correct Your Posture. Track Your Communications in Smart Cities, IEEE Commun. Mag. 54 (7) (2016) 122–128.
Activity, Sleep, UV Exposure, and More, 2019, [Online] https://opterlife.com/ [181] C. Mouradian, D. Naboulsi, S. Yangui, R.H. Glitho, M.J. Morrow, P.A. Polakos,
(Accessed March 3, 2021). A Comprehensive Survey on Fog Computing: State-of-the-Art and Research
[154] V. Hyndavi, N.S. Nikhita, S. Rakesh, Smart Wearable Device for Women Safety Challenges, IEEE Commun. Surv. Tutor. 20 (1).
Using IoT, in: Proc. of 5th International Conference on Communication and [182] P. Vilela, J. Rodrigues, R. Righi, S. Kozlov, V. Rodrigues, Looking at Fog
Electronics Systems (ICCES), IEEE, 2020, pp. 459–463. Computing for e-Health through the Lens of Deployment Challenges and
[155] S. Draper, Wearable Ultrasound Patch Monitors Blood Pressure in Deep Arteries, Applications, Sensors (20).
2019, [Online] https://www.wearable-technologies.com/2018/11/wearable- [183] F. Kraemer, A. Bråten, N. Tamkittikhun, D. Palma, Fog computing in healthcare
ultrasound-patch-monitors-blood-pressure-in-deep-arteries/ (Accessed March 3, – a review and discussion, IEEE Access PP (2017) 9206–9222, http://dx.doi.
2021). org/10.1109/ACCESS.2017.2704100.
[156] M. Klum, M. Urban, T. Tigges, A.-G. Pielmus, A. Feldheiser, T. Schmitt, R. [184] F. Bonomi, R. Milito, J. Zhu, S. Addepalli, Fog computing and its role in the
Orglmeister, Wearable Cardiorespiratory Monitoring Employing a Multimodal internet of things, in: Proceedings of the First Edition of the MCC Workshop on
Digital Patch Stethoscope: Estimation of ECG, PEP, LVETand Respiration Using Mobile Cloud Computing, MCC ’12, Association for Computing Machinery, New
a 55 mm Single-Lead ECG and Phonocardiogram, Sensors 20 (7) (2020) 2033. York, NY, USA, 2012, pp. 13–16, http://dx.doi.org/10.1145/2342509.2342513.
[157] C. Russey, These Smart Patches Monitor Your Stress to Help You [185] F. Bonomi, R. Milito, P. Natarajan, J. Zhu, Fog Computing: A Platform for
Lead a Happier, Healthier Life, 2018, [Online] https://www.wearable- Internet of Things and Analytics, Springer International Publishing, Cham, 2014,
technologies.com/2018/11/these-smart-patches-monitor-your-stress-to-help- pp. 169–186, http://dx.doi.org/10.1007/978-3-319-05029-4_7.
you-lead-a-happier-healthier-life/ (Accessed March 3, 2021). [186] ZIGURAT Innovation School, Exploring beyond cloud fog & edge computing –
[158] R. Long, M. McShane, Optical Instrument Design for Interrogation of Dermally- innovation blog, 2020, [Online] https://www.e-zigurat.com/innovation-school/
Implanted Luminescent Microparticle Sensors, in: Proc. of 30th Annual blog/beyond-cloud-fog-edge-computing/#comments (Accessed March 3, 2021).
International Conference of the IEEE Engineering in Medicine and Biology [187] A. Yousefpour, C. Fung, T. Nguyen, K. Kadiyala, F. Jalali, A. Niakanlahiji, J.
Society, IEEE, 2008, pp. 5656–5659. Kong, J.P. Jue, All One Needs to Know about Fog Computing and Related Edge
[159] D. Meetoo, L. Wong, B. Ochieng, Smart Tattoo: Technology for Monitoring Computing Paradigms: A Complete Survey, J. Syst. Archit. 98 (2019) 289–330.
Blood Glucose in the Future, Br. J. Nurs. 28 (2) (2019) 110–115. [188] J. Ren, D. Zhang, S. He, Y. Zhang, T. Li, A Survey on End-Edge-Cloud
[160] D. Pal, A. Tassanaviboon, C. Arpnikanondt, B. Papasratorn, Quality of Experi- Orchestrated Network Computing Paradigms: Transparent Computing, Mobile
ence of Smart-Wearables: From Fitness-Bands to Smartwatches, IEEE Consum. Edge Computing, Fog Computing, and Cloudlet, ACM Comput. Surv. 52 (6)
Electron. Mag. 9 (1) (2019) 49–53. (2019) 1–36.

30
A. Ometov et al. Computer Networks 193 (2021) 108074

[189] K. Dolui, S.K. Datta, Comparison of Edge Computing Implementations: Fog [216] P.A. Catherwood, J. Rafferty, S. McComb, J. McLaughlin, LPWAN Wearable
Computing, Cloudlet and Mobile Edge Computing, in: Proc. of Global Internet Intelligent Healthcare Monitoring for Heart Failure Prevention, in: Proc. of 32nd
of Things Summit (GIoTS), IEEE, 2017, pp. 1–6. International BCS Human Computer Interaction Conference Vol. 32, 2018, pp.
[190] A. Rubin, J. Ophoff, nvestigating Adoption Factors of Wearable Technology in 1–4.
Health and Fitness, in: Proc. of Open Innovations Conference (OI), 2018, pp. [217] R. Fernández-Garcia, I. Gil, An Alternative Wearable Tracking System Based on
176–186. a Low-Power Wide-Area Network, Sensors 17 (3) (2017) 592.
[191] H. Cao, M. Brown, L. Chen, R. Smith, M. Wachowicz, Lessons Learned from [218] Y. Hao, Y. Jiang, M.S. Hossain, A. Ghoneim, J. Yang, I. Humar, Data-Driven
Integrating Batch and Stream Processing Using IoT Data, in: Proc. of 6th Resource Management in a 5G Wearable Network Using Network Slicing
International Conference on Internet of Things: Systems, Management and Technology, IEEE Sens. J. 19 (19) (2018) 8379–8386.
Security (IOTSMS), 2019, pp. 32–34. [219] Y. Hao, D. Tian, G. Fortino, J. Zhang, I. Humar, Network Slicing Technology
[192] M. Obalı, B. Dursun, Z. Erdem, A.K. Görür, A Real Time Data Ware- in a 5G Wearable Network, IEEE Commun. Stand. Mag. 2 (1) (2018) 66–71.
house Approach for Data Processing, in: Proc. of 21st Signal Processing and [220] Y. Li, L. Sun, W. Wang, Exploring Device-to-Device Communication for
Communications Applications Conference (SIU), 2013, pp. 1–4. Mobile Cloud Computing, in: Proc. of IEEE International Conference on
[193] T. Abughofa, F. Zulkernine, Towards Online Graph Processing with Spark Communications (ICC), IEEE, 2014, pp. 2239–2244.
Streaming, in: Proc. of IEEE International Conference on Big Data (Big Data), [221] Z. Cheng, P. Li, J. Wang, S. Guo, Just-in-Time Code Offloading for Wearable
2017, pp. 2787–2794. Computing, IEEE Trans. Emerg. Top. Comput. 3 (1) (2015) 74–83.
[194] M. Terzi, A. Cenedese, G. Susto, A Multivariate Symbolic Approach to Activ- [222] D. Huang, L. Yang, S. Zhang, Dust: Real-time Code Offloading System for
ity Recognition for Wearable Applications, IFAC-PapersOnLine 50 (1) (2017) Wearable Computing, in: Proc. of IEEE Global Communications Conference
15865–15870. (GLOBECOM), IEEE, 2015, pp. 1–7.
[195] Q. Wen, L. Sun, X. Song, J. Gao, X. Wang, H. Xu, Time Series Data [223] B. Shi, J. Yang, Z. Huang, P. Hui, Offloading Guidelines for Augmented
Augmentation for Deep Learning: A Survey, arXiv preprint arXiv:2002.12478. Reality Applications on Wearable Devices, in: Proc. of 23rd ACM international
[196] Z. Hussain, Q.Z. Sheng, W.E. Zhang, A Review and Categorization of Techniques conference on Multimedia, 2015, pp. 1271–1274.
on Device-Free Human Activity Recognition, J. Netw. Comput. Appl. (2020) [224] M. Devos, A. Ometov, N. Mäkitalo, T. Aaltonen, S. Andreev, Y. Koucheryavy,
102738. D2D Communications for Mobile Devices: Technology Overview and Prototype
[197] L.M. Dang, K. Min, H. Wang, M.J. Piran, C.H. Lee, H. Moon, Sensor-Based and Implementation, in: Proc. of 8th International Congress on Ultra Modern
Vision-Based Human Activity Recognition: A Comprehensive Survey, Pattern Telecommunications and Control Systems and Workshops (ICUMT), IEEE, 2016,
Recognit. 108 (2020) 107561. pp. 124–129.
[198] S. Pardoel, J. Kofman, J. Nantel, E.D. Lemaire, Wearable-Sensor-Based Detection [225] Y. Yang, Y. Geng, L. Qiu, W. Hu, G. Cao, Context-Aware Task Offloading
and Prediction of Freezing of Gait in Parkinson’s Disease: A Review, Sensors for Wearable Devices, in: Proc. of 26th International Conference on Computer
19 (23) (2019) 5141. Communication and Networks (ICCCN), IEEE, 2017, pp. 1–9.
[199] M. Zhang, Z. Cui, M. Neumann, Y. Chen, An End-to-End Deep Learning [226] M. Usman, A.A. Gebremariam, U. Raza, F. Granelli, A Software-Defined Device-
Architecture for Graph Classification, in: Proc. of 32nd AAAI Conference on to-Device Communication Architecture for Public Safety Applications in 5G
Artificial Intelligence, 2018. Networks, IEEE Access 3 (2015) 1649–1654.
[200] J. Guo, L. Yang, R. Bie, J. Yu, Y. Gao, Y. Shen, A. Kos, An XGBoost- [227] M.M. Alam, D.B. Arbia, E.B. Hamida, Research Trends in Multi-Standard
Based Physical Fitness Evaluation Model Using Advanced Feature Selection Device-to-Device Communication in Wearable Wireless Networks, in: Proc.
and Bayesian Hyper-Parameter Optimization for Wearable Running Monitoring, of International Conference on Cognitive Radio Oriented Wireless Networks,
Comput. Netw. 151 (2019) 166–180. Springer, 2015, pp. 735–746.
[201] F.J. Ordóñez, D. Roggen, Deep Convolutional and lstm Recurrent Neural [228] J. Foerster, J. Lansford, J. Laskar, T. Rappaport, S. Kato, Realizing Gbps Wireless
Networks for Multimodal Wearable Activity Recognition, Sensors 16 (1) (2016) Personal Area Networks, IEEE J. Sel. Areas Commun. 27 (8) (2009) 1313–1317.
115. [229] C. Park, T.S. Rappaport, Short-Range Wireless Communications for Next-
[202] A.A. Aljarrah, A.H. Ali, Human Activity Recognition using PCA and BiL- Generation Networks: UWB, 60 GHz Millimeter-Wave WPAN, and ZigBee, IEEE
STM Recurrent Neural Networks, in: Proc. of 2nd International Conference Wirel. Commun. 14 (4) (2007) 70–78.
on Engineering Technology and Its Applications (IICETA), IEEE, 2019, pp. [230] K. Venugopal, R.W. Heath, Millimeter Wave Networked Wearables in Dense
156–160. Indoor Environments, IEEE Access 4 (2016) 1205–1221.
[203] L. Eren, T. Ince, S. Kiranyaz, A Generic Intelligent Bearing Fault Diagnosis [231] G.H. Sim, A. Loch, A. Asadi, V. Mancuso, J. Widmer, 5G millimeter-Wave and
System Using Compact Adaptive 1D CNN Classifier, J. Signal Process. Syst. 91 D2D Symbiosis: 60 GHz for Proximity-based Services, IEEE Wirel. Commun. 24
(2) (2019) 179–189. (4) (2017) 140–145.
[204] F. Samie, L. Bauer, J. Henkel, IoT Technologies for Embedded Computing: A [232] Y. Niu, Y. Li, D. Jin, L. Su, A.V. Vasilakos, A Survey of Millimeter Wave
Survey, in: Proc. of International Conference on Hardware/Software Codesign Communications (mmWave) for 5G: Opportunities and Challenges, Wirel. Netw.
and System Synthesis (CODES+ ISSS), IEEE, 2016, pp. 1–10. 21 (8) (2015) 2657–2676.
[205] E. Ferro, F. Potorti, Bluetooth and Wi-Fi Wireless Protocols: A Survey and a [233] F.O. Ombongi, H.O. Absaloms, P.L. Kibet, Resource Allocation in Millimeter-
Comparison, IEEE Wirel. Commun. 12 (1) (2005) 12–26. Wave Device-to-Device Networks, Mob. Inf. Syst. (2019).
[206] A. Ometov, E. Olshannikova, P. Masek, T. Olsson, J. Hosek, S. Andreev, Y. [234] D. Singh, A. Chattopadhyay, S.C. Ghosh, Distributed Relay Selection in Presence
Koucheryavy, Dynamic Trust Associations over Socially-Aware D2D Technology: of Dynamic Obstacles in Millimeter Wave D2D Communication, arXiv preprint
A Practical Implementation Perspective, IEEE Access 4 (2016) 7692–7702. arXiv:1910.14367.
[207] Y. Ghasempour, C.R. da Silva, C. Cordeiro, E.W. Knightly, 802.11ay: Next- [235] L. Atzori, A. Iera, G. Morabito, M. Nitti, The Social Internet of Things (SIoT)–
Generation 60 GHz Communication for 100 GB/s Wi-Fi, IEEE Commun. Mag. When Social Networks Meet the Internet of Things: Concept, Architecture and
55 (12) (2017) 186–192. Network Characterization, Comput. Netw. 56 (16) (2012) 3594–3608.
[208] M.S. Elbamby, C. Perfecto, M. Bennis, K. Doppler, Toward Low-Latency and [236] M. Roopa, S. Pattar, R. Buyya, K.R. Venugopal, S. Iyengar, L. Patnaik, Social
Ultra-Reliable Virtual Reality, IEEE Netw. 32 (2) (2018) 78–84. Internet of Things (SIoT): Foundations, thrust areas, systematic review and
[209] D. Solomitckii, V. Semkin, R. Naderpour, A. Ometov, S. Andreev, Comparative future directions, Comput. Commun. 139 (2019) 32–57.
Evaluation of Radio Propagation Properties at 15 GHz and 60 GHz Frequencies, [237] D. Moltchanov, R. Kovalchukov, M. Gerasimenko, S. Andreev, Y. Koucheryavy,
in: Proc. of 9th International Congress on Ultra Modern Telecommunications M. Gerla, Socially Inspired Relaying and Proactive Mode Selection in mmWave
and Control Systems and Workshops (ICUMT), IEEE, 2017, pp. 91–95. Vehicular Communications, IEEE Internet Things J. 6 (3) (2019) 5172–5183.
[210] H. Elayan, O. Amin, B. Shihada, R.M. Shubair, M.-S. Alouini, Terahertz Band: [238] S. Pizzi, C. Suraci, A. Iera, A. Molinaro, G. Araniti, A Sidelink-Aided Approach
The Last Piece of RF Spectrum Puzzle for Communication Systems, IEEE Open for Secure Multicast Service Delivery: From Human-Oriented Multimedia Traffic
J. Commun. Soc. 1 (2019) 1–32. to Machine Type Communications, IEEE Trans. Broadcast, 2020.
[211] P.H. Pathak, X. Feng, P. Hu, P. Mohapatra, Visible Light Communication, [239] M. Nitti, G.A. Stelea, V. Popescu, M. Fadda, When Social Networks Meet D2D
Networking, and Sensing: A Survey, Potential and Challenges, IEEE Commun. Communications: A Survey, Sensors 19 (2) (2019) 396.
Surv. Tutor. 17 (4) (2015) 2047–2077. [240] J. Li, M. Liu, J. Lu, F. Shu, Y. Zhang, S. Bayat, D.N.K. Jayakody, On Social-
[212] R. Bridgelall, Wearable RFID reader and system, 2006. Aware Content Caching for D2D-enabled Cellular Networks with Matching
[213] V. Coskun, B. Ozdenizci, K. Ok, A Survey on Near Field Communication (NFC) Theory, IEEE Internet Things J. 6 (1) (2017) 297–310.
Technology, Wirel. Pers. Commun. 71 (3) (2013) 2259–2294. [241] L. Militano, A. Orsino, G. Araniti, M. Nitti, L. Atzori, A. Iera, Trusted D2D-based
[214] H. Sun, Z. Zhang, R.Q. Hu, Y. Qian, Wearable Communications in 5G: Data Uploading in in-Band narrowband-IoT with Social Awareness, in: Proc. of
Challenges and Enabling Technologies, IEEE Veh. Technol. Mag. 13 (3) (2018) 27th Annual International Symposium on Personal, Indoor, and Mobile Radio
100–109. Communications (PIMRC), IEEE, 2016, pp. 1–6.
[215] C. Orfanidis, K. Dimitrakopoulos, X. Fafoutis, M. Jacobsson, Towards Battery- [242] G. Araniti, A. Orsino, L. Militano, L. Wang, A. Iera, Context-Aware Information
Free LPWAN Wearables, in: Proc. of 7th International Workshop on Energy Diffusion for Alerting Messages in 5G Mobile Social Networks, IEEE Internet
Harvesting & Energy-Neutral Sensing Systems, 2019, pp. 52–53. Things J. 4 (2) (2016) 427–436.

31
A. Ometov et al. Computer Networks 193 (2021) 108074

[243] C. Suraci, S. Pizzi, A. Iera, G. Araniti, Enhance the Protection of Transmitted [271] B.C. Fargas, M.N. Petersen, GPS-Free Geolocation Using LoRa in Low-Power
Data in 5G D2D Communications through the Social Internet of Things, in: WANs, in: Proc. of Global Internet of Things Summit (GIoTS), IEEE, 2017, pp.
Proc. of 29th Annual International Symposium on Personal, Indoor and Mobile 1–6.
Radio Communications (PIMRC), IEEE, 2018, pp. 376–380. [272] A. Mackey, P. Spachos, LoRa-Based Localization System for Emergency Ser-
[244] B. Han, V. Sciancalepore, O. Holland, M. Dohler, H.D. Schotten, D2D-based vices in GPS-Less Environments, in: Proc. of IEEE Conference on Computer
Grouped Random Access to Mitigate Mobile Access Congestion in 5G Sensor Communications Workshops (INFOCOM WKSHPS), IEEE, 2019, pp. 939–944.
Networks, IEEE Commun. Mag. 57 (9) (2019) 93–99. [273] C. Duan, L. Yang, Q. Lin, Y. Liu, Tagspin: High Accuracy Spatial Calibration of
[245] sony.com, Sony to Release High-Precision GNSS Receiver LSIs for IoT and RFID Antennas via Spinning Tags, IEEE Trans. Mob. Comput. 17 (10) (2018)
Wearable Devices, 2020, [Online] https://www.sony.com/ (Accessed March 3, 2438–2451.
2021). [274] F. Xiao, Z. Wang, N. Ye, R. Wang, X.-Y. Li, One More Tag Enables Fine-
[246] Ó. Belmonte-Fernández, A. Puertas-Cabedo, J. Torres-Sospedra, R. Montoliu- Grained RFID Localization and Tracking, IEEE/ACM Trans. Netw. 26 (1) (2017)
Colás, S. Trilles-Oliver, An Indoor Positioning System Based on Wearables for 161–174.
Ambient-Assisted Living, Sensors 17 (1) (2017) 36. [275] Y. Zhao, Y. Liu, L.M. Ni, VIRE: Active RFID-Based Localization Using Vir-
[247] J. Conesa, A. Pérez-Navarro, J. Torres-Sospedra, R. Montoliu, Geographical tual Reference Elimination, in: Proc. of International Conference on Parallel
and Fingerprinting Data for Positioning and Navigation Systems: Challenges, Processing (ICPP), IEEE, 2007, p. 56.
Experiences and Technology Roadmap, Academic Press, 2018. [276] Y.-L. Chen, D. Liu, S. Wang, Y.-F. Li, X.-S. Zhang, Self-powered smart active
[248] M.A. Al-Ammar, S. Alhadhrami, A. Al-Salman, A. Alarifi, H.S. Al-Khalifa, A. RFID tag integrated with wearable hybrid nanogenerator, Nano Energy 64
Alnafessah, M. Alsaleh, Comparative Survey of Indoor Positioning Technolo- (2019) 103911.
gies, Techniques, and Algorithms, in: Proc. of International Conference on [277] A. Bröring, S. Schmid, C.-K. Schindhelm, A. Khelil, S. Käbisch, D. Kramer, D.
Cyberworlds, IEEE, 2014, pp. 245–252. Le Phuoc, J. Mitic, D. Anicic, E. Teniente, Enabling IoT Ecosystems through
[249] Z. Kasmi, A. Norrdine, J. Blankenbach, Platform Architecture for Decentralized Platform Interoperability, IEEE software 34 (1) (2017) 54–61.
Positioning Systems, Sensors 17 (5) (2017) 957. [278] N. Pathak, A. Mukherjee, S. Misra, Reconfigure and Reuse: Interoperable
[250] Q. Wang, H. Luo, A. Men, F. Zhao, Y. Huang, An Infrastructure-Free Indoor Wearables for Healthcare IoT, in: Proc. of IEEE Conference on Computer
Localization Algorithm for Smartphones, Sensors 18 (10) (2018) 3317. Communications, IEEE, 2020, pp. 20–29.
[251] V. Zeimpekis, G.M. Giaglis, G. Lekakos, A Taxonomy of Indoor and Outdoor [279] M. Noura, M. Atiquzzaman, M. Gaedke, Interoperability in Internet of Things:
Positioning Techniques for Mobile Location Services, ACM SIGecom Exchanges Taxonomies and Open Challenges, Mob. Netw. Appl. 24 (3) (2019) 796–809.
3 (4) (2002) 19–27. [280] O. Bello, S. Zeadally, M. Badra, Network Layer Inter-Operation of Device-to-
[252] A.B. Adege, H.-P. Lin, G.B. Tarekegn, Y.Y. Munaye, L. Yen, An Indoor and Device Communication technologies in Internet of Things (IoT), Ad Hoc Netw.
Outdoor Positioning using a Hybrid of Support Vector Machine and Deep Neural 57 (2017) 52–62.
Network Algorithms, J. Sens. (2018). [281] S.D. Nagowah, H.B. Sta, B. Gobin-Rahimbux, An Overview of Semantic Inter-
[253] M. Yassin, E. Rachid, A Survey of Positioning Techniques and Location based operability Ontologies and Frameworks for IoT, in: Proc. of 6th International
Services in Wireless Networks, in: Proc. of IEEE International Conference on Conference on Enterprise Systems (ES), IEEE, 2018, pp. 82–89.
Signal Processing, Informatics, Communication and Energy Systems (SPICES), [282] P. Murdock, L. Bassbouss, M. Bauer, M.B. Alaya, R. Bhowmik, P. Brett, R.N.
IEEE, 2015, pp. 1–5. Chakraborty, M. Dadas, J. Davies, W. Diab, et al., Semantic interoperability for
[254] A.K. Paul, T. Sato, Localization in Wireless Sensor Networks: A Survey on the Web of Things, Médiathèque Télécom Sudparis & Institut Mines-Télécom
Algorithms, Measurement Techniques, Applications and Challenges, J. Sens. Business School, 2016, [Online] (Accessed March 3, 2021) https://tel.archives-
Actuator Netw. 6 (4) (2017) 24. ouvertes.fr/UNIV-PARIS-SACLAY/hal-01362033v1.
[255] G.M. Mendoza-Silva, J. Torres-Sospedra, J. Huerta, A Meta-Review of Indoor [283] O. Vermesan, Advancing IoT Platforms Interoperability, Gistrup: River
Positioning Systems, Sensors 19 (20) (2019) 4507. Publishers, 2018.
[256] P. Pascacio, S. Casteleyn, J. Torres-Sospedra, E.S. Lohan, J. Nurmi, Collaborative [284] N.-G. Wunsch, LexisNexis PatentSight, Largest Wearables Patent Owners
Indoor Positioning Systems: A Systematic Review, Sensors 21 (3) (2021) 1002. Worldwide from 2010 to 2019, Statista.
[257] P. Barsocchi, et al., A privacy-by-design architecture for indoor localization [285] J.E. Mück, B. Ünal, H. Butt, A.K. Yetisen, Market and Patent Analyses of
systems, in: Quality of Information and Communications Technology, Springer Wearables in Medicine, Trends Biotechnol. 37 (6) (2019) 563–566.
International Publishing, Cham, 2020, pp. 358–366. [286] S. Osman, R. Marks, Methods for Optimizing Positioning of Content on a Screen
[258] H. KhademSohi, S. Sharifian, K. Faez, Accuracy-Energy Optimized Location of a Head Mounted Display, 2019.
Estimation Method for Mobile Smartphones by GPS/INS Data Fusion, in: Proc. [287] T. Tamai, T. Oya, Information Processing Device, Information Processing
of 2nd International Conference of Signal Processing and Intelligent Systems Method, Information Processing Program, and Recording Medium, 2018.
(ICSPIS), IEEE, 2016, pp. 1–5. [288] C. Russey, RealWear Raises $80 million for Augmented Reality Headset for Con-
[259] M. Kjaergaard, Location-Based Services on Mobile Phones: Minimizing Power nected Workers, Wearabe Technologies Show Medica [Online] (Accessed March
Consumption, IEEE Pervasive Comput. 11 (1) (2010) 67–73. 3, 2021), 2019, https://www.wearable-technologies.com/2019/07/realwear-
[260] A. Ometov, S.V. Bezzateev, J. Kannisto, J. Harju, S. Andreev, Y. Koucheryavy, raises-80-million-for-augmented-reality-headset-for-connected-workers/.
Facilitating the Delegation of Use for Private Devices in the Era of the Internet [289] S. Draper, Kopin Unveils World’s First Voice-Controlled AR Headset,
of Wearable Things, IEEE Internet Things J. 4 (4) (2016) 843–854. Wearabe Technologies Show Medica [Online] (accessed march 3, 2021),
[261] S. He, S.-H.G. Chan, Wi-Fi Fingerprint-based Indoor Positioning: Recent 2018, https://www.wearable-technologies.com/2018/06/kopin-unveils-worlds-
Advances and Comparisons, IEEE Commun. Surv. Tutor. 18 (1) (2015) 466–490. first-voice-controlled-ar-headset/.
[262] D. Vasisht, S. Kumar, D. Katabi, Decimeter-Level Localization with a Single [290] S. Draper, HMDmd-Kopin Partnership to Develop Next Gen Wearable Displays
WiFi Access Point, in: Proc. fo 13th USENIX Symposium on Networked Systems for MedTech, Wearabe Technologies Show Medica [Online] (accessed march 3,
Design and Implementation, 2016, 165–178. 2021), 2019, https://www.wearable-technologies.com/2019/06/hmdmd-kopin-
[263] R. Bharadwaj, C. Parini, A. Alomainy, Experimental Investigation of 3-D Hu- partnership-to-develop-next-gen-wearable-displays-for-medtech/.
man Body Localization Using Wearable Ultra-wideband Antennas, IEEE Trans. [291] B. Heater, Microsoft and Trimble Made a Hard Hat with HoloLens Built-in,
Antennas and Propagation 63 (11) (2015) 5035–5044. 2019, [Online] (Accessed March 3, 2021) https://techcrunch.com/2019/02/24/
[264] T. Otim, L.E. Díez, A. Bahillo, P. Lopez-Iturri, F. Falcone, Effects of the Body microsoft-and-trimble-made-a-hardhat-with-hololens-built-in/.
Wearable Sensor Position on the UWB Localization Accuracy, Electronics 8 (11) [292] Zebra Technologies Corp. Zebra’S Total Wearable Solutions: Double-digit
(2019) 1351. Productivity Gains over Prior Generation Zebra Wearable Devices, 2019,
[265] D. Dardari, N. Decarli, A. Guerra, F. Guidi, The Future of Ultra-wideband [Online] (Accessed March 3, 2021) https://www.zebra.com/us/en/products/
Localization in RFID, in: Proc. of IEEE International Conference on RFID (RFID), mobile-computers/wearable-computers.html.
IEEE, 2016, pp. 1–7. [293] R. Socher, Vuzix Partners With Plessey to Develop Next-Gen AR
[266] R. Faragher, R. Harle, Location Fingerprinting with Bluetooth Low Energy Smartglasses Using MicroLED Technology, Wearabe Technologies Show
Beacons, IEEE J. Sel. Areas Commun. 33 (11) (2015) 2418–2428. Medica [Online] (accessed march 3, 2021), 2018, https://www.wearable-
[267] D. Giovanelli, E. Farella, RSSI or Time-of-Flight for Bluetooth Low Energy Based technologies.com/2018/06/vuzix-partners-with-plessey-to-develop-next-gen-
Localization? An Experimental Evaluation, in: Proc. of 11th IFIP Wireless and ar-smartglasses-using-microled-technology/.
Mobile Networking Conference (WMNC), IEEE, 2018, pp. 1–8. [294] C. Russey, Tesla Files Patent for Google Glass-Like AR System for
[268] A.K. Ng, L.K. Chan, H.Y. Lau, A Low-Cost Lighthouse-Based Virtual Reality Head Faster, Accurate Vehicle Production, 2018, [Online] (Accessed March 3,
Tracking System, in: Proc. of International Conference on 3D Immersion (IC3D), 2021) https://www.wearable-technologies.com/2018/12/tesla-files-patent-for-
IEEE, 2017, pp. 1–5. google-glass-like-ar-system-for-faster-accurate-vehicle-production/.
[269] S. Shao, A. Khreishah, J. Paez, PassiveRETRO: Enabling Completely Passive [295] Bose Corporation, Bose Frames Rondo: Audio Sunglasses, 2019, [Online]
Visible Light Localization for IoT Applications, in: Proc. of IEEE Conference on (Accessed March 3, 2021) https://www.bose.com/en_us/products/frames/bose-
Computer Communications (INFOCOM), IEEE, 2019, pp. 1540–1548. frames-rondo.html.
[270] Z. Wang, Z. Yang, Q. Huang, L. Yang, Q. Zhang, ALS-P: Light Weight Visible [296] M. Sawh, Nreal Light Smartglasses Provide Hope Mixed Reality for the Masses
Light Positioning via Ambient Light Sensor, in: Proc. of IEEE Conference on Will Be Here Soon, 2019, [Online] (Accessed March 3, 2021) https://www.
Computer Communications (INFOCOM), IEEE, 2019, pp. 1306–1314. wareable.com/ar/nreal-mixed-reality-glasses-review-7018.

32
A. Ometov et al. Computer Networks 193 (2021) 108074

[297] S. Mashal, G. Hoelzl, M. Kranz, Valkyrie Project: Flying Immersion in Virtual [328] RAIS: Real-time Analytics for the Internet of Sports, 2021, [Online] https:
Reality, in: Proc. of 11th International Conference on Virtual Worlds and Games //cordis.europa.eu/project/id/813162 (Accessed March 3, 2021).
for Serious Applications (VS-Games), IEEE, 2019, pp. 1–4. [329] Exceptional Sense of Touch for Robots and Prosthetics, 2021, [On-
[298] C. Russey, This Smart Vest Aims to Predict Your Heart-Attack Before it line] http://news.nus.edu.sg/research/exceptional-touch-robots-prosthetics (Ac-
Happens, Wearabe Technologies Show Medica [Online] (accessed march 3, cessed March 3, 2021).
2021), 2019, https://www.wearable-technologies.com/2019/01/ces-2019-this- [330] S. Lee, D. Sasaki, D. Kim, M. Mori, T. Yokota, H. Lee, S. Park, K. Fukuda,
smart-vest-aims-to-predict-your-heart-attack-before-it-happens/. M. Sekino, K. Matsuura, et al., Ultrasoft Electronics to Monitor Dynamically
[299] K. Matthews, Upright Vs. Lumo Lift: Which Posture Gadget Gets the Best Pulsing Cardiomyocytes, Nature Nanotechnol. 14 (2) (2019) 156–160.
Reviews?, 2017, [Online] (Accessed March 3, 2021) https://geardiary.com/ [331] Don’t Go Breaking My Heart, 2021, [Online] https://www.u-tokyo.ac.jp/focus/
2017/11/18/upright-vs-lumo-lift-posture-gadget-gets-best-reviews/. en/press/z0508_00021.html (Accessed March 3, 2021).
[300] J. Stables, Best Heart Rate Monitor 2019: HRM Watches and Chest Straps [332] B. Zhang, D.K. Sodickson, M.A. Cloos, A High-Impedance Detector-Array Glove
Compared, 2019, [Online] (Accessed March 3, 2021) https://www.wareable. for Magnetic Resonance Imaging of the Hand, Nat. Biomed. Eng. 2 (8) (2018)
com/fitness-trackers/best-heart-rate-monitor-and-watches. 570–577.
[301] J. Su, Apple Watch 4 Is Now An FDA Class 2 Medical Device: Detects [333] A.J. Bandodkar, P. Gutruf, J. Choi, K. Lee, Y. Sekine, J.T. Reeder, W.J.
Falls, Irregular Heart Rhythm, 2018, [Online] (Accessed March 3, Jeang, A.J. Aranyosi, S.P. Lee, J.B. Model, et al., Battery-Free, Skin-Interfaced
2021) https://www.forbes.com/sites/jeanbaptiste/2018/09/14/apple- Microfluidic/Electronic Systems for Simultaneous Electrochemical, Colorimetric,
watch-4-is-now-an-fda-class-2-medical-device-detects-falls-irregular-heart- and Volumetric Analysis of Sweat, Sci. Adv. 5 (1) (2019) eaav3294.
rhythm/#113ff69a2071. [334] Blood and Sweat Take Training App to the Next Level, 2021, [Online]
[302] S. Draper, FDA Approves First Cochlear Implants for Single-Sided Deaf- https://www.kth.se/en/aktuellt/nyheter/blod-och-svett-tar-traningsappen-till-
ness and Asymmetric Hearing Loss, 2019, [Online] (Accessed March nasta-niva-1.901859 (Accessed March 3, 2021).
3, 2021) https://www.wearable-technologies.com/2019/07/fda-approves-first- [335] H. Garcia-Molina, M. Joglekar, A. Marcus, A. Parameswaran, V. Verroios,
cochlear-implants-for-single-sided-deafness-and-asymmetric-hearing-loss/. Challenges in Data Crowdsourcing, IEEE Trans. Knowl. Data Eng. 28 (4) (2016)
[303] W.W. Lee, Y.J. Tan, H. Yao, S. Li, H.H. See, M. Hon, K.A. Ng, B. Xiong, J.S. 901–911.
Ho, B.C. Tee, A Neuro-Inspired Artificial Peripheral Nervous System for Scalable [336] M. Al-Soh, I.A. Zualkernan, An MQTT-based Context-Aware Wearable Assess-
Electronic Skins, Sci. Robot. 4 (32) (2019) eaax2198. ment Platform for Smart Watches, in: Proc. of 17th International Conference
[304] M.-K. Schöppl, EVER Pharma Launches Innovative, Patient Friendly Subcuta- on Advanced Learning Technologies (ICALT), IEEE, 2017, pp. 98–100.
neous Pump for Parkinson’s Disease, 2019, BusinessWire [Online] (Accessed [337] W. Feng, Z. Yan, H. Zhang, K. Zeng, Y. Xiao, Y.T. Hou, A.Survey.on. Security,
March 3, 2021) https://www.businesswire.com/news/home/20190630005001/ Privacy, A Survey on Security Privacy and Trust in Mobile Crowdsourcing, IEEE
en/Pharma-Launches-Innovative-Patient-Friendly-Subcutaneous-Pump. Internet Things J. 5 (4) (2017) 2971–2992.
[305] A. Gruentzig, Wearable Device: WO2015183470, 2017. [338] D. Ravi, C. Wong, B. Lo, G.-Z. Yang, A Deep Learning Approach to On-Node
[306] D.M. Kirsch, B. Chen, H. Vashi, System and Method for Vehicle Control Sensor Data Analytics for Mobile or Wearable Devices, IEEE J. Biomed. Health
Integrating Health Priority Alerts of Vehicle Occupants: US20160297359A1, Inf. 21 (1) (2016) 56–64.
2016. [339] A. Harris, H. True, Z. Hu, J. Cho, N. Fell, M. Sartipi, Fall Recognition
[307] D. Lumb, We Need Sony’s Wearable Air Conditioner Right Now, Techradar Using Wearable Technologies and Machine Learning Algorithms, in: Proc. of
[Online] (Accessed March 3, 2021), 2019, https://www.techradar.com/news/ International Conference on Big Data (Big Data), IEEE, 2016, pp. 3974–3976.
we-need-sonys-wearable-air-conditioner-right-now. [340] J. Ker, L. Wang, J. Rao, T. Lim, Deep Learning Applications in Medical Image
[308] Embr Labs, Your Personal Thermostat, 2020, [Online] (Accessed March 3, 2021) Analysis, IEEE Access 6 (2017) 9375–9389.
https://embrlabs.com. [341] O.D. Lara, M.A. Labrador, A Survey on Human Activity Recognition Using
[309] IMeasureU, The Best Wearable Technology for Athletes, 2019, [Online] Wearable Sensors, IEEE Commun. Surv. Tutor. 15 (3) (2012) 1192–1209.
(Accessed March 3, 2021) https://imeasureu.com/2019/08/29/best-wearable- [342] Q. Zhu, Z. Chen, Y.C. Soh, A Novel Semisupervised Deep Learning Method for
technology-athletes/. Human Activity Recognition, IEEE Trans. Ind. Inf. 15 (7) (2018) 3821–3830.
[310] Catapult, Wearable Technology. Products, 2020, [Online] (Accessed March 3, [343] Y. Ma, H. Ghasemzadeh, LabelForest: Non-Parametric Semi-Supervised Learning
2021) https://www.catapultsports.com/products. for Activity Recognition, in: Proc. of AAAI Conference on Artificial Intelligence,
[311] STATSports, STATSports Products, 2020, [Online] (Accessed March 3, 2021) 33, 2019, 4520–4527.
https://statsports.com/products/. [344] M. Lv, L. Chen, T. Chen, G. Chen, Bi-View Semi-Supervised Learning Based
[312] P.R. Bailey, Wearable Device for Fishing: US20150057968A1, 2019. Semantic Human Activity Recognition Using Accelerometers, IEEE Trans. Mob.
[313] Xsens, 2020, [Online] (Accessed March 3, 2021) https://www.xsens.com/. Comput. 17 (9) (2018) 1991–2001.
[314] Allied Business Intelligence, Inc, in: In-Vehicle Wearable Integration, Tech. Rep. [345] Q.W. Oung, M. Hariharan, H.L. Lee, S.N. Basah, M. Sarillee, C.H. Lee, Wearable
Technical Report, 2014. Multimodal Sensors for Evaluation of Patients with Parkinson Disease, in:
[315] K. Roberts, SMK Electronics and CAARESYS Partner to Develop Biometric Proc. of IEEE International Conference on Control System, Computing and
Passenger Monitoring Technology, BusinessWire, 2019, [Online] (Accessed Engineering (ICCSCE), IEEE, 2015, pp. 269–274.
March 3, 2021) https://www.zepp.com/en-us/golf/smart-coach/. [346] N. Lu, H. Jeong, Wireless Electronic Tattoos, Proc. of Structural Health
[316] S. Vhaduri, C. Poellabauer, Biometric-based Wearable User Authentication Monitoring..
During Sedentary and Non-Sedentary Periods, arXiv preprint arXiv:1811.07060. [347] S. Geisler, Body Art Gets Smart, Cell 174 (6) (2018) 1331.
[317] Y. Chen, G.A. Yuval, M.J. Sinclair, Mobile Device Security Using Wearable [348] A. Tastanova, M. Folcher, M. Müller, G. Came-nisch, A. Ponti, T. Horn, M.S.
Security Tokens, 2015, uS Patent 9, 135, 620. Tikhomirova, M. Fus-senegger, Synthetic Biology-Based Cellular Biomedical
[318] X. Huang, Z. Nie, H. Yi, Wearable Security Device, 2017, uS Patent App. Tattoo for Detection of Hypercalcemia Associated with Cancer, Sci. Transl. Med.
15/229, 393. 10 (437) (2018) 8562.
[319] D.M. Lerner, User-Wearable Secured Devices Provided Assuring Authentication [349] K.Y. Goud, C. Moonla, R.K. Mishra, C. Yu, R. Narayan, I. Litvan, J. Wang,
and Validation of Data Storage and Transmission, 2018, uS Patent 10, 154, Wearable Electrochemical Microneedle Sensor for Continuous Monitoring of
031. Levodopa: Toward Parkinson Management, ACS Sens. 4 (8) (2019) 2196–2204.
[320] N. Tanaka, V. Elgort, J. Danielson, A. Kalachev, J. Wong, U.R. Bhat, L. Copere, [350] K. Kaewkannate, S. Kim, A Comparison of Wearable Fitness Devices, BMC Public
M. Kataoka, et al., Smart Wearable Devices and Methods for Acquisition of Health 16 (1) (2016) 433.
Sensorial Information from Wearable Devices to Activate Functions in Other [351] L. Klus, E.S. Lohan, C. Granell, J. Nurmi, Crowdsourcing Solutions for Data
Devices, 2017, uS Patent App. 15/229, 393. Gathering from Wearables, in: Proc. of XXXV Finnish URSI Convention on Radio
[321] A. Parara, S. Sekka, Wearable Care Security Smart Watch Device, 2016, uS Science, 2019.
Patent App. 15/162, 121. [352] D.R. Witt, R.A. Kellogg, M.P. Snyder, J. Dunn, Windows Into Human Health
[322] B. Murison, Remote Drive and Wearable Adult Devices, 2018, uS Patent App. Through Wearables Data Analytics, Curr. Opinion Biomed. Eng. 9 (2019) 28–46.
15/764, 201. [353] B.R. Bloem, E.J. Henderson, E.R. Dorsey, M.S. Okun, N. Okubadejo, P. Chan,
[323] F. Sedic, Wearable Massager for Couples, 2016, uS Patent 9, 474, 681. J. Andrejack, S.K. Darweesh, M. Munneke, Integrated and Patient-centred
[324] A. Scheuring, A. Guldahl, L.I. Andersen, K.R. Ellekrans, Devices and Methods Management of Parkinson’s Disease: A Network Model for Reshaping Chronic
for Sexual Wellness, 2018, uS Patent 10, 022, 293. Neurological Care, The Lancet Neurology.
[325] Wearable Electroactive Fabrics Integrated in Garments, 2021, [Online] https: [354] P. Stanley-Marbell, A. Alaghi, M. Carbin, E. Darulova, L. Dolecek, A. Gerstlauer,
//cordis.europa.eu/project/id/825232 (Accessed March 3, 2021). G. Gillani, D. Jevdjic, T. Moreau, A. Daglis, et al., Exploiting Errors for
[326] Smart and Flexible Energy Supply Platform for Wearable Electronics, 2021, Efficiency: A Survey from Circuits to Algorithms, ACM Comput. Surv. (CSUR)
[Online] https://cordis.europa.eu/project/rcn/219473/factsheet/en (Accessed 53 3:51.
March 3, 2021). [355] G. Wilson, J. Bryan, K. Cranston, J. Kitzes, L. Nederbragt, T.K. Teal, Good
[327] Personalised Body Sensor Networks with Built-In Intelligence for Real-Time Enough Practices in Scientific Computing, PLoS Comput. Biol. 13 (6) (2017)
Risk Assessment and Coaching of Ageing workers, in All Types of Working e1005510.
and Living Environments, 2021, [Online] https://cordis.europa.eu/project/id/ [356] Q. Xu, T. Mytkowicz, N.S. Kim, Approximate Computing: A Survey, IEEE Des.
826304 (Accessed March 3, 2021). Test 33 (1) (2015) 8–22.

33
A. Ometov et al. Computer Networks 193 (2021) 108074

[357] S. Mittal, A Survey of Techniques for Approximate Computing, ACM Comput. [384] O. Badreddin, R. Castillo, L. Lessard, M. Albanese, Towards Improved Per-
Surv. 48 (4) (2016) 1–33. formance and Compliance in Healthcare Using Wearables and Bluetooth
[358] N.T. Ti, L.B. Le, Computation Offloading Leveraging Computing Resources from Technologies, in: Proc. of 25th Annual International Conference on Computer
Edge Cloud and Mobile Peers, in: Proc. of IEEE International Conference on Science and Software Engineering, 2015, 239–242.
Communications (ICC), IEEE, 2017, pp. 1–6. [385] D.R. Seshadri, R.T. Li, J.E. Voos, J.R. Rowbottom, C.M. Alfes, C.A. Zorman, C.K.
[359] F. Jalali, O.J. Smith, T. Lynar, F. Suits, Cognitive IoT Gateways: Automatic Task Drummond, Wearable Sensors for Monitoring the Physiological and Biochemical
Sharing and Switching Between Cloud and Edge/Fog Computing, in: Proc. of Profile of the Athlete, NPJ Digital Med. 2 (1) (2019) 1–16.
the SIGCOMM Posters and Demos, 2017, 121–123. [386] M. D’Arco, A. Renga, A. Ceccarelli, F. Brancati, A. Bondavalli, Enhancing
[360] F. Jalali, T. Lynar, O.J. Smith, R.R. Kolluri, C.V. Hardgrove, N. Waywood, F. Workers Safety in Worksites through Augmented GNSS Sensors, Measurement
Suits, Dynamic Edge Fabric Environment: Seamless and Automatic Switching 117 (2018) 144–152.
among Resources at the Edge of IoT Network and Cloud, in: Proc. of IEEE [387] L. Guanke, Y. Jianan, Y. Wen, Z. Yanling, Research on Seamless Positioning
International Conference on Edge Computing (EDGE), IEEE, 2019, pp. 77–86. of Power Wearables Based on GPS/UWB Combination, in: Proc. of IEEE Inter-
[361] A. Das, A. Leaf, C.A. Varela, S. Patterson, Skedulix: Hybrid Cloud Scheduling national Conference on Computer and Communication Engineering Technology
for Cost-Efficient Execution of Serverless Applications, arXiv preprint arXiv: (CCET), IEEE, 2018, pp. 123–127.
2006.03720. [388] H. Yoon, S.-H. Park, K.-T. Lee, Exploiting Ambient Light Sensor for Authen-
[362] O. Chukhno, N. Chukhno, G. Araniti, C. Campolo, A. Iera, A. Molinaro, Optimal tication on Wearable Devices, in: Proc. of 4th International Conference on
Placement of Social Digital Twins in Edge IoT Networks, Sensors 20 (21) (2020) Cyber Security, Cyber Warfare, and Digital Forensic (CyberSec), IEEE, 2015,
6181. pp. 95–100.
[363] C. Campolo, A. Iera, A. Molinaro, G. Ruggeri, MEC Support for 5G-V2X Use [389] S.K. Polu, S. Polu, NFC Based Smart Healthcare Services System, Int. J. Innov.
Cases through Docker Containers, in: Proc. of IEEE Wireless Communications Res. Sci. Technol. 5 (7) (2018) 45–48.
and Networking Conference (WCNC), IEEE, 2019, pp. 1–6. [390] C. Hill, Wearables–The Future of Biometric Technology? Biometric Technol.
[364] Q. Fan, N. Ansari, On Cost Aware Cloudlet Placement for Mobile Edge Today 2015 (8) (2015) 5–9.
Computing, IEEE/CAA J. Autom. Sinica 6 (4) (2019) 926–937. [391] H. Sellahewa, N. Ibrahim, S. Zeadally, Biometric Authentication for Wearables,
[365] S. Wang, J. Xu, N. Zhang, Y. Liu, A Survey on Service Migration in Mobile in: Biometric-Based Physical and Cybersecurity Systems, Springer, 2019, pp.
Edge Computing, IEEE Access 6 (2018) 23511–23528. 355–386.
[366] X. Yu, M. Guan, M. Liao, X. Fan, Pre-Migration of Vehicle to Network Services [392] M. Nguyen, M.O. Gani, V. Raychoudhury, Yours Truly? Survey on Accessibility
Based on Priority in Mobile Edge Computing, IEEE Access 7 (2018) 3722–3730. of Our Personal Data in the Connected World, in: Proc. of IEEE International
[367] L. Wang, L. Jiao, J. Li, J. Gedeon, M. Mühlhäuser, MOERA: Mobility-agnostic Conference on Pervasive Computing and Communications Workshops (PerCom
Online Resource Allocation for Edge Computing, IEEE Trans. Mob. Comput. 18 Workshops), IEEE, 2019, pp. 292–297.
[393] A. Ometov, P. Masek, L. Malina, R. Florea, J. Hosek, S. Andreev, J. Hajny,
(8) (2018) 1843–1856.
J. Niutanen, Y. Koucheryavy, Feasibility Characterization of Cryptographic
[368] S. Wang, R. Urgaonkar, M. Zafer, T. He, K. Chan, K.K. Leung, Dynamic
Primitives for Constrained (Wearable) IoT Devices, in: Proc. of International
Service Migration in Mobile Edge Computing Based on Markov Secision Process,
Conference on Pervasive Computing and Communication Workshops (PerCom
IEEE/ACM Trans. Netw. 27 (3) (2019) 1272–1288.
Workshops), IEEE, 2016, pp. 1–6.
[369] S. Padalkar, A. Korlekar, U. Pacharaney, Data Gathering in Wireless Sensor
[394] M. Wang, Z. Yan, A Survey on Security in D2D Communications, Mob. Netw.
Network for Energy Efficiency with and without Compressive Sensing at Sensor
Appl. 22 (2) (2017) 195–208.
Node, in: Proc. of International Conference on Communication and Signal
[395] R.I. Ansari, C. Chrysostomou, S.A. Hassan, M. Guizani, S. Mumtaz, J. Rodriguez,
Processing (ICCSP), IEEE, 2016, pp. 1356–1359.
J.J. Rodrigues, 5G D2D networks: Techniques, Challenges, and Future Prospects,
[370] A. Fragkiadakis, P. Charalampidis, E. Tragos, Adaptive Compressive Sensing for
IEEE Syst. J. 12 (4) (2017) 3970–3984.
Energy Efficient Smart Objects in IoT Applications, in: Proc. of 4th International
[396] S.R. Jondhale, M. Sharma, R. Maheswar, R. Shubair, A. Shelke, Comparison of
Conference on Wireless Communications, Vehicular Technology, Information
Neural Network Training Functions for RSSI Based Indoor Localization Problem
Theory and Aerospace & Electronic Systems (VITAE), IEEE, 2014, pp. 1–5.
in WSN, in: Handbook of Wireless Sensor Networks: Issues and Challenges in
[371] M. Liyanage, I. Ahmad, A.B. Abro, A. Gurtov, M. Ylianttila, A Comprehensive
Current Scenario’s, Springer, 2020, pp. 112–133.
Guide To 5G Security, John Wiley & Sons, 2018.
[397] X.B. Maxama, E.D. Markus, A Survey on Propagation Challenges in Wireless
[372] P. Datta, A.S. Namin, M. Chatterjee, A Survey of Privacy Concerns in Wearable
Communication Networks over Irregular Terrains, in: Proc. of Open Innovations
Devices, in: 2018 IEEE International Conference on Big Data (Big Data), IEEE,
Conference (OI), IEEE, 2018, pp. 79–86.
2018, pp. 4549–4553.
[398] K. Yu, K. Wen, Y. Li, S. Zhang, K. Zhang, A Novel NLOS Mitigation Algorithm
[373] E. Huang, J.-P. Onnela, Activity Classification Using Smartphone Gyroscope and
for UWB Localization in Harsh Indoor Environments, IEEE Trans. Veh. Technol.
Accelerometer Data, arXiv preprint arXiv:1903.12616.
68 (1) (2018) 686–699.
[374] L. Simpson, M.M. Maharaj, R.J. Mobbs, The Role of Wearables in Spinal Posture [399] W. Wei, X. Jin-Yu, Z. Zhong-Liang, A New NLOS Error Mitigation Algorithm in
Analysis: A Systematic Review, BMC Musculoskelet. Disorders 20 (1) (2019) 55. Location Estimation, IEEE Trans. Veh. Technol. 54 (6) (2005) 2048–2053.
[375] A.M. Sabatini, V. Genovese, A Sensor Fusion Method for Tracking Vertical [400] H. Miao, K. Yu, M.J. Juntti, Positioning for NLOS propagation: Algorithm
Velocity and Height Based on Inertial and Barometric Altimeter Measurements, derivations and Cramer–Rao Bounds, IEEE Trans. Veh. Technol. 56 (5) (2007)
Sensors 14 (8) (2014) 13324–13347. 2568–2580.
[376] L. Ozella, L. Gauvin, L. Carenzo, M. Quaggiotto, P.L. Ingrassia, M. Tizzoni, [401] R.M. Vaghefi, R.M. Buehrer, Cooperative Localization in NLOS Environ-
A. Panisson, D. Colombo, A. Sapienza, K. Kalimeri, et al., Wearable Proximity ments Using Semidefinite Programming, IEEE Commun. Lett. 19 (8) (2015)
Sensors for Monitoring a Mass Casualty Incident Exercise: Feasibility Study, J. 1382–1385.
Med. Internet Res. 21 (4) (2019) e12251. [402] V. Shubina, A. Ometov, S. Andreev, D. Niculescu, E.S. Lohan, Privacy Versus Lo-
[377] S. Bian, B. Zhou, H. Bello, P. Lukowicz, A Wearable Magnetic Field Based cation Accuracy in Opportunistic Wearable Networks, in: Proc. of International
Proximity Sensing System for Monitoring COVID-19 Social Distancing, in: Proc. Conference on Localization and GNSS (ICL-GNSS), IEEE, 2020, pp. 1–6.
of International Symposium on Wearable Computers, 2020, 22–26. [403] F. Potortì, A. Crivello, P. Barsocchi, F. Palumbo, Evaluation of Indoor Lo-
[378] J. Martín-Vaquero, A. Hernández Encinas, A. Queiruga-Dios, J. José Bullón, calisation Systems: Comments on the ISO/IEC 18305 Standard, in: Proc. of
A. Martínez-Nova, J. Torreblanca González, C. Bullón-Carbajo, Review on International Conference on Indoor Positioning and Indoor Navigation (IPIN),
Wearables to Monitor Foot Temperature in Diabetic Patients, Sensors 19 (4) 2018, 1–7.
(2019) 776. [404] ISO, Information Technology – Real Time Localization System – Test and
[379] H. Han, H. Jang, S.W. Yoon, Novel Wearable Monitoring System of Forward Evaluation of Localization and Tracking Systems, Standard, International
Head Posture Assisted by Magnet-Magnetometer Pair and Machine Learning, Organization for Standardization, Geneva, CH, 2016.
IEEE Sens. J., 2019. [405] F. Furfari, A. Crivello, P. Barsocchi, F. Palumbo, F. Potortí, What is next for
[380] W. Zhuang, Y. Chen, J. Su, B. Wang, C. Gao, Design of human activity Indoor Localisation? Taxonomy, Protocols, and Patterns for Advanced Location
recognition algorithms based on a single wearable IMU sensor, Int. J. Sens. Based Services, in: Proc. of International Conference on Indoor Positioning and
Netw. 30 (3) (2019) 193–206. Indoor Navigation (IPIN), 2019, 1–8.
[381] A. Mihoub, G. Lefebvre, Wearables and Social Signal Processing for Smarter [406] A. Ometov, P. Masek, J. Urama, J. Hosek, S. Andreev, Y. Koucheryavy,
Public Presentations, ACM Trans. Interactive Intell. Syst. (TiiS) 9 (2–3) (2019) Implementing Secure Network-Assisted D2D Framework in Live 3GPP LTE
1–24. Deployment, in: Proc. of IEEE International Conference on Communications
[382] V. Dissanayake, D.S. Elvitigala, H. Zhang, C. Weerasinghe, S. Nanayakkara, Workshops (ICC), IEEE, 2016, pp. 749–754.
CompRate: Power Efficient Heart Rate and Heart Rate Variability Monitoring on [407] K. Wang, J. Alonso-Zarate, M. Dohler, Energy-Efficiency of LTE for Small
Smart Wearables, in: Proc. of 25th ACM Symposium on Virtual Reality Software Data Machine-to-Machine Communications, in: Proc. of IEEE International
and Technology, 2019, 1–8. Conference on Communications (ICC), IEEE, 2013, pp. 4120–4124.
[383] C.L. Karmen, M.A. Reisfeld, M.K. McIntyre, R. Timmermans, W. Frishman, The [408] G. Rigazzi, N.K. Pratas, P. Popovski, R. Fantacci, Aggregation and Trunking of
Clinical Value of Heart Rate Monitoring Using an Apple Watch, Cardiol. Rev. M2M Traffic via D2D Connections, in: Proc. of IEEE International Conference
27 (2) (2019) 60–62. on Communications (ICC), IEEE, 2015, pp. 2973–2978.

34
A. Ometov et al. Computer Networks 193 (2021) 108074

[409] Z. Zhang, C. Wang, H. Yu, M. Wang, S. Sun, Power Optimization Assisted [431] K.M. Rashid, J. Louis, Times-Series Data Augmentation and Deep Learning
Interference Management for D2D Communications in mmWave Networks, IEEE for Construction Equipment Activity Recognition, Adv. Eng. Inform. 42 (2019)
Access 6 (2018) 50674–50682. 100944.
[410] A. Gupta, R.K. Jha, A Survey of 5G Network: Architecture and Emerging [432] J.V. Jacobs, L.J. Hettinger, Y.-H. Huang, S. Jeffries, M.F. Lesch, L.A. Simmons,
Technologies, IEEE Access 3 (2015) 1206–1232. S.K. Verma, J.L. Willetts, Employee Acceptance of Wearable Technology in the
[411] Y. He, J. Ren, G. Yu, Y. Cai, D2D Communications Meet Mobile Edge Computing Workplace, Applied Ergon. 78 (2019) 148–156.
for Enhanced Computation Capacity in Cellular Networks, IEEE Trans. Wireless [433] K.W. Ching, M.M. Singh, Wearable Technology Devices Security and Privacy
Commun. 18 (3) (2019) 1750–1763. Vulnerability Analysis, Int. J. Netw. Secur. Appl. 8 (3) (2016) 19–30.
[412] G. Araniti, A. Orsino, L. Militano, G. Putrino, S. Andreev, Y. Koucheryavy, A. [434] D.F.S. Santos, H.O. Almeida, A. Perkusich, A Personal Connected Health System
Iera, Novel D2D-based relaying method for multicast services over 3GPP LTE-A
for the Internet of Things Based on the Constrained Application Protocol,
systems, in: Proc. of IEEE International Symposium on Broadband Multimedia
Computers and Electrical Engineering 44.
Systems and Broadcasting (BMSB), IEEE, 2017, pp. 1–5.
[435] A. Martins, D. Santos, A. Perkusich, H. Almeida, IEEE 11073 and Connected
[413] L. Militano, M. Condoluci, G. Araniti, A. Molinaro, A. Iera, G.-M. Muntean,
Health: Preparing Personal Health Devices for the Internet, in: Proc. of IEEE
Single Frequency-based Device-to-Device-enhanced Video Delivery for Evolved
International Conference on Consumer Electronics (ICCE), 2014.
Multimedia Broadcast and Multicast Services, IEEE Trans. Broadcast. 61 (2)
(2015) 263–278. [436] iTouch Wearables, Customer support, 2017, [Online] (Accessed March 3, 2021)
[414] M. Kritzler, M. Bäckman, A. Tenfält, F. Michahelles, Wearable Technology as https://itouchwearables.com/pages/support.
a Solution for Workplace Safety, in: Proc. of 14th International Conference on [437] ATT, Learn About and Get Support for Wearables, 2017, [Online] (Accessed
Mobile and Ubiquitous Multimedia, 2015, 213–217. March 3, 2021) https://www.att.com/support/article/wireless/KM1008691/.
[415] E. Svertoka, A. Rusu-Casandra, I. Marghescu, State-of-the-Art of Industrial [438] R. Snader, R. Kravets, A.F. Harris III, Cryptocop: Lightweight, Energy-Efficient
Wearables: A Systematic Review, in: Proc. of 13th International Conference Encryption and Privacy for Wearable Devices, in: Proc. of Workshop on
on Communications (COMM), IEEE, 2020, pp. 411–415. Wearable Systems and Applications, 2016, 7–12.
[416] S. Kim, Y. Sharif, I. Nasim, Human Electromagnetic Field Exposure in Wearable [439] A. Ometov, Y. Bardinova, A. Afanasyeva, P. Masek, K. Zhidanov, S. Vanurin, M.
Communications: A Review, arXiv preprint arXiv:1912.05282. Sayfullin, V. Shubina, M. Komarov, S. Bezzateev, An Overview on Blockchain
[417] I. Nasim, S. Kim, Human EMF Exposure in Wearable Networks for Internet for Smartphones: State-of-the-Art, Consensus, Implementation, Challenges and
of Battlefield Things, in: Proc. of IEEE Military Communications Conference Future Trends, IEEE Access 8 (2020) 103994–104015.
(MILCOM), IEEE, 2019, pp. 1–6. [440] O. Arias, J. Wurm, K. Hoang, Y. Jin, Privacy and Security in Internet of Things
[418] G. Kiruthiga, A. Sharmila, P. Mahalakshmi, M. Muruganandam, Power Optimi- and Wearable Devices, IEEE Trans. Multi-Scale Comput. Syst. 1 (2) (2015)
sation for Wearable Heart Rate Measurement Device with Wireless Charging, 99–109.
J. Med. Eng. Technol. 41 (4) (2017) 288–297. [441] R. Fallahzadeh, H. Ghasemzadeh, Trading Off Power Consumption and Pre-
[419] M.A. Razzaque, S. Dobson, Energy-Efficient Sensing in Wireless Sensor Networks diction Performance in Wearable Motion Sensors: An Optimal and Real-Time
Using Compressed Sensing, Sensors 14 (2) (2014) 2822–2859.
Approach, ACM Trans. Des. Autom. Electron. Syst. (TODAES) 23 (5) (2018)
[420] Y. Zhang, Y. Zhao, J. Ren, W. Weng, H. Peng, Advances in Wearable
1–23.
Fiber-Shaped Lithium-Ion Batteries, Adv. Mater. 28 (22) (2016) 4524–4531.
[442] H. Yin, N.K. Jha, A Health Decision Support System for Disease Diagnosis Based
[421] M. Magno, D. Boyle, Wearable Energy Harvesting: From Body to Battery, in:
on Wearable Medical Sensors and Machine Learning Ensembles, IEEE Trans.
Proc. of 12th International Conference on Design & Technology of Integrated
Multi-Scale Comput. Syst. 3 (4) (2017) 228–241.
Systems in Nanoscale Era (DTIS), IEEE, 2017, pp. 1–6.
[443] L.C. Kourtis, O.B. Regele, J.M. Wright, G.B. Jones, Digital biomarkers for
[422] M. Magno, D. Kneubühler, P. Mayer, L. Benini, Micro Kinetic Energy Harvesting
for Autonomous Wearable Devices, in: Proc. of International Symposium on Alzheimer’s Disease: The Mobile/Wearable Devices Opportunity, NPJ Digital
Power Electronics, Electrical Drives, Automation and Motion (SPEEDAM), IEEE, Med. 2 (1) (2019) 1–9.
2018, pp. 105–110. [444] Z. Qin, Y. Zhang, S. Meng, Z. Qin, K.-K.R. Choo, Imaging and Fusing Time
[423] P. Jokic, M. Magno, Powering Smart Wearable Systems with Flexible Solar Series for Wearable Sensor-based Human Activity Recognition, Inf. Fusion 53
Energy Harvesting, in: Proc. of International Symposium on Circuits and (2020) 80–87.
Systems (ISCAS), IEEE, 2017, pp. 1–4. [445] C. Bormann, A.P. Castellani, Z. Shelby, COAP: An Application Protocol for
[424] L.V. Thekkekara, M. Gu, Large-Scale Waterproof and Stretchable Textile- Billions of Tiny Internet Nodes, IEEE Internet Comput. (2) (2012) 62–67.
Integrated Laser-Printed Graphene Energy Storages, Sci. Rep. 9 (1) (2019) [446] R.T. Fielding, Architectural Styles and the Design of Network-Based Software
1–7. Architectures (Ph.D. thesis), University of California, Irvine, 2000.
[425] K.N. Bocan, E. Sejdić, Adaptive Transcutaneous Power Transfer to Implantable [447] M. Jeronimo, J. Weast, UPnP Design By Example: A Software Developer’s Guide
Devices: A State of the Art Review, Sensors 16 (3) (2016) 393. to Universal Plug and Play, Intel Press, 2003.
[426] K. Agarwal, R. Jegadeesan, Y.-X. Guo, N.V. Thakor, Wireless Power Transfer [448] H. Khattak, M. Ruta, E. Di Sciascio, CoAP-based Healthcare Sensor Networks: A
Strategies for Implantable Bioelectronics, IEEE Rev. Biomed. Eng. 10 (2017) Survey, in: Proc. of 11th International Bhurban Conference on Applied Sciences
136–161. and Technology (IBCAST), 2014.
[427] T. Ghomian, S. Mehraeen, Survey of Energy Scavenging for Wearable and [449] M.R. Palattella, M. Dohler, A. Grieco, G. Rizzo, J. Torsner, T. Engel, L. Ladid,
Implantable Devices, Energy 178 (2019) 33–49. Internet of Things in the 5G Era: Enablers, Architecture, and Business Models,
[428] J. Portilla, G. Mujica, J.-S. Lee, T. Riesgo, The Extreme Edge at the Bottom of IEEE J. Sel. Areas Commun. 34 (3) (2016) 510–527.
the Internet of Things: A Review, IEEE Sens. J. 19 (9) (2019) 3179–3190. [450] Fischer Connectors, Internet of Things (IoT) and Wearables, 2020, [On-
[429] G. Aloi, G. Caliciuri, G. Fortino, R. Gravina, P. Pace, W. Russo, C. Savaglio, line] (Accessed March 3, 2021) https://www.fischerconnectors.com/global/en/
Enabling IoT Interoperability through Opportunistic Smartphone-based Mobile applications/internet-things-iot-wearables.
Gateways, J. Netw. Comput. Appl. 81 (2017) 74–84. [451] D. Miorandi, S. Sicari, F. De Pellegrini, I. Chlamtac, Internet of Things: Vision,
[430] X. Sun, N. Ansari, Avaptive Avatar Handoff in the Cloudlet Network, IEEE Applications and Research Challenges, Ad Hoc Networks 10 (7).
Trans. Cloud Comput. 7 (3) (2017) 664–676. [452] J. Kim, J. Lee, J. Kim, J. Yun, M2 Service Platforms: Survey, Issues, and
Enabling Technologies, IEEE Communications Surveys Tutorials 16 (1).

35
A. Ometov et al. Computer Networks 193 (2021) 108074

Aleksandr Ometov is a Postdoctoral Research Fellow at Darwin Quezada Gaibor is an Early Stage Researcher and
Tampere University (TAU), Finland. He is currently working Ph.D. student at Universitat Jaume I (Spain) and Tampere
on H2020 MCSA ITN/EJD A-WEAR project. He received University (Finland). He received his bachelor’s degree
D.Sc. (Tech) in Telecommunications and M.Sc. is Informa- in Mechatronic Engineering from Universidad Tecnológica
tion Technology from Tampere University of Technology América, Ecuador, 2013 and his Master’s Degree in Ra-
(TUT), Finland in 2018 and 2016, correspondingly. He dioengineering – GNSS receivers: Hardware and Software
also holds his Specialist degree in Information Security from Samara National Research University, Russia, 2017.
from Saint Petersburg State University of Aerospace Instru- His main interests are VoIP, Cloud Computing, Networking,
mentation (SUAI), Russia, 2013. His research interests are Servers, and open-source software.
wireless communications, information security, blockchain
technology and wearable applications.
Nadezhda Chukhno is an Early Stage Researcher at A-
WEAR and Doctoral Researcher at Mediterranea University
Viktoriia Shubina is a Doctoral student at Tampere Univer- of Reggio Calabria, Italy and Jaume I University, Spain.
sity (TAU), and an Early Stage Researcher at H2020 MCSA She graduated from RUDN University, Russia, and re-
ITN/EJD A-WEAR project. She received her Double M.Sc. in ceived her B.Sc. in Business Informatics (2017) and M.Sc.
Engineering from the University of Applied Sciences Tech- in Fundamental Informatics and Information technologies
nikum Wien, Austria, and M.Sc. in Business Informatics from (2019). Her current research activity mainly focuses on
National Research University Higher School of Economics, wireless communications, 5G networks, D2D, and wearable
Russia in 2019. Topics, which interest her most are location technologies.
privacy, privacy-preserving mechanisms, indoor positioning,
and wearable technologies.
Olga Chukhno is an Early Stage Researcher within H2020
MCSA ITN/EJD A-WEAR project and a Ph.D. student at
Lucie Klus is the Early Stage Researcher at Tampere Uni- Mediterranea University of Reggio Calabria, Italy and Tam-
versity, Finland and Jaume I University, Spain as a part pere University, Finland. She received M.Sc. (2019) in
of H2020 MCSA ITN/EJD A-WEAR project since Septem- Fundamental Informatics and Information Technologies and
ber 2019. She received her M.Sc. and B.Sc. degrees from B.Sc. (2017) in Business Informatics from RUDN Univer-
Electronics and Communications at Brno University of Tech- sity, Russia. Her current research interests include wireless
nology in 2019 and 2017. Her research interests include communications, social networking, edge computing, and
modern approaches in wireless communications, data pro- wearable applications.
cessing and analytics, crowdsourcing methods and machine
learning techniques.
Asad Ali is an Early Stage Researcher in the H2020
MCSA ITN/EJD A-WEAR project and a Doctoral Researcher
Justyna Skibińska is the Early Stage Researcher and Ph.D. at Tampere University, Finland, and Brno University of
student at Brno University of Technology, Czech Republic Technology, Czech Republic. He pursued his B.Sc. in Elec-
and Tampere University, Finland as a part of H2020 MCSA trical Engineering with majors in Telecommunication from
ITN/EJD A-WEAR project. She obtained her M.Sc. and B.Sc. COMSATS University Islamabad, Pakistan (2015), and M.Sc.
degrees from Biomedical Engineering with specialization in in Electrical Engineering from the University of Rostock,
Computer Science and Electronics in Medicine at University Germany (2019). His research interests include, but are not
of Science and Technology in Kraków, Poland in 2018 and limited to, mmWave communication, wearable technology,
2017 respectively. Her research focus is on machine learning and 5G network.
techniques and wearable technologies.

Asma Channa is an Early Stage Researcher in the H2020

Salwa Saafi is a joint Ph.D. student at the Department MCSA ITN/EJD A-WEAR project and a Doctoral Researcher
of Telecommunications at Brno University of Technology, at University Politehnica of Bucharest, and Mediterranea
Czech Republic and the Unit of Electrical Engineering at University of Reggio Calabria, Italy. In 2015, she graduated
Tampere University, Finland. She received her engineer- from Mehran University of Engineering and Technology
ing degree (2017) in telecommunications from the Higher Jamshoro, Pakistan with B.Sc. degree in Electronic Engi-
School of Communication of Tunis, Tunisia. Her research neering. She holds her M.Sc. degree in Electronic Systems
interests include cellular radio access technologies, future Engineering from the same university (2018). Her research
wireless architectures, and wearable applications. interests include, biomedical signal processing, wearables,
eHealth and artificial intelligence.

Ekaterina Svertoka is a Ph.D. candidate at Politehnica

Pavel Pascacio is currently an Early Stage Researcher of the
University of Bucharest, Romania as part of H2020 ITN/EJD
H2020 MCSA ITN/EJD A-WEAR project at the Universitat
A-WEAR project. She received her M.Sc. in Radioengi-
Jaume I (Spain) and Tampere University (Finland) respec-
neering from Saint Petersburg Electrotechnical University
tively. He received his M.Sc. in Automation and Control
(LETI), Russia in 2018. Her research interests are wireless
Engineering from Politecnico di Milano, Italy, in 2019 and a
communications and wearable technology for work safety.
B.Sc. in Electronics Engineering from Instituto Tecnológico
de Tuxtla Gutiérrez, México, in 2009. His main research
interests are in the areas of infrastructure-less indoor loca-
tion & positioning, machine learning, collaborative indoor
positioning systems, peer-to-peer communication and mesh
networks.
Waleed Bin Qaim is an Early Stage Researcher (ESR) in the
Laura Flueratoru is currently pursuing her joint Ph.D. H2020 MSCA ITN/EJD A-WEAR project. He is a double de-
degree at University Politehnica of Bucharest and Tampere gree Ph.D. student affiliated with Tampere University (TAU),
University as part of the H2020 MSCA ITN/EJD A-WEAR Finland and University Mediterranea of Reggio Calabria
project. She received her M.Sc. degree in Electrical En- (UNIRC), Italy. He received his M.Sc. in Computer Science
gineering from ETH Zürich in 2019 and, prior to this, and Engineering from Koc University, Turkey in 2018.
her B.Eng. degree (with highest honors) in Electronics He completed his B.Sc. in Telecommunication Engineering
and Telecommunications from University Politehnica of from the National University of Computer and Emerging
Bucharest in 2017. Her main research interests are in wire- Sciences (NUCES), Pakistan in 2011. His research interests
less communications, localization systems, ultra-wideband include Wireless Communication, IoT, Wearable Networks,
communications, and embedded systems. and Distributed Computing Systems.

36
A. Ometov et al. Computer Networks 193 (2021) 108074

Raúl Casanova-Marqués is a joint-Ph.D. student at Brno Giuseppe Araniti (Senior Member, IEEE) received the Lau-
University of Technology (BUT) and Universitat Jaume I rea degree and the Ph.D. degree in electronic engineering
(UJI), working as an Early Stage Researcher (ESR) within from the University Mediterranea of Reggio Calabria, Italy,
the A-WEAR project. He graduated (2017) in Computer in 2000 and 2004, respectively. He is currently an Assis-
Engineering at Universitat Rovira i Virgili (URV) in Spain tant Professor of telecommunications with the University
and he received a M.Sc. (2018) in Information and Com- Mediterranea of Reggio Calabria. His major area of re-
munication Technology Security from Universitat Oberta de search is on 5G/6G networks and it includes personal
Catalunya (UOC), and a M.Sc. (2018) Indra Master’s in communications, enhanced wireless and satellite systems,
Cybersecurity from U-tad. He is currently involved as a traffic and radio resource management, multicast and broad-
researcher in several scientific projects focused on novel cast services, device-to-device (D2D), and machine-type
cryptographic protocols for attribute-based authentication, communications (M2M/MTC).
privacy protection mechanisms of users in electronic systems
and secure algorithms on wearable devices.

Radim Burget is an Assoc. Professor at Brno University of

Sylvia Holcer is currently an Early Stage Researcher and
Technology and is heading Signal processing program at SIX
a Ph.D. student at Universitat Jaume I (UJI) in Castellón,
Research Centre. He has been involved in research of arti-
Spain with a secondment at Brno University of Technology
ficial intelligence for many years and in plenty of research
(BUT) in Brno, Czech Republic. She is focusing her work on
projects which include projects funded on European level,
location privacy in wearables. She majored in Geoinforma-
national level or privately funded projects. Companies he
tion for her M.Sc. and B.Sc. at Adam Mickiewicz University
is cooperating with include Honeywell, Mitsubishi Electric,
in Poznań, Poland. Her primary research interests are new
Rapidminer, Konica-Minolta and others.
technologies, machine learning, and spatial analysis.

Joaquín Torres-Sospedra is CEO at UBIK Geospatial So-

lutions. He has a Ph.D. since 2011 about Ensembles of Jiri Hosek (Senior Member, IEEE) received the M.Sc. and
Neural Networks and Machine Learning from Universitat Ph.D. degrees in electrical engineering from the Faculty
Jaume I. He has authored more than 120 articles in journals of Electrical Engineering and Communication, Brno Univer-
and conference proceedings. His current research interests sity of Technology (BUT), Czech Republic, in 2007 and
include indoor positioning solutions based on Wi-Fi & BLE, 2011, respectively. He is currently an Associate Professor
Machine Learning and Evaluation. Dr. Torres-Sospedra is the and the Deputy Vice-Head for Research and Development
chair of the Smartphone-based track of IPIN Competition and International Relations at the Department of Telecom-
since 2015. He is also the chair of the IPIN International munications, BUT. He is also coordinating the WISLAB
Standards Committee since 2018. Research Group, where he deals mostly with industry-
oriented projects in the area of future mobile networks, the
Internet of Things, and home automation services. He has
Sven Casteleyn received his master and Ph.D. in Computer (co)authored more than 130 research works on networking
Science from the Vrije Universiteit Brussel, Belgium. He is technologies, wireless communications, quality of service,
currently an associate professor at the Universidad Jaime quality of experience, and IoT applications.
I, CastellÃşn, Spain, and previously held a Ramon y Cajal
(Spain) and a Marie Curie (Europe) individual post-doctoral
fellowship. Sven supervised 6 Ph.D. and over 50 master
theses. He published over 100 scientific articles in the
Elena Simona Lohan is a Professor at Tampere University
broad fields of the Web, mobile computing, geographical
(TAU), Finland. She received an M.Sc. degree in electri-
information science and technology, and their application
cal engineering from Polytechnics University of Bucharest,
fields.
Romania, in 1997, a DEA degree (French equivalent of
master) in econometrics at Ecole Polytechnique, Paris,
Giuseppe Ruggeri received the master degree in electronics France, in 1998, and a Ph.D. degree in telecommunica-
engineering in 1998. In 2002 he received the Ph.D. in tions from Tampere University of Technology in 2003. She
electronics, computer science and telecommunications engi- is now a professor at the Electrical Engineering unit at
neering. He is currently assistant professor at the University Tampere University, Finland and the coordinator of the
Mediterranea of Reggio Calabria. His current interests in- MSCA EU A-WEAR network. Her current research interests
clude self organizing networks, Internet of Things, Social include wireless location techniques, wearable computing,
Internet of Things. and privacy-aware positioning solutions.

37