ArticlePDF Available

Toward Embodied Intelligence: Smart Things on the Rise

Authors:
Article

Toward Embodied Intelligence: Smart Things on the Rise

Abstract

Embodied intelligence will unleash the potential of artificial intelligence by providing it with a body and impact every area of life. The success of the adoption of the embodied intelligence market depends on solving the technical, legal, economic, and social issues it brings with it.
Towards Embodied
Intelligence – Smart Things on
the Rise
A. Br¨
oring
Siemens AG
C. Niedermeier
Siemens AG
I. Olaru
Deutsches Dialog Institut
U. Sch¨
opp
fortiss GmbH
K. Telschig
Siemens AG
M. Villnow
Siemens AG
Abstract—Embodied intelligence will unleash the potential of AI by providing it with a body and
will impact every area of life. It will help us overcome major societal challenges of our time, such
as climate change and population growth. The success of EI market adoption depends on
solving the technical, legal, economic, and social issues it brings with it.
WIT HIN THE N EXT TWO DECA DES a new
class of cyber-physical systems will disrupt al-
most every domain: Artificial embodied intel-
ligence (EI). The key characteristics of these
cyber-physical systems are that they can decide
on concrete actions based on vague challenges
and execute these actions in open, unstructured
environments. Our paper describes this ongoing
development and its disruptive potential for se-
lected application areas. Before diving deeper
into the definition of EI, let us look at the big
challenges of our time, i.e., why we need EI and
its impact areas. Roland Busch, Deputy CEO of
Siemens AG, identifies five megatrends [1], which
are driving the transformation of markets:
Demographic Changes. The global population
is expected to continue growing (8.5 billion
people in 2030, 9.7 billion in 2050) [2] and
aging (e.g., in 2050, the number of people
Computer Published by the IEEE Computer Society © 2021 IEEE 1
This is the accepted draft of this paper.
The final version of this paper can be found at:
https://ieeexplore.ieee.org/document/9473215
The citation is:
A. Broering, C. Niedermeier, I. Olaru, U. Schopp, K. Telschig and M. Villnow, "Toward Embodied Intelligence: Smart
Things on the Rise," in Computer, vol. 54, no. 7, pp. 57-68, July 2021, doi: 10.1109/MC.2021.3074749.
Department Head
above 65 will be about twice as high as the
number of children under 5). [3]
Urbanization. The world is expected to have
43 megacities in 2030 (2018: 33 megacities),
which is putting huge stress on infrastructures.
[4]
Climate change, pollution, and scarcity of re-
sources. Humans are consuming more than
earth can reproduce. By 2030, global energy
demand will be 17% higher than in 2015 and
by 2050, 68% of the population is expected to
live in regions plagued by water stress or water
scarcity. [5]
Glocalization. While globalization has been
shaping markets and supply chains for decades,
more recently, political forces in various coun-
tries countered this concept towards more local
emphasis (e.g., Brexit). Also, the Covid-19
pandemic has shown that overlong interna-
tional supply chains are not resilient and value
creation has to be redistributed.
Digitalization. The fifth megatrend describes
the combination of the physical with the digital
world through information and communication
technology. This has broad and rapid impact
on the transformation of many business models
and is increasingly disrupting existing markets.
All five megatrends are reshaping economies
and societies globally and are forcing technolog-
ical developments. They pose challenges for ed-
ucation, health and work systems, energy supply,
as well as food production and transportation,
to name just some of the more visible impacted
areas. For companies, digitalization is posing
new challenges as well as new opportunities.
Significant opportunities are the possibility to
optimize processes, to reach many users via dig-
ital platforms, and the customization of product-
service systems (PSS) to comprehensively fulfill
a certain need (e.g., car-sharing). Rethinking the
product-service portfolio and adapting to changes
in their ecosystems are key elements of the digital
transformation.
With these megatrends in mind, EI – as the
next big thing in digitalization – is about to enter
the stage. So, what is EI? The amalgamation of
the physical and the digital world leads to an
increase in the intelligence of physical devices
towards smarter and more autonomous behavior.
An artificial Embodied Intelligence (EI) is a thing
with a tight integration of physical capabilities
such as sensing and actuating, perception and in-
teraction capabilities, and AI algorithms enabling
decision making and learning (see Figure 1). The
emerging autonomous car is a good example:
It has a multitude of sensors and actuators, the
capability to perform perception, learning and
reasoning based on actual and previously acquired
sensor data, and the ability to autonomously
perform driving related actions as well as to
interact with other cars and the user. In addition
to these capabilities, an EI thing must also have
a goal that provides guidance for its autonomous
behavior. Intelligent things (e.g. autonomous cars,
autonomous trucks) are self-contained forms of
EI: they have both a “body” with sensors and
actuators and a “brain” tightly interconnected
with that body. Furthermore, we envision so-
called Collaborative Intelligent Systems (CIS)
consisting of at least two Intelligent Things that
collaborate with each other and thus likely form
a higher level of EI (e.g., smart factory, smart
home, smart city) with a common goal. While
today such systems are engineered in advance,
we expect that future CIS will be composed
autonomously (Self-organizing CIS). Thus, we
extend the definition of Cangelosi et al. [6],
who define EI only for agents, namely individual
things such as cars.
The concept of EI is closely related to the
digital twin concept. However, while the digital
twin is always a digital representation of a phys-
ical object, an EI is a cyber-physical object (it
might use a digital twin, though). Moreover, EI
combines technologies of Artificial Intelligence
with those of the Internet of Things.
Several technological developments leading
to the emergence of EI are currently underway.
Thus, it is important to understand the relevance
of megatrends for EI and their impact on markets
and application.
EMBODIED INTELLIGENCE FOR
HEALTHY LIVING
As an effect of the megatrend demographic
change, elder and long-term care is rapidly be-
coming one of the most significant healthcare
challenges. According to Forbes [7], between
2015 and 2030, the number of people in the world
2Computer
Figure 1. Properties and types of artificial Embodied
Intelligence (EI)
aged 60 years or over is expected to grow by 56%,
from just over 900 million to nearly 1.5 billion.
As older people are particularly vulnerable for
diseases, a healthier lifestyle preventing diseases
as well as early diagnosis and treatment of dis-
eases is crucial to contain the cost of healthcare
systems. Embodied intelligence will help to cope
with this challenge in several ways: Smart Wear-
ables will not only increasingly support people in
tracking and improving their fitness and engaging
in social interactions but will also help to prevent
more serious diseases as well as provide support
for efficient diagnosis and treatment of diseases
as part of a larger effort called Smart Healthcare.
Everyday life robots will increasingly be able
to take over normal household work and even
provide basic care to disabled and frail persons.
As a result of these developments, elderly and
disabled people can continue to live in their
homes rather than having to move to an expensive
nursing home.
Wearable technologies that have been uti-
lized for many years by the military and health-
care industries are increasingly emerging in main-
stream consumer products due to ongoing ad-
vances in key enabling technologies such as low-
cost sensors, wireless connectivity, active ma-
terials, and energy. Facilitated by the miniatur-
ization of components and driven by the abil-
ity to interconnect with key modern trends of
healthcare, fitness, messaging and socialization,
a growing number of wearable devices such as
smartwatches, fitness trackers, smart glasses or
hearables appear on the market. The wearable
technology market will grow at a CAGR of 15.9%
between 2020 to 2027 [8]. By the end of 2027,
wearable technologies will have a market worth
of $104 billion. The COVID-19 crisis is expected
to significantly drive the further development of
the wearable market. The evolution of wearable
technology is likely to occur as follows [9]: As
components get smaller, wearables become more
efficient and powerful. Future Smart Wearables
will be more sentient and adaptive, i.e. be able
to sense and analyze the situation of the user and
learn how to adjust to changing behaviors and
needs. Furthermore, wearables will be multipoint,
i.e. multiple sensors and devices will cooperate to
perform increasingly complex tasks. Ultimately,
wearables will seamlessly interact with environ-
ments such as smart homes or smart offices.
The following examples illustrate the disruptive
potential of EI for smart wearables:
Implanted intelligent micro-devices will au-
tonomously monitor and influence physiologi-
cal parameters such as heart rate, blood sugar
level etc. Based on machine learning and rea-
soning, they will be able to consider changes
of a person condition and behavior and develop
strategies how to adjust their actions appropri-
ately.
Associations of wearables and other devices
(smartphone, smart home etc.) may form a
distributed EI system. For instance, a personal
trainer drone will be able to either follow or
lead an athlete (e.g. a hiker, runner, or cyclist)
while monitoring his heart rate, breathing rate,
blood pressure and body temperature as well
as his running pace, step length and step fre-
quency supported by wearable sensors. The
drone may act as a coach, tour guide, and
pacesetter. It may even make autonomous deci-
sions about the route to be taken by the athlete
May/June 2021 3
Department Head
based on the training goal, the fitness and
performance of the athlete, and the weather
situation. In particular, it will have the task to
watch the heart rate for signs of overexertion
or dehydration and prevent unhealthy behavior
by appropriately guiding the athlete.
In the face of these developments the follow-
ing questions arise: How will the use of wearable
EI influence the way humans behave and interact
at home, at work and in public spaces? What are
the psychological, social, and legal implications
of the omnipresence of such devices? Which
value propositions are feasible, and which kind
of ecosystems are needed to enable these value
propositions? Which challenges and risks (e.g.
regarding safety and privacy) must be considered
and mitigated?
Smart Healthcare (SH) denotes a future
health service system that uses technology such
as wearables, IoT, and the mobile internet to
dynamically access information, connect people,
resources and institutions associated with health-
care, and then actively responds to medical needs
in an intelligent manner [10]. SH employs various
kinds of advanced technology (e.g., IoT, 5G, AI,
or modern biotechnology) for aspects of health-
care such as disease prevention and monitoring,
diagnosis and treatment, hospital management,
health decision-making, and medical research to
help involved stakeholders to make informed de-
cisions and facilitate the rational allocation of
resources.
Smart Wearables in combination with Quanti-
fied Self technologies have the potential to form
a cornerstone of future SH. A crucial element
of SH is the ability to correlate and analyze the
vast amount of unstructured data by using AI
algorithms that enable prediction, alerting, and
taking both corrective and proactive action.
The global wearable medical devices market
has been valued $10.6 billion in 2019, and is pro-
jected to reach $67.2 billion by 2030, at an 18.3%
CAGR between 2020 and 2030 [11]. Examples
for SH applications under development are:
Cancer treatment using smart monitoring sys-
tem
Smart glucose monitoring and insulin pens
Closed loop (automated) insulin delivery
Connected inhalers
Ingestible biosensors
Connected contact lenses
Smartwatch app monitoring depression
Coagulation testing
Asthma monitor
While these applications require some intel-
ligence for analyzing medical data, suggesting
reasonable measures, or performing appropriate
actions, the disruptive potential of EI systems in
healthcare lies in autonomous medical robots. We
envision those robots to be able to continuously
monitor a patient’s health status (supported by
medical wearables), autonomously perform di-
agnosis and treatment (not only in emergency
situations) and involve human medical staff if
required. Autonomous surgical robots are a spe-
cialized, more advanced form of medical robots
that may not only assist but even replace human
surgeons in the future. The introduction of med-
ical robots will be crucial for compensating the
staff shortage in healthcare that will become more
and more severe as the share of population aged
65 or older increases.
Key questions regarding the future of SH are:
Will there be a paradigm shift towards decentral-
ization of healthcare, i.e. diagnosis and treatment
will increasingly occur outside the hospital at
home? How can safety and security of such
systems be safeguarded? Which personal data
will be collected, and how can misuse of these
data be prevented? How do hospitals, physicians,
health care service providers and device vendors
reasonably collaborate? Which rules and regula-
tions are required to ensure smooth and secure
operation of SH ecosystems?
Everyday life robots are among the most in-
fluential innovations of the next 10-15 years: ap-
plying EI in everyday life situations. These cyber-
physical systems will engage in open, dynamic
environments to provide and utilize digital and
physical offerings (data, energy, assistance, . . . ).
While everyday life situations are manageable
for humans, they are among the most difficult
problems for robots: Due to their unstructured na-
ture it is almost impossible to describe (i.e., pro-
gram) the intended behavior unambiguously. In-
stead, perception, planning and execution must be
implemented using statistical approaches based
on assumptions about the expected environment,
4Computer
which is challenging and error prone. However,
current advances in AI – consider autonomous
driving – indicate that robots might soon be
widely applicable in such situations: Robotic as-
sistants could do the dishes and clean up the
bathroom or do other simple assistive tasks for
elderly at home. A robot might even give first
aid until the emergency physician arrives. Becom-
ing cheaper, robotic cooks and waiters might be
employed as well as robotic shopping helpers.
Once these robots are able to operate in public
space, more and more application fields can be
established, obviously not limited to health and
assistance (e.g. robotic police officers).
We believe that all these examples will be
possible in a few years from now. It will be
important for the acceptance of these robots, that
humans can understand and trust their behavior,
especially if the robot does things interactively
with a human (not only via user interfaces, but via
natural language, gestures, or actions). Probably,
this means that collaborative everyday life robots
must behave similarly to humans. Finally, we
think it is among the most important questions,
in how far humans can control these robots:
Must robots obey to every human (e.g. when he
or she shouts “STOP”) as long as it does not
harm anybody, as suggested by Isaac Asimov.
And how can owners and passers-by keep self-
determination w.r.t. to their free will and data
about them?
EMBODIED INTELLIGENCE FOR
CITIES
EI is emerging as an important tool to address
the challenges of urbanization.
CIS for utilities networks: through the cou-
pling of IoT and AI within EIs, things within
cities can be represented as EIs (e.g., utility or
mobility networks) to help cities manage the chal-
lenges that come with rising population density.
E.g., an EI that combines sensors and actuators
(e.g., traffic light controllers) of a city’s traffic
network with AI data analytics can optimize
traffic management. Also, utility networks, such
as for water supply, can be represented in an EI
to implement self-regulation (e.g., optimization of
water pressure), self-diagnosis (e.g., monitoring
of water pressure and quality), and steps to elim-
inate malfunctions.
Intelligent electrical grids could increase the
reliability of energy networks and the energy
supply security by helping to optimize energy
generation, consumption and costs (based on
energy supply, demand, and price predictions),
and by autodetecting their own status and taking
first steps towards maintenance and repair when
needed (e.g., by sending notifications).
The global energy production is undergoing a
change towards renewable energy sources, such
as wind, photovoltaic, hydroelectric, or geother-
mal power. These energy sources are more of-
ten exploited through small-scale and distributed
power generating plants. This trend pushes the
demand for advanced electrical grid infrastruc-
ture. New technologies for building a smart grid
infrastructure are coming up, driven by digitaliza-
tion that leverages on sensors, and software-based
control. The market for smart grid technology is
expected to grow from USD 23.8 billion to USD
61.3 billion between 2018 and 2023 with a CAGR
of 20.9% [12].
Thereby, a key factor is the utilization of
two-way communication between consumers and
utilities. I.e., the smart grid system can access
data about the current electricity demand and
supply, with smart energy meters as a central part
of such systems.
In the long run, we predict that the smart grid
will develop towards an aggregate that is repre-
sented by an EI as collaborative intelligent system
(CIS). There will be a central point of interaction
between the manager of the grid and this CIS,
which will implement autonomous behavior that
fosters efficient usage of energy and production of
electricity. Thereby, smart metering devices will
act as the nervous system of such smart grid
EIs. Key questions in this regard relate rather
to the political and organizational level, than to
technological issues. E.g.: How can we bypass
institutional barriers that currently prevent open
platforms and ecosystem on top of smart grid
infrastructures for making such an EI commer-
cially feasible? This relates to the question of the
feasibilities of giving access to the relevant data,
e.g., from different smart grid sources.
Intelligent things that can eliminate the
need for extensive utilities networks represent
an application area with large market poten-
tial in developing economies, where lack of
May/June 2021 5
Department Head
financial resources hampers investments in in-
frastructure. For example, photovoltaic systems
already allow for electricity generation at the
point of consumption. Devices for waste treat-
ment, water filtering, or gaining water out of
waste (http://www.gatesnotes.com/development/
omniprocessor-from-poop-to-potable) not only
help eliminate infrastructure network costs, they
also help tackle pollution and water scarcity. The
elimination of utilities networks allows for faster,
more flexible expansion of access to electricity,
heating, and fresh water.
Another area for EI in cities is public services:
the dog-resembling, remotely-controlled robot pa-
trolling a highly-populated area in Singapore to
help enforce social distancing [13] may well be
a forerunner of intelligent robots that will help
ensure the safety or that will do various repetitive
chores (street cleaning for example) in the cities
of the future.
Autonomous vehicles are perhaps the most
known application of EI. The global market for
fully autonomous vehicles with no manual inter-
vention by the driver is expected to reach $95
billion by 2030; by 2035, it is expected to reach
around $260 billion.[14]
With autonomous vehicles, individual mobil-
ity as we know it may no longer exist. In the fu-
ture, autonomous vehicles may be able to connect
to car-sharing platforms and communicate among
each other to optimally pick up and take travelers
to their destinations. Autonomous vehicles could
even themselves become platforms for product
and service placements. A mix of autonomous
cars and buses could ensure better public trans-
portation, not only in cities but also in rural areas,
increasing the attractivity of rural areas; this may
even reverse the urbanization trend, leading to
more people remaining in or moving to rural
areas. The advantage intelligent buses offer in
rural areas is that they drive down personnel
costs for bus operators (no driver needed). More
autonomous buses (than human-driven buses) can
travel at similar or lower costs, hence they hold
the potential for significantly better public trans-
portation services in areas with low passenger
numbers, where profitability is currently too low
for buses to travel as often as demand would
dictate. More efficient public transportation and
the use of autonomous taxis or car-sharing would
lead to decongestion of traffic due to fewer ve-
hicles on roads. Looking from the perspective
of vehicle producers: as the vehicle-use intensity
would increase, the average car lifetime would
decrease, hence less vehicles on roads would not
necessarily mean lower demand for vehicles.
Another advantage of new mobility models
enabled by autonomous vehicles is the elimi-
nation of parking lots, which offers additional
space for other purposes. Autonomous vehicles
could not only help tackle urbanization issues
such as traffic jams and dearth of public space:
autonomous vehicles are expected to drive safer,
and this could lead to reduction in service and
insurance costs (due to less accidents) as well as
to savings in production costs and higher design
flexibility by allowing for lighter (less-resistant to
impact) auto bodies.
Another advantage that drives the adoption
of autonomous vehicles is the reduction in en-
ergy consumption: platoon driving (travelling in
swarms or convoys with the help of car-to-car
communications) is known to improve efficiency
and traffic flow.
People could spend their time travelling in
more productive ways than driving: working,
reading, eating, etc. Autonomous vehicles would
not only disrupt the mobility of people, but also
the logistics industry as they open new possibili-
ties for the transportation of goods, energy, water,
and raw materials.
Another example of autonomous vehicles
in cities, outside mobility, are waste collection
trucks: with no need for a driver to supervise
them, they could operate without stopping, for
more hours during the day/week, meaning that
less trucks would be needed on roads.
Intelligent, fully autonomous cars are not far
off into the future. They are expected to hit the
roads in the late 2020s, as several companies
are looking into the technology. Already since
October 2020, after years of on-road real-world
tests, Alphabet subsidiary Waymo offers fully
driverless rides (there is no human driver in
the car, although the vehicle is still supervised
remotely by humans) to the general public in
Phoenix (Arizona, US).
The adoption of autonomous vehicles poses
a range of questions for the automotive industry
and policymakers, such as: Who will be liable for
6Computer
accidents caused by self-driving vehicles? How
will the business models for self-driving vehicles
look like? How will the partner and competitor
ecosystems look like?
CIS and intelligent things open new oppor-
tunities for better planning of the resources a
city has at its disposal, resource savings and
reduction of CO2 emissions. However, there are
many questions to be answered on the way to the
cities of the future: How can the various intelli-
gent things and CIS owned/operated by different
organizations communicate with one another/how
could the different systems by integrated? What
legal changes are needed and how should they be
introduced? What are the data privacy issues and
how should they be addressed?
EMBODIED INTELLIGENCE FOR
PRODUCTION
Three selected developments related to pro-
duction show the potential impact of EI on
the megatrends glocalization and scarcity of re-
sources: microfactories, production as a service,
and future food production.
Microfactories will enable decentralization
and individualization of production that has not
been possible before. Currently, most manufac-
turing is still either industrial or artisanal. Indus-
trial manufacturing can produce standard items
in large quantities. However, setting up industrial
production runs requires large up-front invest-
ments and thus business risks. Artisanal manu-
facturing, the manual production of custom-made
items, better meets the needs of customers, but
it is expensive and does not scale beyond small
quantities. Microfactories [15] aim at filling the
gap between industrial and artisanal production.
This can substantially reduce the risk of innova-
tion. New products can be produced in medium
quantities (in volumes of hundreds to thousands
of pieces) in microfactories to test marketability
before making the step to industrial production,
which requires a substantial up-front investment
with the associated risks.
The idea is to set up miniaturized produc-
tion systems close to customers that allow the
decentralized production of individualized prod-
ucts. This idea of microfactories was already put
forward in the 1990s when the miniaturization of
manufacturing machines became possible. Recent
developments are driven by advances in digital
technologies, such as 3D printers, 3D scanners,
production software, CNC machines, or laser
cutters. Then, representing such a microfactory
through an EI and using IoT and Artificial In-
telligence makes flexible, efficient manufacturing
of products accessible. The progress on intelligent
things will drive the decentralized production of
highly customized products to meet local needs,
with short lead and delivery times. First examples
are Local Motors in the automotive sector and
Adidas Speedfactory in the textile sector.
Production as a Service. Driven by advances
in industrial IoT, AI, and microfactory technolo-
gies, things within factories such as smart robots
(19.6% annual growth [16]) and UAVs (15.5%
annual growth [17]) become increasingly flexible
and adaptable. Today these machines are used
to save engineering time. In the future, EI in
autonomous production cells and in autonomous
intra- and inter-logistics could enable fully auto-
mated production of a large variety of products
given a bill of materials (BOM) and a bill of
process (BOP), only. In such a future, products
can be designed and optimized digitally and in-
dependently from the processes of planning, exe-
cuting, and maintaining its production: production
becomes a service. Production workloads could
be allocated flexibly to smart factories all over the
world, so new products can ’go live’ and scale
with demand immediately - like cloud services.
Production as a service could even become faster
and cheaper than product-specific mass produc-
tion due to economies of scale.
Manufacturing accounts for at least 15.2% of
the global GDP [18]. The following transition
scenario shows how EI could gradually disrupt
this sector. Initially, it should be uneconomical
for most manufacturers to build smart factories
for specific production problems. However, smart
factories could reduce cost and increase adap-
tation speed for companies, which have a huge
variety of products (e.g. retail trade companies
with own products) or complex value chains
(e.g. automotive industry). After maturing, smart
factories might be opened for third-party prod-
uct designers: from internal reuse over outsourc-
ing to an ecosystem of shared factories. Then,
they might be connected by integrating platforms
(solving multi-factory problems). Finally, smart
May/June 2021 7
Department Head
factories might be provided by national or munic-
ipal enterprises as part of the public infrastructure
for local citizens.
With such scenarios in mind – what, if for in-
stance 80% of all manufacturing can be achieved
completely as a service by 2035? Who will
(be able to) build and operate this infrastruc-
ture, and where? What are the key technologies
and competencies? Which (new) business models
and allocation mechanisms ensure that everybody
(product developers, manufacturers, technology
providers and consumers world-wide) can benefit
from these advances? And how will the digital-
ized production change our way to work, to live
together, and maybe even our values?
Future Food Production. The growth of the
world population will drive up demand for food
by 50% by 2050. Due to the growing prosperity
in developing countries, the composition of the
food components will change towards a more
meat-oriented diet, leading to an 88% increase
in the demand for meat and meat products over
this period. However, only 3% of the global
calorie consumption is obtained from meat with
a comparatively poor CO2 balance, and 50% of
all calories come from only three plant-based
food sources, namely rice, wheat and maize [19].
Consequently, the global calorie demand can only
be met by basic foods from optimized tradi-
tional agriculture. Hence, there are three main
issues that need to be addressed: increasing the
production of basic ingredients for staple foods,
finding ways to provide more meat or meat-
substitute products and improving the quality
of food production to increase health. To meet
these demands, a paradigm shift is needed, away
from steadily increasing mass production towards
demand-oriented, prediction-based food produc-
tion, supported by sustainable and optimized pro-
cesses and climate-neutral production, adaptable
processing and autonomous distribution. This can
only be achieved by making the local food pro-
duction more intelligent by using AI and Embod-
ied Intelligence.
Nowadays, modern digital farming uses GPS-
controlled autonomous cultivation of the arable
land to maximize the yield. Customer-specific,
optimized control of pests and weeds using
drones with image recognition enable a targeted
use of pesticides. In the context of precision agri-
culture, area-specific cultivation of the fields is
used, as these mostly have uneven soil conditions.
By measuring the moisture and temperature of
the soil by autonomous agricultural robots, in
combination with the use of satellite imagery and
AI, growing conditions can be optimized. It is
now known that AI-assisted irrigation of fields
can reduce the consumption of water and fertilizer
dramatically [20].
For future, we propose the need for a reduc-
tion of food waste by prediction using AI and
the reduction of meat consumption as the food
with the worst ecological balance. New regional
production processes, such as in-vitro meat, meat
substitutes and others, will lead to independence
from classic livestock farming. An intelligent
demand management can optimize the local and
temporal distribution of food. And finally, in-
creasing food quality leads to better health of
the population, supported by home robots that
prepare individual meals and sustainably replace
unhealthy instant food.
CONCLUSION AND FUTURE WORK
We predict that over the next decade, driven
by further development of specialized AI systems
(for particular applications) and the potential for
significant benefits (discussed in the previous sec-
tions), EI will start to become a part of everyday
life.
EI is the logical next step in technology de-
velopment and may help address the tremendous
societal and economical challenges we are facing,
stemming from the described megatrends. EI in
general goes well hand in hand with digitaliza-
tion. Microfactories and production as a service
support glocalization and help tackle resource
scarcity by increasing the efficiency of local,
medium-volume production. Robots that can take
care of household work and provide basic health-
care could help elderly people live longer in their
own homes (instead of elderly care facilities)
and hence help address the challenges of aging
population. New forms of intelligent things that
can eliminate the need for infrastructure, along
with CIS such as electricity and water grids, smart
homes and cities can help tackle resource scarcity
(via resource optimization) and climate change as
well as help optimize consumption, traffic, and
public services in increasingly crowded cities.
8Computer
EI could not only help tackle challenges
brought by megatrends, but it also has the poten-
tial of reversing some megatrends. For example,
EI that eliminates the need for infrastructure,
along with teleworking and autonomous vehicles
could stop urbanization or even lead to the oppo-
site megatrend: return of people into rural areas.
EI could also drive the emergence of completely
new megatrends, which we are currently unaware
of.
In the previous sections we have showed that
the emergence of EI poses challenges for com-
panies, governments, research institutions, and
the civil society at large. Summed up, these
challenges are:
Self-determination issue: what are the ”free
decision” limits of EI?
Operational safety and data security: how can
the safety and data security of EI and its
environment be safeguarded?
Privacy and data ownership issues: How can
data privacy be ensured? Who will own and
have the right to use the massive amounts of
data that EI will gather and have access to?
How can data misuse be prevented?
Standards and regulatory issues: Which rules
are required to ensure smooth and secure op-
eration of EI and their surrounding ecosystems
(e.g., open standards)? Who will be liable for
accidents caused by EI?
Infrastructure: how should the existing infras-
tructure (communication, roads, energy, etc.)
be adapted for EI? Are new forms of infras-
tructure needed? Will infrastructure become
obsolete?
Business models: what are the optimal ways
to employ and monetize the different forms of
EI? How will the new ecosystems look like?
Strategy: What steps should governments,
companies, and the civil society at large take
to speed up and smoothen the transition to EI?
Social impact: How will EI influence the way
humans feel, behave and interact at home, at
work, and in public spaces?
Given the complexity of these issues, and the
wide range of economic actors EI would impact,
it is required that companies, research institutions,
governments, and NGOs work together if they
want to speed up the adoption of EI and ensure
that it is used in a sustainable manner that benefits
society. The foresighting project conducted by
Siemens, Fortiss and Deutsches Dialog Institut
is exploring the future of EI by looking into
future EI scenarios (including societal, business
and technological aspects), understanding the per-
spectives of the various groups of stakeholders
(governments, companies, research institutions
and civil society), and proposing recommenda-
tions for the way forward.
ACKNOWLEDGMENT
This work received funding from Germany’s
Federal Ministry for Economic Affairs and En-
ergy under grant agreement No. 01MT20005.
REFERENCES
1. S. Singh, “Roland Busch: The fifth mega trend that is
changing Siemens’ future,Forbes, 2019. [Online].
Available: https://www.forbes.com/sites/sarwantsingh/
2019/05/29/roland-busch-coo-cto-of- siemens-the-
fifth-mega- trend-that-is- changing-siemens-future
2. World Bank, “Population estimates and projections
database,” 2020. [Online]. Available:
https://databank.worldbank.org/source/population-
estimates-and- projections
3. United Nations Department of Economic and Social
Affairs, “World population prospects 2019 – Highlights,
2019. [Online]. Available: https://population.un.org/
wpp/Publications/Files/WPP2019 Highlights.pdf
4. United Nations, “World urbanization prospects: The
2018 revision – Key facts,” 2018. [Online]. Available:
https://population.un.org/wup/Publications/Files/
WUP2018-KeyFacts.pdf
5. C. Krys and K. Fuest, “Megatrend scarcity of
resources,Roland Berger Trend Compendium, 2017.
[Online]. Available:
http://www.rolandberger.com/publications/publication
pdf/roland berger trend compendium 2030 trend
3 scarcity of resources 1.pdf
6. A. Cangelosi, J. Bongard, M. H. Fischer, and S. Nolfi,
“Embodied intelligence,” in Springer Handbook of
Computational Intelligence. Springer, 2015, pp.
697–714.
7. W. A. Haseltine, “Aging populations will challenge
healthcare systems all over the world,Forbes, 2018.
[Online]. Available: https://www.forbes.com/sites/
williamhaseltine/2018/04/02/aging-populations- will-
challenge-healthcare- systems-all-over-the-world
May/June 2021 9
Department Head
8. Grand View Research, “Wearable technology market
size, share & trends analysis report,” 2020. [Online].
Available: https://www.grandviewresearch.com/
industry-analysis/wearable-technology-market
9. J. Quinlan, “The future of wearable Tech,” 2015.
[Online]. Available: https://www.wired.com/insights/
2015/02/the-future- of-wearable-tech
10. S. Tian, W. Yang, J. M. LeGrange, P. Wang, W. Huang,
and Z. Ye, “Smart healthcare: making medical care
more intelligent,Global Health Journal, vol. 3, no. 3,
pp. 62–65, 2019. [Online]. Available:
https://www.sciencedirect.com/science/article/pii/
S2414644719300508
11. P&S Intelligence, “Wearable medical devices market
research report,” 2020. [Online]. Available:
https://www.globenewswire.com/news-release/2020/
04/13/2014947/0/en/Wearable-Medical-Devices-
Market-to- Generate-67-2- Billion-Revenue-by-2030- P-
S-Intelligence.html
12. Markets and Markets, “Smart grid market – Global
Forecast to 2023,” 2018. [Online]. Available:
https://www.marketsandmarkets.com/Market-Reports/
smart-grid-market-208777577.html
13. J. Nalewicki, “Singapore is using a robotic dog to
enforce proper social distancing during covid-19,” 2020.
[Online]. Available: http://www.smithsonianmag.com/
smart-news/singapore-using-robotic-dog-enforce-
proper-social- distancing-during-covid-19-180974912
14. M. R¨
omer, S. Gaenzle, and C. Weiss, “How
automakers can survive the self-driving era.” 2016.
[Online]. Available:
https://www.kearney.com/automotive/article?/a/how-
automakers-can-survive-the-self- driving-era
15. J. O. Montes and F. X. Olleros, “Microfactories and the
new economies of scale and scope.Journal of
Manufacturing Technology Management, 2019.
16. Allied Market Research, “Smart robot market
overview.” 2018. [Online]. Available: https:
//www.alliedmarketresearch.com/smart-robot-market
17. Markets and Markets, “Unmanned aerial vehicle
market – Global forecast to 2025,” 2019. [Online].
Available:
https://www.marketsandmarkets.com/Market-Reports/
unmanned-aerial- vehicles-uav-market-662.html
18. World Bank, “Manufacturing, value added (percent of
GDP),” 2018. [Online]. Available:
https://data.worldbank.org/indicator/NV.IND.MANF.ZS
19. T. Searchinger, R. Waite, C. Hanson, J. Ranganathan,
P. Dumas, E. Matthews, and C. Klirs, “Creating a
sustainable food future. a menu of solutions to feed
nearly 10 billion people by 2050,” 2019. [Online].
Available: https://research.wri.org/sites/default/files/
2019-07/creating- sustainable-food-future 2 5.pdf
20. W. Aulbur, “Farming 4.0: How precision agriculture
might save the world,” 2019. [Online]. Available:
https://www.rolandberger.com/publications/
publication pdf/roland berger precision farming.pdf
Arne Br ¨
oring is a Senior Researcher at Siemens’
corporate research unit in Munich (Germany). Arne
is a computer scientist and received his MSc from
University of M¨
unster (Germany), and a Ph.D. from
the University of Twente (Netherlands). Contact him
at arne.broering@siemens.com.
Christoph Niedermeier is with Siemens Technology,
Munich, Germany. Christoph received his diploma in
Physics at Technical University of Munich in 1989,
and a Ph.D. in Biophysics from Ludwig-Maximilians-
Universit¨
at Munich in 1995. At Siemens Technology,
he currently works on IoT systems, performing R&D
on semantic models, knowledge-based processes,
and software architectures for the IoT. Contact him at
christoph.niedermeier@siemens.com
Ioana Olaru is with Deutsches Dialog Institut (DDI),
Frankfurt am Main, Germany. Ioana received her
MBA from Pforzheim University in 2016. At DDI,
she supports companies with innovation manage-
ment, as well as with development and implemen-
tation of digital transfer strategies. Contact her at
ioana.olaru@dialoginstitut.de
Ulrich Sch ¨
opp is with fortiss GmbH, Munich, Ger-
many. Ulrich received a Ph.D. in Computer Science
from University of Edinburgh in 2006 and a Habilita-
tion in Computer Science from Ludwig-Maximilians-
Universit¨
at M¨
unchen in 2015. He currently works as
a research scientist on safety and security at fortiss.
Contact him at schoepp@fortiss.org.
Kilian Telschig is with Siemens Technology, Munich,
Germany. Kilian received his M.Sc. in Software Engi-
neering from University of Augsburg, Technical Uni-
versity of Munich and Ludwig-Maximilians-Universit¨
at
Munich in 2015. He is working on his Ph.D. on up-
dating distributed embedded applications during op-
eration under supervision by Prof. Alexander Knapp
at the ISSE institute, University of Augsburg. Contact
him at kilian.telschig@siemens.com.
Michael Villnow is with Siemens Technology, Er-
langen Germany. Michael received his diploma in
10 Computer
Mechatronics from Friedrich-Alexander University of
Erlangen-Nuremberg in 2008. At Siemens Technol-
ogy, he currently works on IoT systems, constrained
field devices and strategic topics regarding sen-
sor nodes and embedded software. Contact him at
michael.villnow@siemens.com.
May/June 2021 11
... AI achieves information learning and evolution with a cloud brain and a neural network. Their combination will promote the development of ward-round robots that can realize real-time remote ward-round to reduce the burden on nurses (28,29). In terms of disease prevention, diagnosis, treatment, and nursing, 5G network technology supports real-time transmission of huge data on human health and helps medical institutions in continuous physical monitoring of wearers. ...
... Based on these data, AI can record and analyze individual patients comprehensively and continuously so that personalized healthcare schemes can be formulated. Meanwhile, doctors can make remote judgments and analyses according to relevant data such as medical records and images, and provide personalized treatment and nursing schemes in time (28,30,31). Additionally, the 5G+AI system enables multi-party simultaneous consultation and connection with other hospitals and peers in a 7 × 24 manner, and multidisciplinary consultation can be held anytime, thus enhancing diagnostic accuracy. ...
... Digi robot It is used in reminding staff to provide treatment or medicine to patients and can help enforce social distancing and ensure the safety (28,97). ...
Article
Full-text available
The medical and healthcare industry is currently developing into digitization. Attributed to the rapid development of advanced technologies such as the 5G network, cloud computing, artificial intelligence (AI), and big data, and their wide applications in the medical industry, the medical model is shifting into an intelligent one. By combining the 5G network with cloud healthcare platforms and AI, nursing robots can effectively improve the overall medical efficacy. Meanwhile, patients can enjoy personalized medical services, the supply and the sharing of medical and healthcare services are promoted, and the digital transformation of the healthcare industry is accelerated. In this paper, the application and practice of 5G network technology in the medical industry are introduced, including telecare, 5G first-aid remote medical service, and remote robot applications. Also, by combining application characteristics of AI and development requirements of smart healthcare, the overall planning, intelligence, and personalization of the 5G network in the medical industry, as well as opportunities and challenges of its application in the field of nursing are discussed. This paper provides references to the development and application of 5G network technology in the field of medical service.
... A new area of interest revolves around the use of "digital twins" and "serious games" to offer a virtual platform on which to test design scenarios, dynamics and behaviours. Digital twins can be described as a digital copy of a physical thing and can be used to understand the real behaviour of, for example, a machine or an entire manufacturing plant (Broering et al., 2021). They increasingly integrate AI and machine learning within virtual, augmented and mixed reality-based simulation experiences (Callaghan et al., 2021). ...
... There are, however, human, ethical and other challenges posed by the use of AI in food systems. Similar to the results described in Section 2.2, issues of self-determination, operational safety, data security, privacy, data ownership and social impacts are raised when considering robotics (Broering et al., 2021;Tuomi et al., 2021) and drones (Hanrahan et al., 2021). However, in addition to the previous findings on ethics (Section 2.2 -Article #1 -Part 3.4), postcolonial computing concerns are being considered in the approach to AI-powered digital technology taken in countries outside the Global North. ...
... This shift brings a call for new ways of designing. One such call is for embodied intelligence (EI) as a design approach and a research field for the design of intelligent behaviour in embodied agents (Broering et al., 2021). EI for food production takes the shape of intelligent food demand management capable of decentralising and individualising production of highly customised products to meet local needs in a production-as-a-service model where micro-factories only needed a bill of materials and a bill of processes to commence production anywhere in the world. ...
Thesis
Full-text available
In an epoch when our environments are under threat from climate change, resource depletion and pollution, our food systems require urgent transition to sustainable pathways in order to feed 9 billion people by 2050. To do so in such a way that future generations’ social, economic and environmental realities are not compromised, the designers of the innovations and technologies deployed to transform the various socio-technical systems of food production and consumption have a critical role to play. Artificial intelligence (AI) is increasingly disrupting food systems and redefining the practices of entire fields from agriculture, transportation and processing to retail and consumer behaviours. An automated, but not agnostic, analytical and decision-making technology, AI is key to designing more efficient food systems. However, it should be considered one element of designed transition strategies and evaluated within sustainability and ethical frameworks. Consequently this research addresses the question: How might we design with AI for the sustainable and ethical transition of our food systems and practices? The research, conveyed through a series of peer-reviewed and published case studies, focuses on how designers can use AI to transition to sustainable food systems by investigating the use of AI across food systems and its impacts on sustainability outcomes. A systems thinking lens and an action-research approach have been deployed to conduct case studies of different food systems – food rescue, child nutrition in schools, digital agri-food research and farmers’ markets – to explore how AI is applied, the consequences of using it, how designers of transitions work with AI and what sustainable outcomes may plausibly be achieved. Building on design for sustainability theory and practice, the research draws on qualitative primary data from fieldwork, interviews, active AI system design and evaluation, and literature from the field. The research finds that designers of the transition required can use AI to support and scale up sustainable outcomes in food systems by drawing on the ethical and practice frameworks that underpin the field of design for sustainability. Transdisciplinary and participatory design methods integrated through agile iterative learning cycles are shown to support, and extend over time, the designer’s intent for both sustainable transition and the technical development of AI-powered tools. The research extends knowledge in the fields of food system innovation and design for sustainability by developing a method for and demonstrating the use of a bottom-up engagement framework for sustainable and ethical design with AI for sustainable transitions. Further, it presents a prototypical case and method for nuanced approaches to the relationship between AI and design so as to contribute a granular and practice-focused definition of designing with AI based on sustainable and ethical precepts. Finally, the research provides an ontological design reflection on what design with AI means in the context of being “designed by our designing” (Willis, 2006, p. 70) and the moral and ethical consequences of designing with AI and creating futures. It finds that design intent, and particularly the targeting of deep leverage points in systems such as goals, mindsets and paradigms, enables the designer to acquire agency within complex and increasingly automated systems of production, exchange and consumption.
Article
Full-text available
In the industrial Internet of Things domain, applications are moving from the Cloud into the Edge, closer to the devices producing and consuming data. This means that applications move from the scalable and homogeneous Cloud environment into a potentially constrained heterogeneous Edge network. Making Edge applications reliable enough to fulfill Industry 4.0 use cases remains an open research challenge. Maintaining operation of an Edge system requires advanced management techniques to mitigate the failure of devices. This article tackles this challenge with a twofold approach: (1) a policy-enabled failure detector that enables adaptable failure detection and (2) an allocation component for the efficient selection of failure mitigation actions. The parameters and performance of the failure detection approach are evaluated, and the performance of an energy-efficient allocation technique is measured. Finally, a vision for a complete system and an example use case are presented.
Article
Full-text available
With the development of information technology, the concept of smart healthcare has gradually come to the fore. Smart healthcare uses a new generation of information technologies, such as the internet of things (loT), big data, cloud computing, and artificial intelligence, to transform the traditional medical system in an all-round way, making healthcare more efficient, more convenient, and more personalized. With the aim of introducing the concept of smart healthcare, in this review, we first list the key technologies that support smart healthcare and introduce the current status of smart healthcare in several important fields. Then we expound the existing problems with smart healthcare and try to propose solutions to them. Finally, we look ahead and evaluate the future prospects of smart healthcare.
Article
Full-text available
Purpose – The purpose of this paper is to explore the microfactory model, the elements that enable it and its implications. The authors argue that microfactories reduce the risks and costs of innovation and that they can move various industries toward more local, adaptive and sustainable business ecosystems. Design/methodology/approach – This conceptual paper explores several processes and practices that are relatively new; hence, it uses online secondary sources (e.g. interviews with CEOs, videos, blogs and trade magazine articles) extensively. Findings – Given its versatility and high automation levels, the microfactory model can fill the gap between artisanal and mass production processes, boost the rate of innovation, and enable the local on-demand fabrication of customized products. Practical implications – Currently, manufacturers generally need to make large investments when launching a new product, despite high uncertainty about customer acceptance, thus risking considerable losses. The microfactory model offers a safer alternative by allowing a firm to develop and fabricate new products and test their acceptance in a local market before mass producing them. Microfactories also enable the local on-demand fabrication of highly customized products. Originality/value – This paper contributes to the discussion on the economic advantages and disadvantages of scale and scope, which have been insufficiently explored in the digital domain.
Article
Full-text available
Functioning of the Internet is persistently transforming from the Internet of computers (IoC) to the ‘Internet of things (IoT)’. Furthermore, massively interconnected systems, also known as cyber-physical systems (CPSs), are emerging from the assimilation of many facets like infrastructure, embedded devices, smart objects, humans, and physical environments. What the authors are heading to is a huge ‘Internet of Everything in a Smart Cyber Physical Earth’. IoT and CPS conjugated with ‘data science’ may emerge as the next ‘smart revolution’. The concern that arises then is to handle the huge data generated with the much weaker existing computation power. The research in data science and artificial intelligence (AI) has been striving to give an answer to this problem. Thus, IoT with AI can become a huge breakthrough. This is not just about saving money, smart things, reducing human effort, or any trending hype. This is much more than that – easing human life. There are, however, some serious issues like the security concerns and ethical issues which will go on plaguing IoT. The big picture is not how fascinating IoT with AI seems, but how the common people perceive it – a boon, a burden, or a threat.
Article
Full-text available
Humankind has the ability of learning new things automatically due to the capacities with which we were born. We simply need to have experiences, read, study... live. For these processes, we are capable of acquiring new abilities or modifying those we already have. Another ability we possess is the faculty of thinking, imagine, create our own ideas, and dream. Nevertheless, what occurs when we extrapolate this to machines? Machines can learn. We can teach them. In the last years, considerable advances have been done and we have seen cars that can recognise pedestrians or other cars, systems that distinguish animals, and even, how some artificial intelligences have been able to dream, paint, and compose music by themselves. Despite this, the doubt is the following: Can machines think? Or, in other words, could a machine which is talking to a person and is situated in another room make them believe they are talking with another human? This is a doubt that has been present since Alan Mathison Turing contemplated it and it has not been resolved yet. In this article, we will show the beginnings of what is known as Artificial Intelligence and some branches of it such as Machine Learning, Computer Vision, Fuzzy Logic, and Natural Language Processing. We will talk about each of them, their concepts, how they work, and the related work on the Internet of Things fields.
Article
Full-text available
Technologies for mapping the “digital twin” have been under development for approximately 20 years. Nowadays increasingly intelligent, individualized products encourages companies to respond innovatively to customer requirements and to handle the rising product variations quickly. An integrated engineering network, spanning across the entire value chain, is operated to intelligently connect various company divisions, and to generate a business ecosystem for products, services and communities. The conditions for the digital twin are thereby determined in which the digital world can be fed into the real, and the real world back into the digital to deal such intelligent products with rising variations. The term digital twin can be described as a digital copy of a real factory, machine, worker etc., that is created and can be independently expanded, automatically updated as well as being globally available in real time. Every real product and production site is permanently accompanied by a digital twin. First prototypes of such digital twins already exist in the ESB Logistics Learning Factory on a cloud- and app-based software that builds on a dynamic, multidimensional data and information model. A standardized language of the robot control systems via software agents and positioning systems has to be integrated. The aspect of the continuity of the real factory in the digital factory as an economical means of ensuring continuous actuality of digital models looks as the basis of changeability. For the indoor localization sensor combinations that in addition to the hardware already contain the software required for the sensor data fusion should be used. Processing systems, scenario-live-simulations and digital shop floor management results in a mandatory procedural combination. Essential to the digital twin is the ability to consistently provide all subsystems with the latest state of all required information, methods and algorithms.
Article
Full-text available
Embodied intelligence is the computational approach to the design and understanding of intelligent behavior in embodied and situated agents through the consideration of the strict coupling between the agent and its environment (situatedness), mediated by the constraints of the agent’s own body, perceptual and motor system, and brain (embodiment). The emergence of the field of embodied intelligence is closely linked to parallel developments in computational intelligence and robotics, where the focus is on morphological computation and sensory-motor coordination in evolutionary robotics models, and in neuroscience and cognitive sciences where the focus is on embodied cognition and developmental robotics models of embodied symbol learning. This chapter provides a theoretical and technical overview of some principles of embodied intelligence, namely morphological computation, sensory-motor coordination, and developmental embodied cognition. It will also discuss some tutorial examples on the modeling of body/brain/environment adaptation for the evolution of morphological computational agents, evolutionary robotics model of navigation and object discrimination, and developmental robotics.
Article
Full-text available
We provide an overview of the evolutionary approach to the emergence of artificial intelligence in embodied behavioral agents. This approach, also known as Evolutionary Robotics, builds and capitalizes upon the interactions between the embodied agent and its environment. Although we cover research carried out in several laboratories around the world, the choice of topics and approaches is based on work carried out at EPFL. We describe a large number of experiments including evolution of single robots in environments of increasing complexity, competitive and cooperative evolution, evolution of vi-sion-based systems, evolution of learning, and evolution of electronics and morphologies for autonomous robots
Article
The Digital Twin (DT) is commonly known as a key enabler for the digital transformation, however, in literature is no common understanding concerning this term. It is used slightly different over the disparate disciplines. The aim of this paper is to provide a categorical literature review of the DT in manufacturing and to classify existing publication according to their level of integration of the DT. Therefore, it is distinct between Digital Model (DM), Digital Shadow (DS) and Digital Twin. The results are showing, that literature concerning the highest development stage, the DT, is scarce, whilst there is more literature about DM and DS.
Aging populations will challenge healthcare systems all over the world
  • W A Haseltine
W. A. Haseltine, "Aging populations will challenge healthcare systems all over the world," Forbes, 2018. [Online]. Available: https://www.forbes.com/sites/ williamhaseltine/2018/04/02/aging-populations-willchallenge-healthcare-systems-all-over-the-world