POSITION PAPER FOR COGNITIVE SYSTEMS
AND ROBOTICS
ROBOCLUSTER, JANUARY 2011
Position paper on Robotics and Cognitive Systems
2
I NDEX
Abstract .................................................................................................................................... 3
Introduction ............................................................................................................................. 4
Research challenges ................................................................................................................. 6
Societal Challenges .................................................................................................................. 9
Competitive challenges ............................................................................................................ 9
State of the art and beyond ................................................................................................... 11
International Trends toward 2020 ......................................................................................... 13
References to European Research programmes ................................................................... 18
Danish competences and interests ........................................................................................ 19
Recommendations/Conclusion .............................................................................................. 22
Position paper on Robotics and Cognitive Systems
3
A BSTRACT
This position paper seeks to contribute to the task of DI-ITEK on aligning future
Danish ICT research with European research policies and agendas, and tries to
outline potential national research contributions to European Research Area (ERA).
Since the national research area is relatively small, researcher‟s network and main
research partners are just as often international. In consequence it is not possible to
point out several national research clusters but to present diverse capacities from
the universities involved as well as their main interests.
RoboCluster has interviewed leading researchers in the field that we find to have
complementary competences. The participating researchers are not composing the
whole picture of Danish research but they represent the diversity in Danish research
on robotics and cognitive systems.
In regard to recommendation for future strategic foci this paper suggest
establishment of a national multidisciplinary cognitive center for robotics as an
answer to interviewed researchers‟ wish to further coordinate and unite multi
disciplinary perspectives on the development of robotics and as a platform to build
new core competences in contribution to ERA.
Recommendations also concerns application areas which we find to have great notobtained potential.
We believe, that in order to push impact of Danish research further - both in a
national and European reality - it is necessary to coordinate research more, exploit
and develop core competences and through political encouragement increase focus
on „post prototype phase‟ and ensure foundations for „testing‟.
The way to any robotic breakthrough depends on a pallet of sub competences from
product design to an increase in cognitive abilities, but first of all it depends on a
test willing approach among end-users and developers in order to reap the benefits
of the last years of research.
Position paper on Robotics and Cognitive Systems
4
I NTRODUCTION
This paper sums up status on robotic research in a Danish academic setting with its
relations to cognitive research. Content of this paper is based on interviews,
European reports and prior knowledge given to RoboCluster through years of work.
The paper provides an overview on the Danish research network, its competences
and interests and points out strength positions based on interviews regarding R & D
activities.
Based on the existing network of the Danish national robotic cluster RoboCluster
and its closest partners in research, a line of interviews with researchers has taken
direction as a relay in order to follow researchers ways of collaboration and to
pursue their interests. New possible ways of bridging interest and research in
robotics has thereby appeared in the making of this paper.
RoboCluster is, as a national network of innovation funded by the Danish Agency for
Science Technology and Innovation, defined by its consortia, the partners‟ activities
and interests. The consortia consist of Danish Industrial Robot Association,
Technical University of Denmark, Aalborg University, University of Southern
Denmark, Kolding School of Design and Danish Technological Institute. 1
The main task of a national network of innovation is to strengthen the collaboration
on research, innovation and technological development among industrial partners,
knowledge institutions, Advanced Technology Groups and the public sector.
Perspectives and contributions from these sectors are included in this paper.
Future „break throughs‟ and Danish contributions to European research are given in
the following based on 11 qualitative interviews with key researchers‟ answers on
the state of the art. Representatives from this multiple disciplinary domain have also
given their recommendations on future policy and direction on relevant research.
The funding of the current research effort is presented in chapter 7 along with main
targets in Danish and European research strategies in this domain.
Challenges in the production of this paper centers on:



1
2
3
4
The relatively small amount of experience with EU funding
Researchers wish to research rather than participate in these meta descriptions
of the field
The national research field is neither sufficiently united nor sufficiently
specialized in terms of coordinated research areas.
www.robocluster.dk
Definition from www.wikipedia.co.uk
www.oxforddictionaries.com
www.oxforddictionaries.com
Position paper on Robotics and Cognitive Systems
5
PARTICIPATING RESEARCHERS, COMPETENCES AND INTERESTS
Lars Kai Hansen, Professor, Head of Section for Cognitive Systems, Center of Neuro informatics, Danish
Technical University
Cognitive component analysis, audio and music modeling, functional neuro imaging, molecular brain imaging,
intelligent processing of signals
Ole Ravn, Associated Professor, Head of section Automation and Control, Department of Electrical
Engineering, Danish Technological University
Adaptive control, The Adaptive Blockset, Computer aided control engineering, vision-based automation and
control, Control of robots and autonomous systems
Ole Caprani, Associated Professor, Department of Computer Science, University of Aarhus
Play and Learning/edutainment
Ole Madsen, Professor, Department of Production, University of Aalborg
Automation, robotics, sensors, small batch production
Thomas Bak, Professor, Department of Electronic systems, University of Aalborg
Automation, control engineering and systems theory, hybrid systems and sensor information fusion,
specifically addressed at social intelligent technology, motion control systems
Henrik Gordon Petersen, Professor of Applied Mathematics, Southern University of Denmark
Technologies for next generation of industrial robots, motion planning
Norbert Krüger, Professor, Head of Cognitive Vision Group, Southern University of Denmark
Computer vision, biologically-motivated vision, Robotics, cognitive systems, machine learning
Kasper Støy, Associated Professor, Mærsk McKinney Møller Institute,Southern University of Denmark
Biology inspired multi-robot coordination, modular robots and self-reconfigurable robots
Kasper Hallenborg, Associated Professor, Mærsk McKinney Møller Institute, Southern University of Denmark
Multi agent system architecture, pervasive computing, welfare technology, agents for ambient assisted living
Kim Steenstrup Pedersen, Associated Professor, Department of Computer Science, University of Copenhagen
Human motion modeling, 3D tracking and image descriptors
Søren Kyllingsbæk, Associate Professor, Center for Visual Cognition, Department of Psychology, University of
Copenhagen
Mathematical modeling of visual cognition, behavioral experiments, and neuro imaging of visual cognition
Position paper on Robotics and Cognitive Systems
6
F IGURE 1: P ARTICIPATING RESEARC HERS
R ESEARCH CHALLENGES
COGNITION
The term cognition (Latin: cognoscere, "to know", "to conceptualize" or "to
recognize") refers to a faculty for the processing of information, applying knowledge,
and changing preferences. Cognition, or cognitive processes, can be natural or
artificial, conscious or unconscious. These processes are analyzed from different
perspectives within different contexts, notably in the fields of linguistics, anesthesia,
2
neurology, psychology, philosophy, anthropology, systemic and computer science.
In other words cognition covers the mental action or process of acquiring
knowledge and understanding through thought, experience, and the senses.3
Cognitive research in relation to robotics can be approached from different points of
departure e. g. human sciences and technological sciences. Roughly put,
psychologist normally focuses on how the human mind and brain functions, how
and why humans perceive and act . Significant contribution to the development of
robotics regards basic knowledge on abilities to observe, remember, plan, reason,
and make decisions.
Technologists‟ research in cognition for robotics is handled through disciplines like
machine learning (computation neuroscience), and cognitive engineering as the
processing (interpretation) of sensory based input. Cognitive researchers with
technological point of departure focus strive to control how the robot interact and
2
3
Definition from www.wikipedia.co.uk
www.oxforddictionaries.com
Position paper on Robotics and Cognitive Systems
7
understand the real world and succeed in working autonomous among humans
guided by sensors. Cross-over researchers are few but most wanted for the further
development of artificial cognitive skills for robots.
There are still many scientific challenges to be solved, a few components related to
them are listed below.
From a technological perspective:

Robustness of tactile sensors delivering the data for ‘perception’.

Vision systems are matured but also need more robustness.

Getting mathematic models for visual attention to support the rest of the parts in a
robotic system.

Storage of gathered input for making qualified decisions.

Integration in real life is yet to be seen.

Cognitive architectures.
From a psychological perspective:

Finding good mathematic models of human behavior and transferring the knowledge to
the ‘technology side’

Intelligent systems based on self-learning algorithms
ROBOTICS
Robotics as a “the branch of technology that deals with the design, construction,
operation, and application of robots”4 is a broad and multidisciplinary research field
since robotic systems draws on many different technologies regardless of it being a
virtual or mechanical „artificial agent‟.
Cognition
Robotics
Information/
Technologinal
sensor data
sciences
4
Moving robotics further into the human sphere
characteristics of robot technology tends to change
towards ICT systems in general, but remains defined as
a mechanism, that replaces human effort through
automatic control, be it artificial agents or systems such
as social computing/pervasive computing. Embedded
software systems striving to make interaction with
technological hardware intelligent can in this sense also
be considered robot technology.5
On the hardware side of robotics locomotion is striving
to renew hardware by making it lighter and more flexible
through basic research of modular and cellular robotics.
www.oxforddictionaries.com
An embedded system is a computer system designed to perform one or a few dedicated function often
with real-time computing constraints. It is embedded as part of a complete device often including
hardware and mechanical parts. By contrast, a general-purpose computer, such as a personal computer
(PC), is designed to be flexible and to meet a wide range of end-user needs.5 In other words, the term
„embedded system‟ is “design and build (e.g. a microprocessor) as an integral part of a system or device”
(Oxford dictionary)
5
Position paper on Robotics and Cognitive Systems
8
To gain this flexibility development on the software side is inextricably linked to the
possibilities on the hardware side.
Sensors are used to make robots able to extract information from its surroundings.
To expand possible ways in building cognitive abilities for robots in unpredictable
environments development in sensor systems are of course crucial. Cognitive vision
systems, sonar sensors, dexterous hands and artificial skin still have their way to go.
The biggest challenge for robotic researchers and developers concerns the software.
Software must be robust, and based on common open standard so that functional
modules can be shared instead of being re-invented. Such basis would also enhance
further development on robots ability to comprehend and make qualified decisions,
even together.
For classification of this collective research field EUROP, the European Robotics
Technology Platform6 divides robot technology in the following application sectors:
Industry, Professional Service, Domestic Service, Security and Space. For a further
classification into groups of functionality robots are valued as „autonomous workers‟,
robotic co-workers, logistic robots, for surveillance and intervention, exploration and
inspection and for education and entertainment.
INDUSTRIAL PRODUCTION
Since the installation of the first industrial robot at General Motors assembly line in
1961 the classic robot (programmable machines), has been a part of the modern
industry. Today the industrial market is still by far the biggest in Denmark.
Research challenges for industrial robots are currently bin picking, (grasping of
flexible objects in uncontrolled environment), precise path planning, compliance
control, navigation in unknown production environments and improvement of vision
systems in order to develop e. g. autonomous and co-working robots.
Future research on industrial robots though lies more in the development of whole
production systems where robots are included as part of a „smart factories‟ concept
with energy awareness, agile manufacturing and customization. (See more under
Trends). This includes focus on modularity and new methods in ICT aggregation.7
Besides smart factories future research concerns a renewal of agricultural
production and a paradigm shift in food processing which implies cognitive systems
and robots for the purpose of upgrading effective and sustainable food production.
PROFESSIONAL SERVICE AND DOMESTIC SERVICE
This category embraces a broad range of robotic solutions and application areas and
implies for example surgical robots, rehabilitation robots, patient handling robots,
cleaning robots, pervasive computing systems (for elderly care) and lawnmower
robots.
6
http://www.robotics-platform.eu/cms/index.php
http://cordis.europa.eu/fp7/ict/micro-nanosystems/docs/fof-smart-factories-factsheetwp2011_en.pdf
7
Position paper on Robotics and Cognitive Systems
9
There are a range of research challenges in this area since robots in the service
sector still are at an incipient stage. The main and fundamental challenge is
therefore the handling of the so called human-robot problem. How can robots
interact with humans based on mutual understanding on situations and goals. The
human-robot problem is reflected in all challenges concerning cognitive robotic
systems including sensing, comprehension, intentional understanding of
surroundings and decision making. Interplay between these elements is part of the
needed answers to questions on safety in this field of application.
For „co-working‟, end-users want robots with the assurance of no injury in robothuman collisions. More work is still to be done on technologies enabling humanrobot interaction and detection of proximity, but there is a clear direction emerging
within the sector of the development of so called soft robots, not just in material but
also in mutual understanding.8
SECURITY AND SPACE
(Left out because Denmark is not playing a role in this research field)
S OCIETAL C HALLENGES
Denmark is challenged by its small workforce, high salaries, high expectations on
welfare but little will to take risks. The latter might be changing though due
consequences of the global financial crises.
Therefore major challenges for Denmark are:

Maintenance and redefinition of production in general, such as the rest of Europe.

Demographic shifts cause great pressure on welfare services – a redefinition of
healthcare includes moving new technological healthcare services into private homes.
9
This causes also ethical challenges in regard to e. g. surveillance and management of
now more self supporting patients.

Transfer of robot technology from the industrial application area into new generation
of hi-tech public hospitals. The aim is to get national SME’s (robot suppliers) involved in
this process of opening up a new market, which would strengthen the supplier chain as
well as the whole robot eco system.

Establishment of sustainable farming and food production
Challenges on todays production systems and healthcare systems calls for robot
technology as a driver for development of new intelligent methods.
C OMPETITIVE CHALLENGES
IN REGARDS TO INDUSTRIAL PRODUCTION
When production moves out of DK/Europe R & D research and knowledge intensive
jobs follows.
8
Examples include Shadow (http://www.shadowrobot.com ), Barrett Technology
(http://www.barrett.com/robot/index.htm), and KUKA (http://www.kukarobotics.
com/en/pressevents/news/NN_060515_Automatica_02.htm)
9 examples of technologies for surveillance involves both intelligent systems outside as well as inside the
body (pervasive computing and brain scanning techniques to detect body conditions
Position paper on Robotics and Cognitive Systems
10
Robots for new production methods will be a part of a holistic answer to global
competition abilities and could be a driver in needed innovation gearing. Due to
massive outsourcing DK is challenged on growth and risks losing jobs in general. So
in order to keep knowledge and important employment in the country/Europe focus
on „smart production‟ is valuable and necessary. This implies progress on e. g.
system architectures and integrated communication, control and cognitive
capabilities and facilitation of cross-fertilization between academic and industrial
research efforts in robotics.10
Normally robots are fitted to huge production volumes. For the needed automation
in small production batches better interface is required to ease threshold for
workers access to use robots. In Denmark a huge number of SME‟s is doing small
batch production and are therefore in some sense used to conditions of
competiveness.
Nevertheless a lower entrance level for automation of production is needed to make
it easier for SME‟s to integrate robots in their production. The Danish robot
developer and supplier Universal Robots11 is a good case to show how successful
work on interface and lower price fits SME users perfectly. For a true break through
open standards are needed.
General demands for future production are:

Development of a new European model of production systems for the factories of the
future (e.g. transformable factories, networked factories, learning factories) depending
on different drivers such as high performance, high customization, environmental
friendliness, high efficiency of resources, human potential and knowledge creation.

ICT-based production systems and high quality manufacturing technologies capable of
optimizing their performance with a high degree of autonomy and adaptability for a
balanced combination of high accuracy production.

Sustainable manufacturing tools, methodologies and processes that have the capability
of cost-efficiently shaping, handling and assembling products composed of complex
12
and novel materials.

Making Human-Robot Interaction more advanced in terms of the level on which
interaction occurs (learning by demonstration, anticipation, language) as well as the
robustness and reliability of interaction.
Cognitive robotic systems can help production industry to meet these necessary
demands if competitiveness is to be regained. This concerns the ability to use more
advanced sensors, the ability to learn as well as the fruitful interaction of robots and
humans.
IN REGARDS TO PROFESSIONAL SERVICE AND DOMESTIC SERVICE
Demands on robotic agents, in order to benefit from already matured technology,
should be met by focusing on cognitive abilities. This implies perceptive abilities in
10
11
12
ICT Work programme 2011-12 6.2 Challenge 2: Cognitive systems and Robotics
www.universal-robots.com
Challenge 7 ICT for the enterprise and manufacturing
http://cordis.europa.eu/fp7/ict/micro-nanosystems/docs/fof-virtual-factories-factsheet-wp2011_en.pdf
Position paper on Robotics and Cognitive Systems
11
learning and recognizing objects in realistic scenes, grasping in unconstrained
environments, 3D tracking of human motions, dealing with flexible objects and
cognitive abilities such as manipulation in unconstrained environments,
understanding of human emotions and behaviors.
These abilities can path the way for necessary safety in human-robot interaction
and thereby further integration of robotics into society.
Human agents on the other hand need to be able to relate to factors as design and
interface to reach a confidence in robot technology as being a partner.
S TATE OF THE ART AND BEYOND
Basically robots carry out three functions - they “sense”, “think” and “act” regardless
of their purpose. Sensors, path planning, vision, grasping etc. have been a part of
the robotics development for many years.
Sensing
Acting
Thinking
SENSING
A range of sensors are on the market (cameras, tactile sensor, temperature sensors,
etc.). All sensors are continuously improved in industrial R & D departments and in
research institutions. State of the art is:

Fusion sensors, combines input from different sensors, to make the robot more
accurate, and to make the robot better to move in an unknown environment. This
fusion of data is in constant focus right now, and several EU project focus on this
subject.

Tactile sensors , this kind of sensor is not developed enough, but will evolve in the
next decade. Sensors can be worn by humans for monitoring and self control with
health condition.

Vision systems, visual sensors: In the last couple of years vision systems has developed
quite a lot. Vision systems today are matured, but are still sensitive towards heavily
changing light conditions. Many simple vision systems with controlled lighting are
successfully integrated in the industry.
13
13
The term tactile sensor usually refers to a transducer that is sensitive to touch, force, or pressure.
Tactile sensors are employed where ever interactions between a contact surface and the environment are
to be measured and registered. Tactile sensors are usefuin a wide variety of applications for robotics and
computer hardware
Position paper on Robotics and Cognitive Systems
12
A cognitive vision system depends on cameras (visual sensors), often many sensors,
cognition, interpretation software, system integration etc. The biggest challenge
right now is the integration of all the components and making it into one intelligent
process.
Development of vision is the „sense‟ Danish researchers are most involved in. Future
robots will need to learn from its sensing. These abilities are challenges of their own
but far more robustness on the hard technology is needed since vision systems still
are sensitive to lighting.
THINKING (PLANNING, PO SITIONING, COGNITION )
Methods are established for well known static environments:

Collision free motion planning

Task planning for optimal motion (Offline)

Simultaneous localization and mapping (SLAM)

Autonomous production system should be able to survey itself in terms of
self control

Grasping abilities on a cognitive level of a 3 year old
14
The research area today is dynamic surroundings where a high level of cognitive
abilities is needed for the robots to be able to navigate and behave reasonable.
Automated Guided Vehicles, AGVs, is driving safely around among humans today
but these robots are only reacting. Future robots must be able to act on its own
initiative that is being capable of understanding abstract signals from its
surroundings.
The big challenge is to make a robot act sensibly on an abstract job description.
This includes abilities to collect data in terms of experience gathering and methods
for decoding surroundings like semantic analysis.
ACTING (GRASPING, HANDLING)
Grasping: Bin picking in unstructured environments is solved and so is new opensource grasp simulators developed for rigid elements. Research today regards non
rigid objects.
Handling: Pick and place problem is solved for uniformed object. Research area is
now the handling of mixed object.
Research is also being done on „flexible assembly‟ performed by robots with fine
motor skills.
14
is a technique used by robots and autonomous vehicles to build up a map within an unknown
environment (without a priori knowledg-.e) or to update a map within a known environment (with a priori
knowledge from a given map) while at the same time keeping track of their current location.
Position paper on Robotics and Cognitive Systems
13
I NTERNATIONAL T RENDS TOWARD 2020
The research fields of robotics will rapidly change towards 2020 and all application
areas will grow but some areas more significantly than others. Therefore only the
application we find most relevant in terms of Danish competences and societal
challenges are described in connection to the following outline.
The outline points out development for the major future market segments expected
in various robotic applications. Further the figure depicts trajectories of evolution for
each key segment across the levels of acceptance, from entry level to mainstream,
via the major growth phase15:
INDUSTRIAL MANUFACTUR ING IN MEDIUM AND LA RGE ENTERPRISES
Development on research areas such as industrial manufacturing in medium and
large enterprises will be limited, since value creation of a separate robot unit is
limited to either a reduction of cost or an increase of flexibility. This improvement
will demand an unrealistic amount of investment.
Instead research will focus on holistic solutions such as concepts for „smart
factories‟. Such concepts are key points of development and the primary frame for
research developments on robotics for industrial purposes.
Smart factories covers ICT-enabled intelligent manufacturing where contribution of
Information and Communication Technologies (ICT) to manufacturing aims to
improve the efficiency, adaptability and sustainability of production systems and
15
JRC Scientific and Technical reports 2010; A Helping hand for Europe: The competitive
outlook for the EU robotic Industry.
Position paper on Robotics and Cognitive Systems
14
their integration within agile business models and processes in an increasingly
globalised industry, requiring continuous change of processes, products and
production volumes. Also the further integration of any newly developed ICT into
the production lines and the industrial environments requires complementary
research and innovation efforts. These integration aspects will play a key role for
generating and using smart production systems for factories in different industrial
sectors. 16
Danish researchers are involved in this approach to renew and restore national and
European production.
INDUSTRIAL MANUFACTURING IN SMALL AND MICRO ENTERPRISES
Instead development is expected in SMEs use of robotics since many small
enterprises will get the opportunity to integrate robots due to lower costs, higher
degree of flexibility and hopefully lower threshold for the worker getting to use it.
Improvement in interface and further upgrading of skills among workers is crucial
for this development.
DOMESTIC USE IN ELDERLY CARE SUPPORT
Domestic use of robotics in elderly care support is part of the overall term welfare
technology which joints smarter working procedures or concepts for service that
sets free resources. Welfare technology covers robot technology, tele medicine, itsolutions and intelligent facilities.
As for today we have a good example in the
vacuum cleaner, but this is only 1 out of a
huge product range that will be developed
in the next decade. Some very different
kinds of products will be „comforting
pervasive systems‟, systems that makes self
control of the state of your heath/illness
possible. Such services are desirable for
both a stressed group of care assistants and
patients striving to maintain a life with
authority and dignity in their own home.
Since these implies holistic systems to surround and monitor the elderly conduct
there are bigger ethic barriers here than technological ditto.
In many cases elderly actually prefers a robot solution to help with e.g. intimate
matters instead of a person.
This marked will boom due to the demographic changes and high payments on
Danish labor. Future populations will live longer and therefore with an increasing
number of diseases. We need technology to help the fewer that will have to take
care of growing group of senior citizens.
General and personalized robots in households require rather complex bodies, high
level of cognitive, emotional and learning abilities.
16
http://ec.europa.eu/research/industrial_technologies/pdf/ppp-factories-of-the-future-strategicmultiannual-roadmap-info-day_en.pdf
Position paper on Robotics and Cognitive Systems
15
National research strategies support this on many levels and cross disciplinary
research are joint in large scale projects as well.
MEDICAL AND HEALTHCARE ROBOTICS
Robotics for this application area are divided in to two sectors:
1) Primary robotics for direct treatment of patients such as surgery robots and
robots for more general handling of patients e.g. human body washing robot. These
robots have been developed but will be further in the future, though Danish
research does not focus on these medico technical tasks.
Instead Danish research revolves around 2) The so called secondary sector which
covers all tasks included in the running of a hospital and the attempt to inter link
technological solutions.
As for today robots solves logistic
tasks such as AGVs transporting
heavy stuff and transporting
blood samples. In the future
there will be a whole lot of robots
“behind the scenes”.
An ongoing research project is a
blood sampling robot. This is an
example of a solution crossing
over two sectors: when a blood
sample has been taken by the
robot it is to be sent to the lab
and test answers should be sent
to the person involved/new department.
Other intelligent solutions to emerge in the near future are a full automatic central
for sterilization, an intelligent pneumatic tube system and a bed making robots. A
solution such as a system for automatic dosing of medicine is already at use.
Denmark is in a unique position with its limited size,
societal challenges, obvious needs for innovative
solutions for welfare services with high standards
and therefore encouragement to be a „test nation‟
on welfare technology. Furthermore Denmark will
make substantial investments in hospitals – building
new, renovating and expanding the existing in the
coming 10-15 years. The ambition is to make these
buildings as intelligent and automated as possible.
National research strategies support this on many
levels and research effort are coordinated in large
scale projects as well.
AGRICULTURE AND FOOD PROCESSING
Robotic solutions such as milking robots are state of the art, but coming up is new
agricultural machinery based on high tech ad on‟s and intelligent technology
transferred from other areas of production.
Position paper on Robotics and Cognitive Systems
16
Unmanned Vehicles collecting data in the field is possible today but to exploit the
full potential adjustments on security as in surveillance of the terrain is under
development by Danish researchers.
A high technological project is developing the next generations of sprayer booms for
agricultural production with the goal to reduce the excessive use of herbicides by at
least a factor 10.17 This Intelligent Sprayer Boom combines micro spraying, vision
systems and decision making into one system saving time and money for farmers.
Other highly innovative national initiatives with the multipurpose of developing a
whole new way of growing food, is being projected right now. This means a shift in
paradigm towards „vertical farming‟ and includes a high tech multilayer production
system for producing food & non-food products positioning Denmark in a leading
role in delivering sustainable and novel technologies to the future world food
market. Intelligent robotics is central for e.g. logistics and plant nursing in this new
method of production.
Danish research groups have great experiences with
plant nursing robots (Hortibot), a computer assisted
slope mowing robot (Casmobot), and a robotic fruit
picker is also under development.
The Danish Agrobotic Network was launched 2009
as part of a program with the main objective to
further expand the corporation and development
within the multidisciplinary agro-robotic field.18
Another Nordic innovation project is User-centric
Mobile Information Management in Automated Plant
Production. This collaboration is a platform in which agricultural and ICT knowledge
from the Nordic countries are combined to form a common system concept for
information management in mobile plant production working environments, where
19
automation can be used to bring more efficiency.
An example of a European project, where Danish research is contributing, is
FutureFarm, funded by the EU as part of the Seventh Research Framework
Programme. The project is about meeting the challenges of the farm of tomorrow
by integrating Farm Management Information Systems to support real-time
20
management decisions and compliance to standards.
To summon up Danish researchers have matured technologies for this application
area during the last years of collaboration and time is now to reap the fruits and
open up traditional methods to new intelligent solutions. End users as farmers and
gardeners are waiting.
TECHNICAL TRENDS IN COGNITIVE SYSTEMS AND ROBOTICS
The scheme below outlines the most likely development on research and
technologies in the next decade. The scheme is divided into cognition-related
subjects and more hardware related subjects. Looking at „natural language
17
18
19
20
http://hoejteknologifonden.dk/projektgalleri/projektgalleri/den_intelligente_sproejtebom/
www.agroboticnetwork.dk
http://www.nordicinnovation.net/prosjekt.cfm?Id=1-4415-286
http://www.futurefarm.eu/
Position paper on Robotics and Cognitive Systems
17
processing‟, robots will only be able to speak simple phrases in 2015, but most likely
be able to speak up to 100 words in 2020.
Cognition
Innovation
Natural language processing for
human interaction
2015
Simple phrases
2020
Reliable 100 word
vocabulary
Higher Cognitive Ability, common
sense, human logic
Laboratory only
Pilot projects
Human Interaction – Robots –
human working together IRL.
Laboratory
Limited pilot projects
Edutainment
Toys
Simple projects
Humanoids
Laboratory
Limited pilot projects
Human interfacing via thoughts
and nerve control
Laboratory
Limited pilot projects
especially with elderly
and disabled
Innovation
New power supplies, solar, fuel
cells etc.
2015
Pilot projects
2020
Limited use
New Sensors, touch sensitive skin,
tactile
Pilot scale projects
Limited use
Soft Robots which can adapt shape
Laboratory
Only in special
applications
Self reconfiguration robots that can
morph with hardware e.g. self
repair
Laboratory
Limited self repair
Robotics /Embedded systems
Source21
Indeed, systems pertaining to any of these areas have to understand their
respective dynamic environments, whether these call for navigation, grasping and
manipulation, interaction with people, or simply the recognition and description of
real or virtual objects and scenes.
Embedded systems will be everywhere in the future. Therefore challenges and
demands on embedded software systems are to make separate devices capable of
communicating unassisted with other intelligent devices in order to find the best
solution to a given task. Each embedded system will have to be integrated to selforganizing networks called cyber physic systems.22
21
22
The Strategic Research Agenda for Robotics in Europe, 07/2009
http://www.comon.dk/nyheder/Danske-og-kinesiske-forskere-sammen-om-indlejret-software1.378297.html
Position paper on Robotics and Cognitive Systems
18
R EFERENCES TO E UROPEAN R ESEARCH PROGRAMMES

ICT-AGRI is the acronym for the ERA-NET which has the full title “Coordination
of European Research within ICT and Robotics in Agriculture and Related
Environmental Issues”.
The overall goal of ICT-AGRI is to strengthen the European Research Area and
develop a common European research agenda concerning ICT and robotics in
agriculture.

NMP - Nanosciences, Nanotechnologies, Materials and New Production
Technologies.
The core objective of the theme is to improve the competitiveness of European
industry and generate the knowledge needed to transform it from a resourceintensive to a knowledge-intensive industry.

ICT Challenge 2: objective 2.1 Cognitive Systems and Robotics

ICT Challenge 5: ICT for Health, Ageing Well, Inclusion and Governance
This challenge addresses advanced ICT research for sustainable high-quality
healthcare, demographic ageing, social and economic inclusion, and the
governance of our societies.

ICT Challenge 7: ICT for the Enterprise and Manufacturing 23
The ICT contribution to this initiative aims at improving the efficiency, adaptability
and sustainability of manufacturing systems as well as their better integration
within business processes in an increasingly globalised industrial context.

PPP =>Research => Industrial technologies => Factories of the future
The "Factories of the Future" is one of the three Public-Private Partnership
included in the Commission's recovery package to support the manufacturing
industry in the development of new and sustainable technologies. The objective is
to help EU manufacturing enterprises, in particular SMEs, to adapt to global
competitive pressures by improving the technological base of EU manufacturing
across a broad range of sectors.24

ARTEMIS Embedded Computing Systems (ARTEMIS)25

AAL - The Ambient Assisted Living Joint Programme

ERC - The European Research Council
23
http://cordis.europa.eu/fp7/ict/micro-nanosystems/docs/fof-smart-factories-factsheetwp2011_en.pdf
24
http://ec.europa.eu/research/industrial_technologies/factories-of-the-future_en.html
25
www.artemis-ju.eu/
Position paper on Robotics and Cognitive Systems
19
Its main aim is to stimulate scientific excellence by supporting and encouraging the
very best, truly creative scientists, scholars and engineers to be adventurous and
take risks in their research. The scientists are encouraged to go beyond established
frontiers of knowledge and the boundaries of disciplines. The ERC complements
other funding activities in Europe such as those of the national research funding
agencies, and is a flagship component of the 'Ideas Programme' of the European
Union's Seventh Research Framework Programme (FP7).
Being 'investigator-driven', or 'bottom-up', in nature, the ERC approach allows
researchers to identify new opportunities and directions in any field of research,
rather than being led by priorities set by politicians. This approach ensures that
funds are channeled into new and promising areas of research with a greater
degree of flexibility.
D ANISH COMPETENCES AN D INTERESTS
Since Denmark is a small research nation „strength positions‟ are determined by
other factors than the amount of clustering research groups.
Danish research groups are engaged in various technologies, each research group
with different foci and disciplinary combinations. Competences and interests are
primarily on sensing and perception, learning, motion planning, locomotion,
navigation, surveillance and system integration 1) systemic integration of sub
systems into whole production systems, 2) multi agent software systems for care
services for example. Most of these sub areas ads up to the development of
improving the human-robot interaction.
Research activities and competences are to some degree divided between the
universities involved.
At Aalborg University research on process and system integration for the purpose of
future production methods is dominant. Research on design of whole flexible
production systems happens „close to industry‟. Research in human-robot interaction
with specific focus on decoding of human intensions social intelligent technology is
also being carried out.
At University of Southern Denmark research clusters around agricultural robotics,
motion planning and cognitive vision systems. The approach is here „close to
industry‟.
Basic research in alternative physic constructions of modular robot bodies that are
less sensitive, more robust and less injuring is far ahead.
At Aarhus University robots for the purpose of „Play and Learning‟/edutainment is
dominant. Focus is to challenge everyday people on their comprehension of
technology through experience technology be it art, toys or other. A close
collaboration with LEGO is going on.
At Technical University of Denmark the Section for Cognitive Systems develops
information processing between man and computer, with a particular focus on the
signals they exchange – audio, imagery, behavior – as well as the opportunities
these signals offer for modeling and prediction. The research is based on statistical
machine learning and signal processing, on quantitative analysis of digital media
and text, on mobility and complex networks, and on cognitive psychology. These
technologies for comprehension of human behavior are interesting perspectives on
the human-robot interaction.
Position paper on Robotics and Cognitive Systems
20
Robot steering, navigation and sensor fusion for autonomous systems is also a
dominant part of research at DTU.
The Center for Visual Cognition (CVC) at the Department of Psychology of the
University of Copenhagen is currently a unit (Center for Integrated Visual Attention
Research) under the Center of Excellence Programme of the University of
Copenhagen. CVC is a world-leading research unit in the study of human visual
cognition. The research at the Center comprises both experimental and theoretical
studies, including the development of mathematical models of human visual
perception, attention, and short-term memory.
To visualize the competences on the universities following scheme is made:
Aalborg University
University of Southern
Denmark
Technical University of
Denmark
Cognition
Sensing
and
perception
Motion
planning
Navigation
Surveillance
System
Integration
GOOD
GOOD
GOOD
GOOD
GOOD
GOOD
EXCELLENT GOOD
EXCELLENT GOOD
OK
GOOD
EXCELLENT GOOD
GOOD
OK
OK
GOOD
Aarhus University
Play and
Learning
GOOD
GOOD
Copenhagen University
-
EXCELLENT EXCELLENT
OK = the university has competences, but no major focus.
Good = the university has projects and researchers in the area.
Excellent = the university has world class research and projects in the area.
Blank = none or only little research.
Since research in welfare technology is so highly profiled and yet still on an early
stage many Danish initiative such as educations and research projects are being
launched. These efforts will gather and get to be coordinated during the next years.
Regardless of the purpose be it welfare or production, robots need cognitive abilities
and social awareness and applications areas in DK are developed enough to use
cognitive systems for welfare, agriculture and industry.
POSSIBLE BREAKTHROUGHS
Whether it regards cellular robotics or cognitive vision researchers are cautious in
predicting future break throughs.
A breakthrough though could happen through extensive collaboration between
human sciences and robotics. Comprehension of context can be transferred to
technological sciences via mathematic algorithms and push research in cognitive
robotics further.
In regard to modular robots break throughs can happen in the redefinition of
flexible production systems, entertainment/toy industry or for the development of
welfare robots not physical dangerous to humans.
Position paper on Robotics and Cognitive Systems
21
Integration of sub technologies into system architecture are expected but still not
realized. Serious breakthrough will happen when open standards are demanded and
the integration of sub technologies is possible.
Technology for the farming industry has matured the last 10 years in a way that has
created great interest from farmers and gardeners.
Sustainable production methods regardless of application area will come in focus.
But demographic and environmental challenges has an obvious push effect on
research since they both demand fundamental changes in our healthcare system
and food production. This will dominate the agenda years to come.
S WOT A N A L Y S I S O F T H E N A T I O N A L P O S I T I O N S I N R O B O T I C S
STRENGTHS
Many unic robotic solutions are
integrated in SMEs
RoboCluster – research cluster as a
platform for national coordination and
knowledge transfer to the industry.
Initiator of cross-disciplinary and
sectoral research projects.
Modular robot technology
Cognitive vision systems
System architecture - holistic approach
for developing new production methods
and welfare technology
GTS‟ (Governmental Technological
Services) secures impact of the
collaboration between universities and
companies
Non-hierarchic structure in society
provides good premises for innovation.
Denmark is a test-willing country with a
population open to use of new
technologies.
WEAKNESSES
Foundations and councils does not
encourage sufficiently to national
collaboration.
No distinct educational profile
divided between the universities nor
research groups
Not too many competitive core
competences. Forces lie in
scientific/interdisciplinary approach
and non-hierarchic culture
Conditions connected to the small
size of Denmark
OPPORTUNITIES
Increased familiarity and coordination
among research groups on the five
engaged universities.
Split of research into either basis
research or fast applicable research
developed with eye for industrial
stakeholders, new national production
or export/sale
Emerging market: Novel robotic
controlled sustainable production
systems
Safe human-robot interaction based on
cognitive abilities
THREATS
Slow readjustment among suppliers
(from industry to welfare
technology)
Still a missing link between SME‟s
and researchers
Lack of coordination can lead to less
competitive core competences
Poor conditions for entrepreneurship
Position paper on Robotics and Cognitive Systems
22
NATIONAL STRATEGIES THAT SUPPORT T HE RESEARCH
Solid financial support is given on a regular basis to Governmental Technological
Services (GTS) to secure transfer of research knowledge to Danish companies and
convert it into actual development and integration of intelligent technological
solutions.
Main focus in Danish research strategies according to the three main funds are
predominantly on usage/integration of robotics in to real life and using robots as
agents for problem solving and discovering new ways of meeting societal needs.
Danish National Advanced Technology Foundation 26, is supportive of new
technological solutions for the international market, development of new markets,
big business potential, research and innovation with emphasize on interplay
between business and research partners.
The Danish Council for Technology and Innovation supports corporation and
dissemination of knowledge between researchers, research and educational
institutions, technological service institutes and companies.
The Danish Council for Production and Technology and Business Innovation Fund
are also relevant foundations where the aim is to promote growth, employment and
export by supporting business opportunities within green growth and welfare.
Nevertheless there are critics from the completed interviews pointing at:



The lack of focus on cognitive systems in DK
The lack of calls who gather researchers on the national scene
The lack of focus on „post-prototype phase‟
R ECOMMENDATIONS /C ONCLUSION
In the coming year‟s national initiatives will focus heavily on environmental
sustainability, food and food production, health and new ways to future welfare
services.
This correlates well with EU strategies and prioritizations.
On these grounds and on the basis of Danish interests and competences
RoboCluster suggest that Danish research will be able to contribute considerably to
competitive European development in mainly two applications areas:
1. Agriculture and food processing
2. Welfare technologies/care support for domestic and public use
Ad 1)
In the effort to meet demands on green production Denmark, as a high tech
developed agricultural country, has great focus on new sustainable production
methods. Many high tech research projects are launched and completed, and
demands from farmers and gardeners on intelligent solutions are great.
26
http://hoejteknologifonden.dk/
Position paper on Robotics and Cognitive Systems
23
Ad 2)
Due to good national funding possibilities, great political attention and will to
maintain welfare services Denmark is on level with technological development with
priority to multidisciplinary collaboration with professionals and managements from
the care sector. National funding is given to studies focused on for instants
pervasive computing for elderly care support or chronic patients in own home.
The approach considers social perspectives on technological care support in a
holistic perspective where psychological effects are studied as another angle to the
development of man-machine interaction.27
Regardless of the recommended application areas researchers have pointed out the
following cases to be of great importance in connection to future research in the
field of cognitive systems and robotics:
 In order to increase focus on the very applications it is necessary to focus on
whole systems and total integration and by doing so break frontiers into new
application areas or markets. Part solution should be replaced by focusing either
on basic research or total architecture of systems. This prioritization should be
backed up politically by foundations, councils and national coordination between
the universities.
 As in a European context Danish research effort also needs a more applicable
focus to connect closer with industrial stakeholders and society as such. Danish
research is ‟good enough‟ on R & D but needs to be focused further on „post
prototype phase‟ with testing, usage in new fields of applications and
commercialization. – Poor possibilities of Danish funding contradict Denmark
being a test willing country.
 Councils such as European Research Council 28 are popular among Danish
researchers due to manageable administration of the funding in contrast to
funding from the 7th framework. Critics about call texts that are too broad and
too general is put forth and seen as part of why funding can be hard to achieve.
 New cross-over research is called for to study better interaction and intimacy
between technologies and humans as such. The pedagogic perspective in
regard to user interface, robots for education, robots for entertainment, and
thereby reaching greater confidentiality with technology and interactive
systems.
 For robots to break-through barriers to new application areas, e. g. the elderly
care sector, user-friendly robots and interface should be developed based on
studies of learning processes, work culture and product psychology.
 On a national research level there is a wish among researchers to do more
cross- over research in regard to development of future cognitive systems and
27
www.partnerskabetunik.dk & www. Intellicare.dk
The ERC is established by the European Commission and funded through the EU 7th Research
Framework Programme
28
Position paper on Robotics and Cognitive Systems
24
robotics. This should be possible and still pay attention to core competences. As
recommendation for new national prioritization we suggest a platform to unit
national research relevant to cognitive robotic systems for close interaction with
humans. This could be a Cognitive Research Center multidisciplinary in its
approach and we believe such a multidisciplinary research center could
contribute in a unique way to European Research Area providing alternative
answers to learning aspects and safety.
Artificial cognitive systems are at the junction of the cognitive, ICT, and natural
sciences:
Design:
National Center
for Applied
Cognitive
Robotics
Technological
Product psychology
Sciences:
Pedagogic interface
Robotics and ICT
Human sciences:
Neuroinformatics
Psychology
Semantics
EDITORS
Marianne Sandgaard, Project Coordinator at RoboCluster
Bjarke Hassert Nielsen, Cluster Manager at RoboCluster