4.2 Challenge 2: Cognitive Systems, Interaction, Robotics
Engineering systems with the capability to sense and understand an unstructured environment is a challenge which goes beyond today's systems engineering paradigm. Present day systems engineering relies on specifying every eventuality a system will have to cope with in the execution of its task(s), and programming the appropriate response in each case. With the abundance of ever cheaper, smaller sensors, actuators and wireless tranceivers that link systems to the real world and with other systems, this approach faces serious limitations:
- The real world is generally too nuanced, too complicated and too unpredictable to be
summarised within a limited set of specifications; there will inevitably be novel situations
and the system will always have gaps, conflicts or ambiguities in its own knowledge and
capabilities.
- Even in situations where unpredictable events are less likely, the problem of extracting
meaning and purpose from bursts of sensor data or strings of computer code arises,
because we don't have a formalisation of information processing that embodies semantics.
Challenge 2 aims to extend systems engineering to the design of systems that can carry out
useful tasks (e.g. manipulation and grasping, exploration and navigation, monitoring and
control, situation assessment, communication and interaction), autonomously or in
cooperation with people, in circumstances that were not planned for explicitly at design time.
Specifically, such systems should be:
- more robust: performance should not degrade when they are presented with unexpected data;
- more adaptive: performance should be open (within reasonable constraints) to changing
service requirements, without the need for extensive human intervention;
- more effective: performance should improve because they can predict or anticipate what
might happen at some point in the future, near or far;
- more natural: performance should be tolerant to the ambiguity and uncertainty that is a
consequence of dealing with humans, and performance should improve with time.

System capabilities in dimensions such as deliberation and learning, and innovation and
creativity, would appear to be necessary to meet this aim. This clearly calls for design that
shares some characteristics with the higher-level cognitive processes of the brain. For the
purposes of this work programme a cognitive system can cope with the uncertainty (in the
system's environment) that makes robust and adaptive performance difficult to achieve. It
should also be borne in mind that it makes no sense to speak of robustness or adaptability
without first specifying the requirements of interest: a robust lawnmower is different to a
robust operating system or a robust planner.
Research and development efforts should aim at generating actual design principles. They
will contribute to establishing scientific foundations for such principles. Alternatively, they
may aim to achieve significant engineering progress, e.g. through integration.
Manufacturers of robots of all sorts, autonomous vehicles, smart cameras and sensor networks
will benefit from R&D efforts. Europe has strong manufacturing capabilities and a significant
share of world market revenues in these sectors. The emergence of service robots and vision
systems that operate outside structured manufacturing environments offer added opportunities
for market expansion.
Likewise automated machine translation stands to profit from more robust and adaptive
methods for natural language understanding. With 23 official languages, the EU is at the
forefront of multi-lingualism and it would be unrealistic to assume that the lingua franca in
machine translation is, or will remain, English. A strategic challenge for Europe in today's
globalised economy is to overcome language barriers through technological means.
Technologies developed under this Challenge are expected to be tailored to meet key societal
and economic needs.

FP7 Cooperation Work Programme: Information and Communication Technologies

4.2 Challenge 2: Cognitive Systems, Interaction, Robotics

Engineering systems with the capability to sense and understand an unstructured environment is a challenge which goes beyond today's systems engineering paradigm. Present day systems engineering relies on specifying every eventuality a system will have to cope with in the execution of its task(s), and programming the appropriate response in each case. With the abundance of ever cheaper, smaller sensors, actuators and wireless tranceivers that link
systems to the real world and with other systems, this approach faces serious limitations:
- The real world is generally too nuanced, too complicated and too unpredictable to be summarised within a limited set of specifications; there will inevitably be novel situations and the system will always have gaps, conflicts or ambiguities in its own knowledge and capabilities.
- Even in situations where unpredictable events are less likely, the problem of extracting meaning and purpose from bursts of sensor data or strings of computer code arises, because we don't have a formalisation of information processing that embodies semantics.

Challenge 2 aims to extend systems engineering to the design of systems that can carry out useful tasks (e.g. manipulation and grasping, exploration and navigation, monitoring and control, situation assessment, communication and interaction), autonomously or in cooperation with people, in circumstances that were not planned for explicitly at design time.
Specifically, such systems should be:
- more robust: performance should not degrade when they are presented with unexpected data;
- more adaptive: performance should be open (within reasonable constraints) to changing service requirements, without the need for extensive human intervention;
- more effective: performance should improve because they can predict or anticipate what might happen at some point in the future, near or far;
- more natural: performance should be tolerant to the ambiguity and uncertainty that is a consequence of dealing with humans, and performance should improve with time.
System capabilities in dimensions such as deliberation and learning, and innovation and creativity, would appear to be necessary to meet this aim. This clearly calls for design that shares some characteristics with the higher-level cognitive processes of the brain. For the  urposes of this work programme a cognitive system can cope with the uncertainty (in the system's environment) that makes robust and adaptive performance difficult to achieve. It should also be borne in mind that it makes no sense to speak of robustness or adaptability without first specifying the requirements of interest: a robust lawnmower is different to a robust operating system or a robust planner.

Research and development efforts should aim at generating actual design principles. They will contribute to establishing scientific foundations for such principles. Alternatively, they may aim to achieve significant engineering progress, e.g. through integration. Manufacturers of robots of all sorts, autonomous vehicles, smart cameras and sensor networks will benefit from R&D efforts. Europe has strong manufacturing capabilities and a significant share of world market revenues in these sectors. The emergence of service robots and vision systems that operate outside structured manufacturing environments offer added opportunities for market expansion. Likewise automated machine translation stands to profit from more robust and adaptive methods for natural language understanding. With 23 official languages, the EU is at the
forefront of multi-lingualism and it would be unrealistic to assume that the lingua franca in machine translation is, or will remain, English. A strategic challenge for Europe in today's globalised economy is to overcome language barriers through technological means.
Technologies developed under this Challenge are expected to be tailored to meet key societal and economic needs.

Objective ICT-2009.2.1: Cognitive Systems and Robotics
Target outcomes
a) New approaches towards understanding and solving key issues related to the engineering of artificial cognitive systems – see above; among these issues are the following:
- representation / categorisation / recognition / interpretation of objects, events, situations, behaviours and affordances in realistically scaled real-world environments;
- the role and implementation of memory and learning in artificial systems;
- adaptive and anticipatory behaviour within incompletely specified environments;
- goal-setting and strategies for achieving goals;
- collective behaviour arising from the interplay of (possibly large numbers of) individual subsystems;
- modelling and design of (multimodal) interaction, communication and collaboration.
Projects are expected to demonstrate measurable progress on a suitable mix of these issues.
b) New approaches towards endowing robots with advanced perception and action capabilities, and towards developing pertinent benchmarks and tests. Of particular interest are:
- 3D sensing for everyday objects and environments;
- motion and affordance perception;
- learning and control strategies for linking perception and action;
- benchmarking with a focus on navigation and autonomy.
Projects are expected to demonstrate measurable progress on at least one of these issues.

Expected impact for a) and b)
• Leading-edge research capacity in Europe in cognitive systems engineering and robotics.
• Innovations in service robots, and industrial production and manufacturing processes.
• Widespread comparative assessment of robot performance (for different tasks and technologies).
• New market opportunities, and technologies for increased productivity and efficiency in EU industries.

c) New ways of designing and implementing complete robotic systems that operate largely autonomously in loosely structured dynamic environments and, where necessary, in close co-operation with people. Systems may be distributed and should integrate rich sensory-motor skills (for example, grasping, manipulation, locomotion) with high level
cognitive competencies (for example, reasoning, planning and decision-making). As appropriate, they should be demonstrably more robust, dependable, flexible and adaptive, and safer than it is possible today, and improve their performance through learning.

d) New, scientifically grounded system architectures integrating communication, control, and cognitive capabilities to enable meaningful and self-sustaining autonomous action in real-world environments, natural interaction with people (where necessary), robust adaptation to changing operating conditions, and self-improvement. The viability and scalability of these architectures will be demonstrated through suitable experiments based on physical implementations and/or simulations of complete systems.

e) A framework to facilitate cross-fertilisation between academic and industrial research efforts in robotics through widespread experimentation with industry-strength platforms in academic research labs and through the joint definition of longer term scenarios and requirements to direct robotics research towards common goals; to assure a comparative assessment of performance through definition of suitable metrics and through benchmarking (supported by competitions or otherwise).

Expected impact for c), d) and e)
• Integrated and consolidated scientific foundations for engineering cognitive systems under a variety of physical instantiations.
• Significant increase of the quality of service of such systems and of their sustainability in terms of, for instance, energy consumption, usability and serviceability, through the integration of cognitive capabilities.
• Innovation capacity in a wide range of application domains through the integration of cognitive capabilities.
• Improved competitive position of the robotics industry in existing and emerging markets for instance in the following sectors: flexible small scale manufacturing;
professional and domestic services; assistance and rehabilitation; construction, maintenance and repair; urban search and rescue; exploration and mining;
entertainment, education and training.
• Consensus by industry on the need (or not) for particular standards. More widely accepted benchmarks. Strengthened links between industry and academia. (especially (e)).
Research and development pertaining to targets (a), (b), (c) and (d) will be guided by
demanding, yet pragmatic, application scenarios. Target environments may be, for example, difficult terrains, buildings, homes, public spaces, shop floors, power plants and other technical infrastructures. Functionalities include: exploration, monitoring, controlling all sorts of sensors and actuators and communication and interaction with people (also including
advanced human-robot interaction).
The applicability of research results is expected to go beyond the scenarios through which they have been obtained. Proposals strictly focusing on applications that are targeted under Challenges other than Challenge 2 are not eligible under Challenge 2.
Pertinent research may be informed by neuro- and behavioural sciences and determine the requirements basic technologies have to meet in order to enable creating the targeted systems.
Systems may for instance employ new sensor and sensor networking technologies or 'intelligent' materials to enhance their functionality, performance, and efficiency of resource usage, and bring new functionalities, like self-configuration and self-repair, within reach of FP7 Cooperation Work Programme: Information and Communication Technologies
industrial realisation. Research will also significantly broaden the remit of machine learning and put stronger emphasis on intelligent process control in real-time.

f) A 'Virtual Institute' integrating diverse research areas whose problems, techniques and solutions need to be brought together to understand cognitive systems and design useful new ones; they will develop a requirements- and capability-led understanding of cognitive systems that can be applied across multiple engineering and application domains.

Expected impact for f)
• Leading-edge research in Europe in cognitive systems engineering and robotics.

g) Co-ordinated co-operation and communication within a multidisciplinary robotics community in Europe, with concomitant outreach to potential users of robotic systems h) Co-ordinated co-operation and communication within a multidisciplinary artificial cognitive systems research community in Europe, with concomitant outreach to potential industrial applications.
Expected impact for g) and h)
• Stronger cohesion among relevant communities; awareness built among wider (including non-professional) audiences of the potential of the technologies at issue.
Where and as appropriate, activities under this objective, and in particular those aiming at targets g) and h), are expected to contribute to a better understanding of the ethical, social and socioeconomic issues related to the design, deployment and operation of robotic and cognitive systems.

Funding schemes
a)-b): STREP; c)-e): IP; f) NoE; g)-h) CA
Indicative budget distribution10
EUR 153 million
Calls:
• ICT call 4: target outcomes (b), (d), (f), (g)
- IP/STREP: EUR 65 million of which a minimum of 50% to IPs and a minimum of 30% to STREPs
- NoE: EUR 6 million
- CA: EUR 2 million
• ICT call 6: target outcomes (a), (c), (e), (h)
- IP/STREP: EUR 78 million of which a minimum of 50% to IPs and a minimum of 30% to STREPs
- CA: EUR 2 million