Commission Européenne, présentations de la Délégation générale Société de l’Information à la délégation d’Aquitains conduite par AEC, 30 janvier 2012: robotique
1. Soutien européen à la robotique
Olivier DA COSTA
Project Officer
Commission Européenne
Direction générale Société de l‘Information et Médias
Unité E5 – Systèmes Cognitifs Systems, Interaction,
Robotiques
Olivier.da-costa@ec.europa.eu
••• 1
2. La robotique : Technologie essentielle
pour l’Europe
Marché en forte croissance, porteur d’emploi:
– robots industriels, de service, ménager: croissance attendue de
60% jusqu’à 2015
– tiré par une demande croissante
• Productivité, qualité, qualité de travail, …
• Besoins sociétaux: santé, démographie, sécurité, ..
– alimenté par le développement technologique
• plus “intelligents”, adaptifs, robustes, intuitifs, moins chers…
Pas d’acteurs industriels dominant (jusqu’à présent)
Technologie déterminante pour l’avenir du secteur ICT
– Espaces intelligents, interfaces intuitives, apprenant par
apprentissage…
Technologie essentielle pour le maintien de la production en
Europe
••• 2
3. Participation en Robotique et Systèmes cognitifs
ETATS-MEMBRES
Participation dans
les projets européens
1 GERMANY 27,4%
2 UNITED KINGDOM 17,3%
3 ITALY 12,4%
4 SWITZERLAND 6,2%
5 FRANCE 5,6%
6 SWEDEN 4,9%
7 SPAIN 3,7%
8 BELGIUM 3,2%
9 NETHERLANDS 3,2%
10 AUSTRIA 3,1% 3
•••
4. Objectif “Cognitive Systems & Robotics”
Trois types de soutien
Recherche:
Fondamentale
RDI industrielle
Réseautage & renforcement de la
communauté
Collaboration plus efficace entre
recherches académique et industrielle
••• 4
5. Objectif “Cognitive Systems & Robotics”
► About 150 funded projects Portefeuille de projets
UNDERSTANDING
► About 100 ongoing projects • Recognising
► With about 1000 • Interpreting
participants UNDERSTANDING • Adapting
• Planning
► Total EU funding over 500 • Modelling
M€ • Cognitive architectures
COGNITIVE
PERCEIVING SYSTEMS
PERCEIVING ACTING
• Touching &
• Seeing ROBOTICS
• Hearing
• Distributed sensing
• Advanced sensing
ACTING
LEARNING • Manipulating
APPLICATION AREAS • Navigating
•Aerial • Interacting
•Underwater • Collaborating
•Industry and manufacturing • Monitoring
•Professional and domestic services
•Medical and rehabilitation
•Monitoring and surveillance
••• 5
http://cordis.europa.eu/fp7/ict/cognition/projects/areas-projects_en.html
6. Objectif “Cognitive Systems & Robotics”
Réseautage & renforcement de la
communauté
EUROP - European Robotics Technology
Platform
EURON - EUropean RObotics research
Network
euRobotics - European Robotics Coordination
Action
EUCogIII - European Network for the
Advancement of Artificial Cognitive Systems,
Interaction and Robotics
+ Echord - European Clearing House for Open
Robotics Development
••• 6
7. Objectif “Cognitive Systems & Robotics” - Call 9
OBJECTIVE: 2.1 Cognitive Systems and Robotics
PUBLICATION: 18/01/2012
DEADLINE: 17/04/2012
INDICATIVE BUDGET: 82 M€ (80M€ (STREPS + IPs) – 2M€CAs)
• Target (b) Cognition and control in complex
systems - STREPs + IPs
• Target (c) Gearing up and accelerating cross-
fertilisation between academic and industrial
robotics research - IPs
• Target (e) Speeding up progress towards smarter
robots through targeted competitions - CAs ••• 7
9. Les chiffres
Les robots industriels
– 2010: ± 118,000 nouvelles unités vendues,
• ~ $5.7 milliards, +97% comparé à 2009
– Marché mondial estimé à $17.5 milliards
(2011)
• Croissance de 40% d’ici 2014
Les robots de service professionnel
– Croissance de 7% en 2010 et de 60% d’ici
2014
– Marché mondial estimé à $16 milliards
• (± 90,000 unités sans les robots ménagers et
de loisir)
Les robots ménagers
– Croissance de 26% en 2010
– > 60% entre 2011et 2015
••• 9
10. Call 9 - Target (b): Cognition and control in
complex systems (STREPs + IPs)
• Acquisition and application of cognitive
capabilities (e.g., perception,
conceptualisation, reasoning, planning)
• Enhance performance and manageability
of complex multi-component and multi-
degree-of freedom artificial systems
• Synergies: cognitive systems - systems
control engineering
••• 10
11. Call 9 - Target (c): Gearing up and accelerating
cross-fertilisation between academic and
industrial robotics research (IPs)
• Joint industrially-relevant scenarios
• Shared research infrastructure
• Experimentation with industrial
platforms
• Benchmarking
••• 11
12. Call 9 - Target (e): Speeding up progress towards
smarter robots through targeted competitions
(CAs)
GOAL –> tool to support science
• Not just ‘nuts-and-bolts’ engineering
Objective Comparison of results
Measure & Ensure progress
Share results
VISIBILITY
••• 12
Notes de l'éditeur
Industrial robotics companies today derive much of their revenues from the automotive manufacturing sector, where robots perform tasks such as material handling, spot-welding, assembly and painting. Overall, there are 1 million industrial robots installed worldwide, valued at about $30 billion. In 2010, 118,000 new units were sold worth $5.7 billion, an increase of 97% compared to 2009. The world market for robotic systems is estimated to be about $17.5 billion in 2011. The sales of industrial robots are expected to increase by 17% this year and by a further 40% until 2014 and are expected to expand further to other manufacturing sectors. European robot manufacturers are strong in industrial robotics with a market share of about 25% in newly sold systems
Research Rationale By promoting research into systems that have cognitive functions normally associated with people or animals and which exhibit a high degree of robustness in coping with unpredictable situations, we seek to overcome limitations of today's computers, robots, and other man-made creations to handle simple everyday situations with common sense and to work without pre-programming in natural surroundings, while maintaining and possibly improving the quality of their services. Unit Mission We support research on the development and construction of robotic systems and other artificial cognitive systems than can process and interpret various kinds of sensor data (images, speech and other forms of sensor data), and act purposefully and autonomously towards achieving goals. Such systems are to operate in dynamic real-life environments and are to be capable of responding timely and sensibly to gaps in their knowledge and to situations that have not been anticipated at design time. These systems should learn and develop through individual or social interaction with their environment. The work provides an enabling technology that applies across domains such as industrial and service robotics, automation, image recognition, systems monitoring and control, automated reasoning and decision support, and natural language understanding. The work could furthermore borrow insights from the bio-sciences, and yield innovative insights about perception, understanding, interaction, learning and knowledge representation to further underpin progress towards engineering systems with the desired capabilities.
Industrial robotics companies today derive much of their revenues from the automotive manufacturing sector, where robots perform tasks such as material handling, spot-welding, assembly and painting. Overall, there are 1 million industrial robots installed worldwide, valued at about $30 billion. In 2010, 118,000 new units were sold worth $5.7 billion, an increase of 97% compared to 2009. The world market for robotic systems is estimated to be about $17.5 billion in 2011. The sales of industrial robots are expected to increase by 17% this year and by a further 40% until 2014 and are expected to expand further to other manufacturing sectors. European robot manufacturers are strong in industrial robotics with a market share of about 25% in newly sold systems
Overhead by infrastructure that had to be built? cooperation with SMEs tend to be more successful academia-driven cooperation tend to be more successful (see US experience)
CHALLENGES better than COMPETITIONS Important to run the competition several times. This should be a firm requirement. Careful about over-fitting on data-sets What are the measures? Tuning of the level of difficulty is critical: it needs to be hard to connect to deep issues, but not too hard to be relevant to current methodologies. - Pascal was a breakthrough and changed the field when introduced. Now it is deteriorated to a stage of adding of a feature to increase performance by Epsylon Challenges increase visibility of the field and give the opportunity to small players to be highlighted Regardless of the difficulty, you should take a risk. You can never predict the outcome.