Embedded Systems Technology Platform First meeting of the Working Group on Application Drivers Abstract These minutes detail the discussions held at a meeting of the Working Group on Application Drivers for the Embedded Systems Technology Platform (ESTP). This meeting involved the Commission and representatives of industry and research to develop ambitious application scenaria that could motivate the future development and integration of embedded systems technologies. Contents Contents... 1 List of attendees... 2 Summary of the meeting... 2 Morning Plenary Session... 3 Main Points from the Morning Plenary Session... 5 Afternoon Plenary Session... 6 Private Infrastructure working group... 6 Mobile Objects... 6 Big Systems... 6 Major Infrastructure... 7 Actions arising and timetable... 8 Conclusions... 8 First meeting ESTP WG on Application Drivers 1/9
List of attendees Name Organisation Country Kostas Glinos European Commission Alkis Konstantellos Europan Commision Marcella Miano European Commission Emile Philips Research NL Hugo De Man IMEC BE Kari-Pekka Estola Nokia FI Adam Wolisz TU Berlin DE Dietmar Schütz Siemens DE Iñaka Larrañaga MCC-Mondragon ES Martin Hurich Bosch DE Harald Hönninger Bosch DE Thomas Kirste Fraunhofer IGD DE François Pilarski Airbus FR Niels Leffler ABB CH Gérard Fargère ST Microelectronics CH Jean-Pierre Lacotte ITEA Oliver Frick SAP DE Hermann Von Hasseln DaimlerChrysler DE Jean-Pierre Desbenoit Schneider Electric FR Sjir van Loo Philips Research NL Francis Wray Rapporteur Summary of the meeting The purpose of the WG is to develop ambitious application scenaria that could motivate the future development and integration of embedded systems technologies. These application drivers should have a visionary, medium to long-term perspective, should raise a number of important technological challenges and should ideally correspond to future product and service markets that are expected to exhibit fast growth rates due to socio-economic trends. The meeting was hosted by the European Commission in room 0/54 of Beaulieu 33. Dr took the chair. The meeting began at 10:30 am and closed at 4:30 pm. The following agenda was agreed: Opening and Introduction (E ) The European Technology Platform for Embedded Systems (K Glinos) The Working Group Application Drivers (E ) The Role of Applications in Embedded Systems Design (H DeMan) Identification of Criteria for Application Classification (Plenary Discussion) Identification of Application Domains (Plenary Discussion) Application Classification (Working Group Discussion) Working Group Results and Plans (Plenary Presentations) Follow-up actions and Timetable Conclusions First meeting ESTP WG on Application Drivers 2/9
Morning Plenary Session Professor welcomed the participants from industry, research and the Commission. He reminded the participants of the objectives of the meeting described above. He also informed the participants of the outcome of this and similar meetings which would be a high-level document containing strategic objectives and the vision of the platform. This would be circulated to CEOs of appropriate companies for their signed approval. This document would be available in the second half of June. A short tour de table to introduce the participants was held. Mr Glinos made a presentation discussing the aim, scope and organisation of the European Technology Platform for embedded systems. This presentation is attached as an appendix to these minutes. Professor made a presentation discussing the aim, scope and organisation of the Working Group for Application Drivers. This presentation is attached as an appendix to these minutes. Professor DeMan made a presentation showing the role of applications in embedded systems design. This presentation is attached as an appendix to these minutes. The following discussion was stimulated by these presentations and covered the agenda items Identification of Criteria for Application Classification and Identification of Application Domains (Plenary Discussion) The report of the ISTAG Working Group on "Grand Challenges in the Evolution of the Information Society" presents 12 Grand Challenges which stretch the vision of what is needed in the area of Embedded Systems. There are enough visions, inventories and other material to work on. The ISTAG report of 2003 presents a vision to address the problems Europe in facing. In particular, ISTAG has identified the need for Experience and Application Research Centres. These vertical integrations of activities try to establish innovation where endusers are involved at an early stage. The Museum of the Future is an example of an initiative which presents future technology directly to the public. The Experience Centre from the automobile industry is an established concept to gain feedback from the public on proposed innovations. Estola Clarification is needed regarding the remit of this Working Group. What are the relationships to other proposed platforms such as nano-technology? The purpose of this Working Group is to put Embedded Systems into a general context, by relating applications to technology. The presentation by Professor DeMan gave a general framework in which to present applications. The diagram which relates technology to application domains is a good example of what this group should be developing. Wolisz Agriculture is an important potential application. Intelligent buildings are an important application which complements the Home and Office applications in the presentation of Professor DeMan. Monitoring of constructions such as buildings and life-cycle monitoring in automobiles are also important applications. Three important activities can be derived from applications: Data collection leading to meta-data generation; Alarm monitoring; First meeting ESTP WG on Application Drivers 3/9
Object tracking. Van Loo It is important to assimilate the challenges from all potential areas, before any are dropped from consideration. DeMan Distributed sensing and actuating is an important domain. Frick Different aspects for industry such as costs and manufacturing issues and for consumers, such as their feedback, need to be taken into account in setting the visionary scenarios. Von Hasseln It should be borne in mind that automotive systems are very different from those in aerospace. Both have very different design methodologies. We should try to group applications so as to have four or five representative clusters. Lacotte We should consider the main differences between the automobile and the aerospace sectors. A domain-based approach (i.e. where the application is used) is preferable to a sector-based approach. Fargère Applications should be the driver for this initiative. What is the timeframe for applications? DeMan The timescales for FP7 mean that would start research in 2007 on aspects which would be exploited between 2010 and 2014. We have to make guesses now on what will be important. However it is hard to convince industry to follow this approach. Fargère Industry looks at timeframes and makes decisions about research or about setting up production lines. The semiconductor industry needs to give input on what are reasonable goals. Application drivers tell the semiconductor industry what is needed to be developed. Glinos Visionary applications stimulate the development of technologies which may be useful in unpredictable ways. It is not necessary to realise the visionary applications per se. We are trying to create the future rather than predict it. The Wellness area will be a consumer-oriented application which has many requirements. There are four classes of application relevant to embedded systems: Big systems from the aerospace, automobile, medical and manufacturing sectors; Mobile objects such as animals and people; Private infrastructure such as homes, houses and offices; Major infrastructure such as airports, cities and highways. First meeting ESTP WG on Application Drivers 4/9
We need to classify visionary applications carefully. Classification and mapping could take a long time. Wolisz The scenario approach may be more useful for developing visionary applications. For example we could envisage a medical scanner as a complex system or similarly the movement of people or livestock. We need to set visions and then identify what steps need to be taken towards those visions. Fargère Application drivers should stimulate technology development. Applications need to be ambitious, challenging developments. We need to set targets which direct developments. Visionary applications are good PR because they can be widely understood. A listing of material already available needs to be brought together. The industries represented at this meeting need to present what the challenges are for them over the next five years. In particular what do these challenges mean for mobile communications and what do these challenges imply for technology developments. Pilarski The aerospace industry covers all the four domains presented by Professor. An issue is to make these four domains cooperate together. The following four domains (chairmen shown in parenthesis) are established for the working group sessions: Big systems from the aerospace, automobile, medical and manufacturing sectors (Leffler); Mobile objects such as animals and people (Estola); Private infrastructure such as homes, houses and offices (van Loo); Major infrastructure such as airports, cities and highways (Wolisz). Main Points from the Morning Plenary Session The following main points arose during the morning session: The outcome of this and other meetings will be a high-level document containing strategic objectives and the vision of the Embedded Systems Technology Platform. This document will be produced towards the middle of June; There is a significant body of existing material on which to base these strategic objectives and vision; The purpose of this working group is to put Embedded Systems into a general context by relating applications to technology. Professor DeMan s diagram relating technology to application domains represents this approach well; The timescales for FP7 mean that research will start in 2007 on areas which would be exploited between 2010 and 2014. Visionary applications stimulate the development of technologies which may be useful in unpredictable ways. It is not necessary to realise the visionary applications per se. First meeting ESTP WG on Application Drivers 5/9
Applications need to be ambitious, challenging developments. We need to set targets which direct developments. Visionary applications are good PR because they can be widely understood. Afternoon Plenary Session The afternoon plenary session began with presentations from each of the four working groups established during the morning. Private Infrastructure working group Dr van Loos presented the results from the Private Infrastructure working group. The vision is of systems which know and recognise individuals and can adapt to their requirements and to longer-term changes in those requirements. Such systems will need significant computational, communication and data-storage capabilities. They will be characterised by ubiquitous sensors and actuators and a high-bandwidth connection to the rest of the world. Technologies will need to be developed which support sensing and tracking, ergonomics and ease of use, security and comfort and multi-modal interaction. Key to this will be the development of both wireless and wired communications and of techniques for managing sensor information including data fusion and sensor overloading. The challenges are to make such systems intelligent, trustworthy, self-installing, self-maintaining, self-repairing and affordable. Mobile Objects Professor DeMan presented the results from the Mobile Objects working group. The vision is of a personal, smart companion linked to the user by body-area smart sensors and connecting to services and resources. Such a device will be able to connect to a range of environments such as the home, cars, trains, aeroplanes and wider area networks. The system will be able to reconfigure itself autonomously depending on patterns of use and the surrounding environment and be able to download new services as they become available. It should be applicable to virtually any form of human activity ranging from games to health and well-being. This device will require significant developments in technology for communications, low-power computing, programming environments, energy management, Grid-computing and security for billing and authorisation. It will be capable of functioning in isolation and connecting to a wide range of services when available. To develop such a system, the gap between large systems and embedded components will need to be bridged. Key enabling technologies for such a device will address software techniques for low-power systems, software adaptable to different architectures, system specification and mapping, and smart interfaces including head-up displays and body-area sensors. Big Systems Mr Leffler presented the results from the Big Systems working group. The vision is of systems with a zero failure rate and reduced life-cycle cost. First meeting ESTP WG on Application Drivers 6/9
Such systems need to be safe, reliable and available. This will be achieved through redundancy, intelligent alarming and self-diagnosis and repair. For complex systems there needs to be a compromise between openness and security. Robustness and testing, the integration of old and new hardware and of heterogeneous systems and the management of obsolescence are important issues. Complex systems will be flexible, adaptive and self-organising. The re-engineering of existing subsystems, the management of diverse communication media and formats and ease of use are central. The design of such systems will need to accommodate smart actuators and sensors, the control of communications and large amounts of electrical power. This sector will need the development of skills to encompass a system-wide perspective, a methodology for design, testing and verification against critical requirements, cooperation between industry and academe and cross-pollination of ideas with other sectors. Major Infrastructure Professor Wolisz presented the results from the Major Infrastructure working group. The vision is of an intelligent infrastructure comprising a set of communicating, intelligent objects which can be used collectively. Examples of such infrastructures include vending machines, buildings, bridges, vehicles, undergrounds, railways and communication networks. Embedded systems can support all aspects of the lifecycle of such infrastructures including ownership, storage of long-term requirements, logging of system data, maintenance, alarms, actions by the emergency services, authorisation of access and usage, and charging and billing under a range of different conditions of usage. Usage of infrastructure is critical to everyday life. Systems must be robust to usage and resistant to malicious attack and fraud. They must also conform to legal frameworks with regard to security, trust, contracts and liability. They must be easily adaptable to new services, new tariff regimes, usage patterns and rules, and billing procedures. Privacy and health aspects are also central considerations. Furthermore, for such systems, it is the overall performance which is important, not the behaviour of individual components. For intelligent infrastructures the following hardware technologies need to be developed: Nodes containing sensors and having a radio and computing capability. These will include very small devices capable of operating at very low power, for example Smart Dust. Hardware platforms which can dynamically adapt their processing to actual requirements and operate over years without service or change of energy source; Scavenging-based energy sources using vibration, solar power, heat and environmental fluctuations; Flexible packaging, effective placement and easy deployment, for example as a component in paint. For intelligent infrastructures the following software technologies need to be developed: Component-based operating systems easily ported to different hardware platforms; Interfaces supporting a sensor/actuator network independent of the underlying hardware platform; Efficient languages supporting the development of distributed monitoring and command applications; Methods for in-network data fusion turning a sensor network into a data bank. First meeting ESTP WG on Application Drivers 7/9
For intelligent infrastructures, system organisation and management tools will need to support: Redundant structures of non-reliable components; Load-sharing across devices to ensure balanced use of resources; Situation-dependent adaptive operation; Effective, precise positioning of devices; Complex models for the interaction and monitoring of complex structures, sensors and actuators; Semantic references to devices rather than network addresses. System engineers and designers will need to develop the following skills for: Cross-layer, hardware-software design; A system as opposed to a component view; Supporting a dialogue between application and system designers. Actions arising and timetable The following actions arose from the afternoon presentations: All working group chairmen should revise their presentations to include: a one-line visionary scenario, the more down-to-earth and realisable the better; a statement of the methodologies, technologies, toolsets and skills which are needed to realise that scenario; a statement of what is important to achieve that vision and of what is less important The objective is to produce a 10 page synthesis of the vision of this working group to support the development of the high level document. Contributions for this 10 page synthesis should be sent to the Commission, Working Group Chairman and Rapporteur by May 25. Mr Glinos informed participants of the event on June 28 and 29 to present the Embedded Systems Technology Platform to high-level representatives of industry and government. Conclusions The meeting of the ESTP Application Working Group held on 18 May 2004 developed four visionary application contexts presented above. The following conclusions can be drawn from the discussions and from the subsequent elaboration of the visionary application contexts: o Visionary applications stimulate the development of technologies which may be useful in unpredictable ways. It is not necessary to realise the visionary applications per se. o Application drivers tell the semiconductor industry what needs to be developed. The semiconductor industry can then give feedback on what are reasonable goals. o Application contexts need to be ambitious, challenging scenaria. They should set targets which direct developments. o The timescales for FP7 mean that research will start in 2007 on areas which would be exploited between 2010 and 2014. Industry will need to make decisions about how this fits into its programmes for both research and implementation. The visionary application contexts chosen address four areas: Big systems from the aerospace, automobile, medical and manufacturing sectors; Mobile objects such as animals and people; Private infrastructure such as homes, houses and offices; First meeting ESTP WG on Application Drivers 8/9
Major infrastructure such as airports, cities and highways. Together these four visionary application contexts will stimulate research and development in the following areas: Low-power computing and energy management; Adaptive, long-lived, zero-maintenance platforms; Smart interfaces, actuators and sensors; Smart dust -like technologies; Scavenging energy sources using vibration and other environmental fluctuations; Situation-dependent adaptive operation of devices; Redundancy, load-sharing, intelligent alarming and self-diagnosis and repair Technologies for managing sensor information including location, tracking, sensor overload and data fusion; Communications, Grid-computing and security including billing and authorisation; Complex models for the interaction and monitoring of complex structures, sensors and actuators; Component-based and easily portable operating systems; Efficient languages supporting the development of distributed monitoring and command applications; Semantic reference to devices rather than network addresses. First meeting ESTP WG on Application Drivers 9/9