Special: Embedded Intelligence24 Ambient Hardware: Embedded Architectures on Demand by Kjetil Svarstad, NTNU, Norway 25 System-Level Design of Fault-Tolerant Embedded Systems by Alain

European Research Consortium for Informatics and Mathematicswww.ercim.org

Number 67, October 2006

Special :Embedded Intelligence

2

CONTENTS

JOINT ERCIM ACTIONS

4 CoreGRID Summer School 2006 - First Steps in NextGeneration Grid Technologiesby Stephan Springstubbe and Wolfgang Ziegler, Fraunhofer SCAI,Germany

5 Cor Baayen Award 2006 for Oliver Heckmann

6 Net-WMS - Towards a New Generation of NetworkedWarehouse Management Systemsby Bruno Le Dantec, ERCIM office

7 Joint DELOS – MUSCLE Summer School on MultimediaDigital Libraries by Remi Ronchaud, ERCIM office

8 ERCIM Working Group on Software Evolution Spawns LocalChaptersby Tom Mens; University of Mons-Hainaut, Belgium; Jean-MarieFavre and Salah Sadou, IMAG, France

8 ERCIM Environmental Modelling Workshopby Thomas Lux, Fraunhofer FIRST, Germany

9 FMICS 2006 - 11th International Workshop on Formal Methodsfor Industrial Critical Systemsby Luboš Brim, Masaryk University / CRCIM, Czech Republic, andMartin Leucker, Technical University Munich, Germany

EUROPEAN SCENE

10 A Security and Dependability Strategic Research Agenda for Europeby Jim Clarke, Waterford Institute of Technology, Ireland

11 European Research Advances in Embedded Intelligence

NEWS FROM W3C

12 W3C Workshop on Mobile Web in Developing Countries

12 Latest W3C Recommendations

12 W3C Launched Secure Browsing Initiative

12 W3C Workshop on XSL-FO 2.0 Requirements

13 COPRAS 'ICT Research and Standardization' Conference

13 W3C Mobile Web Seminar in Paris

13 First W3C Webinar on Mobile Web Design

ERCIM News No. 66, July 2007

Cover illustration by Pertti Jarla

SPECIAL THEME: EMBEDDED INTELLIGENCEIntroduction to the Special Theme

14 Embedded Intelligenceby Erwin Schoitsch, ARC Seibersdorf research / AARIT, Austriaand Amund Skavhaug, NTNU, Norway

Embedded Systems Applications

16 Autonomous Systems - Safety Critical Embedded Systemsand Intelligenceby Erwin Schoitsch and Wilfried Kubinger, ARC Seibersdorfresearch / AARIT, Austria

18 Automotive Visions beyond In-Car Driver Assistance:Integrated Traffic Management with Coopersby Thomas Gruber and Erwin Schoitsch, ARC Seibersdorfresearch / AARIT, Austria

19 Embedded Intelligence for Ambient Assisted Livingby Frank Bomarius, Martin Becker and Thomas Kleinberger,Fraunhofer IESE, Germany

21 iCat - A Friendly Robot that Helps Children and Grown-Ups by Bernt Meerbeek, Jettie Hoonhout, Peter Bingley, and Albertvan Breemen, Philips Research Europe, The Netherlands

Design, Development, Programming

22 The DECOS Tool-Chain: Model-Based Development ofDistributed Embedded Safety-Critical Real-Time Systemsby Wolfgang Herzner, ARC Seibersdorf research/AARIT, Austria;Bernhard Huber, Vienna University of Technology, Austria; andGyörgy Csertan, Andras Balogh, Budapest University ofTechnology and Economics

24 Ambient Hardware: Embedded Architectures on Demandby Kjetil Svarstad, NTNU, Norway

25 System-Level Design of Fault-Tolerant Embedded Systemsby Alain Girault, INRIA, France

27 New Tool to Design the Behaviour of Embedded Systemsby Jan Friso Groote, CWI, The Netherlands

28 Reactive Processing for Reactive Systemsby Reinhard von Hanxleden, Christian-Albrechts-Universität zuKiel, Germany; Xin Li, Partha Roop, University of Auckland, NewZealand, Zoran Salcic and Li Hsien Yoong,

30 TPC: Tagged Procedure Callsby Konstantinos Kapelonis, Sven Karsson, and Angelos Bilas,ICS-FORTH, Greece

Ambient Intelligence Systems

31 A European Ambient Intelligence Research Facility at ICS-FORTHby Constantine Stephanidis, ICS-FORTH, Greece

32 Ecological Approach to Smart Environmentsby Veikko Ikonen, VTT Technical Research Center of Finland

33 Ambiance: A Platform for Macroprogramming PersonalisedAmbient Servicesby Reza Razavi, University of Luxembourg / FNR, Luxembourg

35 Towards a Platform for Widespread Embedded Intelligenceby Simon Dobson, Steve Neely, Graeme Stevenson, LorcanCoyle and Paddy Nixon, University College Dublin / IUA, Ireland

36 Networking Semantic Services for Pervasive Computingby Valerie Issarny, Nikolaos Georgantas and Sonia Ben Mokhtar

ERCIM News No. 65, April 2006 3

Next issue: January 2007Special theme: Traffic Planning and Logistics

R&D AND TECHNOLOGY TRANSFER

50 Nanotechnology

Single Molecule Switch and Memory Elementby Emanuel Lörtscher and Heike Riel, IBM Zurich Research Lab,Switzerland

52 Mobile Hybrid Wireless Networks

HyMN: A Self-Organizing System for Interest-based DataDistribution in Mobile Hybrid Wireless Networksby Steffen Rothkugel, Matthias R. Brust and Adrian Andronache,University of Luxembourg

53 Visualization

Visualization of Colour Information on Highly Detailed 3DModelsby Matteo Dellepiane and Marco Callieri, ISTI-CNR, Italy

55 Databases

Modular Access to Large Socio-Economic Databases - theSedo Project in Luxembourgby Uwe Warner, Centre d'Etude de Population, de Pauvreté et dePolitiques Socio-Economiques, and Thomas Tamisier andFernand Feltz Centre de Recherche Public - Gabriel Lippmann,Luxembourg

37 Materials with Intelligenceby Simon Dobson and Kieran Delaney, University College Dublin/ IUA, Ireland

Wireless sensor networks

39 The CRUISE Project - Network Initiative for CreatingUbiquitos Intelligent Sensing Environmentsby Gabriele Kotsis, Johannes Kepler Universität Linz, Austria;Anelia Mitseva and Neeli R. Prasad, Aalborg University, Denmark

40 IntelliSense RFID - An RFID Platform for Ambient Intelligencewith Sensors Integration Capabilitiesby Ovidiu Vermesan, Nadine Pesonen, Cristina Rusu, Aarne Oja,Peter Enoksson and Helge Rustad, SINTEF ICT, Norway

42 Collaborative Capture: A New Perspective for SensorNetworksby Paul Couderc, INRIA-IRISA, France

Resources

43 Autarkic Power Generation for Networked Systems by Eberhard Mueller and Ulrich Hofman, Salzburg ResearchGmbH, Austria

45 Formally Bounded Embedded Systemsby Kevin Hammond, University of St Andrews, Scotland, UK

Massively Deployed Embedded Systems: Trust and security

46 Adaptable and Context-Aware Trustworthiness Evaluation by Gabriele Lenzini, Telematica Instituut, The Netherlands;Santtu Toivonen and Ilkka Uusitalo, VTT Technical ResearchCenter of Finland

47 Digital Identity in Ambient Environmentsby Ben Schouten, CWI, The Netherlands and Onkar Ambekar,ARC Seibersdorf research / AARIT, Austria

52 Trust4All enhances Trustworthiness of EmbeddedComponent-based Systemsby Gabriele Lenzini, Andrew Tokmakoff, Telematica Instituut; andJohan Muskens, Eindhoven University of Technology, TheNetherlands

56 Digital Libraries

BELIEF: Bringing Europe's Electronic Infrastructures toExpanding Frontiersby Federico Biagini, Pasquale Pagano and Franco Zoppi, ISTI-CNR, Italy

58 Electromagnetic Radiation

Cell Phone Dangerby Martin Röösli, Unversity of Bern, and Harry Rudin, Consultant,Switzerland

60 Satellite Communication Services

SAFE: Satellites for Epidemiology and Health Early Warningby Catherine Chronaki, ICS-FORTH and Laurent Braak,

61 Satellite Devices

Planck Mission 70 GHz Receiversby Jussi Varis, MilliLab-Millimetre Wave Laboratory of Finland

62 Detector Technology

Data Acquisition in Scientific Applicationsby Rob Halsall, CCLC, UK

63 Security

Network Anomaly Detection by Means of Machine Learningby Roland Kwitt, Salzburg Research GmbH, Austria

64 Web 2.0 and Novel Interaction

From Solitary Podcasting to Ambient Pondcasting by Emmanuel Frécon and Pär Hansson, SICS

EVENTS

Reports

66 EPoSS – A new European Technology Platform on SmartSystems Integrationby Erwin Schoitsch, ARC Seibersdorf research / AARIT, Austria

67 ECDL 2006 - 10th European Conference on Digital Libraries by M.Felisa Verdejo, Universidad Nacional de Educación aDistancia, Spain

68 Cross-Language Evaluation Forum - CLEF 2006by Carol Peters, ISTI-CNR, Italy

70 ARTEMIS Annual Conference 2006by Erwin Schoitsch, ARC Seibersdorf research / AARIT, Austria

71 XP2006 - Extreme Programming and Agile Processes inSoftware Engineeringby Outi Salo andPekka Abrahamsson, VTT Technical ResearchCenter of Finland

72 Announcements

73 Euro Legal

75 In Brief

ERCIM News No. 67, October 2006

CSS'06 was organised by theDepartment of Bioinformatics of theFraunhofer Institute SCAI and hosted bythe Bonn-Aachen International Centrefor Information Technology (B-IT). TheB-IT generously supported the event,providing their whole infrastructure.

The 80 participants included 55 studentsand researchers from 12 countries and 25

international experts from eight coun-tries, with approximately two out ofthree of the participants coming frominstitutions involved in CoreGRID. Theyworked on 35 course units focusing onGrid Resource Management, Schedulingand Monitoring, Grid Application in LifeSciences and Industry, as well as FutureGeneration Grids. These course unitsalso included the results of the mostimportant European Grid projects andtwo half-day hands-on exercises. CSS'06was financially supported by IBM andSUN, allowing it to maintain a registra-tion fee appropriate for students, who

indicated on their evaluation forms thatthey were very satisfied with the event.

CSS'06 will have a bearing on severallevels, the most important being the cur-riculum of the B-IT, the further integra-tion of the European Grid research anddevelopment, the transfer of knowledgeto their home institutions via the studentsand researchers and, most importantly,

the growing support for a broader use ofGrid technologies.

For the B-IT the CSS'06 marks thelaunch of a Grid curriculum. For the firsttime, starting with the winter semester2006/07, a course will be offered whichis designed as an introduction to Gridtechnologies and applications. For thefollowing semesters more in-depthcourses are planned, in particularaddressing the application of Grid tech-nology in the Life Sciences. Researchersof the B-IT expect that Grid technologieswill help satisfy the enormous demand

for computing power, data and specificservices in the Life Sciences and, forexample, the pharmaceutical industry,which until now may only be coveredusing supercomputers. Hence, thisaspect of student training has a strategicimpact and helps in assuring competitiveadvantages.

Experience gained during this summerschool and from the previous CSS pointto the creation of initial networksbetween participants from differentEuropean institutions and the lecturers.Participants and lecturers normally con-tinue maintaining these networks afterthe event, thus advancing the integrationof the European Grid research and devel-opment which is - together with excel-lence - the main objective of CoreGRID.

In the home institutions of the partici-pants, the awareness of research anddevelopment around Grid technologiesis increasing through the imminenttransfer of knowledge by the partici-pants.

Finally, the result of the training of sci-entists and students across the threelevels mentioned helps to support thebroader use of Grid technologies in dif-ferent research disciplines and industryin the future.

The CoreGRID Network of Excellenceis administrated by ERCIM.

LLiinnkkss::

CoreGRID: http://www.coregrid.net

2006 Summer School:

http://www.coregrid.net/mambo/content/

view/231/221/

PPlleeaassee ccoonnttaacctt::

Stephan Springstubbe

Fraunhofer-Institute for Algorithms

and Scientific Computing (SCAI)

Tel: +49 2241 14 2337

E-mail:

[email protected]

CoreGRID Summer School 2006 - First Steps in Next Generation Grid Technologiesby Stephan Springstubbe and Wolfgang Ziegler

The CoreGRID Summer School 2006 (CSS'06) took place in Bonn, Germany from24 to 28 July. Organising the yearly summer school, CoreGRID - The EuropeanResearch Network on Foundations, Software Infrastructures and Applicationsfor large-scale distributed GRID and Peer-to-Peer Technologies - aims at providingthe possibility of advanced training to students and researchers in a trend-settingdomain of computer science, which is currently under-represented in the curriculaof universities. CoreGRID summer schools are organised by a different CoreGRIDpartner each year and take place in distinct European cities.

4

JOINT ERCIM ACTIONS

CoreGRIDsummerschoolparticipants.

5 55

JJOOIINNTT EERRCCIIMM AACCTTIIOONNSS

In a particularly tight competition,ERCIM has awarded Oliver Heckmannfor the outstanding originality, highimpact and quality of his work in thefield of network science. To put it in very

simple words, Oliver Heckmann tries togive Internet users better quality of ser-vice for equal or lower costs. He strivesto remove inefficiencies, optimise net-works and achieve technological break-throughs that improve quality of serviceand/or costs for users. He also sheds lighton the trade-off between these two goals.The practical applicability of his resultshas put him in good contact with manyreal providers, helped him in acquiringfunding, and also led to the recent publi-cation of his book "The CompetitiveISP" which also contains many resultsfrom his excellent PhD thesis "ASystem-oriented Approach to Efficiency

and Quality of Service for InternetService Providers".

Oliver Heckmann studied applied eco-nomics and electrical engineering at TUDarmstadt with award for excellence. Hedid his PhD in computer science at TUDarmstadt with Prof. Ralf Steinmetz assupervisor and Prof. Jon Crowcroft fromUniversity Cambridge, UK as secondsupervisor. His PhD thesis was awardedthe prestigious best dissertation award ofthe German Computer ScienceAssociation for 2004. After his PhD, hewas working as an assistant professor /research group head at the MultimediaCommunications Lab (KOM), leading ateam of 7 PhD student. Recently, hejoined Google Labs in Zürich,Switzerland, to continue working onimproving and innovating the Internetand Internet services.

Finalists 2006According to the award rules, each insti-tute was allowed to select up to twofinalists from its country. 19 finalists forthe 2006 Cor Baayen Award have beennominated by the ERCIM institutes:• Alexandre Bergel, Ireland• Rickard Cöster, Sweden• Ton Dieker, The Netherlands• Guido Dornhege, Germany• Olivier Dousse, Switzerland• Sameh El-Ansary, Sweden• Serge Fehr, The Netherlands• Alain Frisch,France• Tudor Gîrba, Switzerland• Oliver Heckmann,Germany• András Kovács, Hungary• Ivan Laptev,France• Taneli Mielikainen,Finland• Harald Øverby, Norway• Tom Schrijvers, Belgium• Giuseppe Scarpa, Italy• Alkis Simitsis, Greece• Vassilios Solachidis, Greece• Gem Stapleton, United Kingdom.

Cor Baayen Award 2006 for Oliver HeckmannOliver HeckmannOliver Heckmann from Technical University Darmstadt, Germany,and Google Labs Zürich, Switzerland, has been awarded the 2006 Cor BaayenAward for a most promising young researcher in computer science and appliedmathematics by ERCIM.

Oliver Heckmann.

The Cor Baayen AwardThe Cor Baayen Award is awardedeach year to a most promising youngresearcher in computer science andapplied mathematics. The award wascreated in 1995 to honour CorBaayen, the first president of ERCIM.

The award consists of a cheque for5000 euro together with an award cer-tificate. The award winner will beinvited to the fall ERCIM meetings.

Cor Baayen Award 2007Conditions:• nominees must have carried out

their work in one of the 'ERCIMcountries': Austria, Belgium, CzechRepublic, Finland, France,Germany, Greece, Hungary,Ireland, Italy, Luxembourg,Norway, Spain, Sweden,Switzerland, The Netherlands andthe United Kingdom

• nominees must have been awardedtheir PhD (or equivalent) after 15April 2004

• a person can only be nominatedonce for the Cor Baayen Award.

Submitting a nomination:• nominations should be made by a

staff member of the university orresearch institute where the nom-inee is undertaking research. Selfnominations are not accepted

• to submit a nomination fill out theCor Baayen Award nominationform at the ERCIM web site (seeURL below).

Selection :• the selection of the Cor Baayen

award winner is the responsibilityof the ERCIM ExecutiveCommittee, who will consult expertopinion in reaching their decision.

Deadline: • nominations must be submitted by

15 April 2007.

LLiinnkk::

http://www.ercim.org/activity/cor-baayen.html

ERCIM News No. 66, Otober 2006

66 6

JOINT ERCIM ACTIONS

Warehouse Management Systems(WMS) are evolving from controllinggoods movement and storage to light

manufacturing, transportation and ordermanagement. Demands from end usersare asking for WMS solutions with deci-sion-making applications and networkedenvironment.

The Net-WMS project proposes a soft-ware solution enabling the expected newgeneration of networked businessesWMS. Net-WMS will handle networkedcommunication and co-operation pro-cesses through the integration of deci-sion-making technologies, generic 3Dplacement primitives, virtual reality for3D visualisation, interactivity to designpacking models, knowledge modelling.

The added value and competitivenessincrease of the Net-WMS solution willbe characterised by shared expertise,

easy deployment and maintenance, flexi-bility, interoperability, better resourcehandling and access to remote services.

The Net-WMS achieved prototype appli-cations for processes in a networkedWMS will include:• a packing modeller of items based on

virtual reality and optimisation tech-niques

• a palletiser tool using optimisationtechniques

• a dispatcher with the virtualisation of atruckload

a• set of interfaces will enable Net-WMScommunication between several plan-ning components across a network.

Net-WMS-generated knowledge will beshared in networked warehouses, thanksto its interoperability and user friendlyinterface designed for plant level techni-cian users. Net-WMS optimisationdevelopments will also be made avail-able for the whole community throughthe enhancement of the Choco open-source system. The achieved optimisa-tion technologies will enable 20% savingfor packaging costs. The project gathersnine European organisations having arecognised expertise in logistical andpacking problems as well as softwaredevelopment in supply chain optimisa-tion. Net-WMS complementary andmultidisciplinary consortium includes 3SMEs, 2 large manufacturers and 4 aca-demics and research centres.

Net-WMS began in September and has aduration of three years. The project ispartly funded by the Sixth FrameworkProgramme of the EuropeanCommission. The project consortium iscomposed of nine partners from researchand industry, including the ERCIMmembers SICS and INRIA.

LLiinnkk::

http://net-wms.ercim.org


Bruno Le Dantec, ERCIM office

E-mail: [email protected]

Net-WMS - Towards a New Generation of Networked Warehouse Management Systemsby Bruno Le Dantec

Net-WMS is a new project managed by ERCIM with the objective to integratevirtual reality and optimisation techniques in a new generation of networkedbusinesses in Warehouse Management Systems under constraints. The projecthas been set up in the frame of the ERCIM Working Group 'Constraints'.


Net-WMS project participants.

77


About one hundred European universi-ties and research centres have beeninvolved over the last few years in initia-tives addressing multimedia data anal-ysis, computer vision and object recogni-tion, multimedia management in largeaudio and video collections.

Currently, two Networks of Excellence,MUSCLE and DELOS, are addressingthese subjects. The networks are par-tially funded by the EuropeanCommission within the sixth EUFramework Programme, and coordi-nated by ERCIM.

In the scope of their respective dissemi-nation and training activities, MUSCLEand DELOS organized a joint summerschool in June in San Vincenzo, Italy,with the support of CNR Pisa, INRIA,CWI and the University of Modena.

The summer school presented softwaretools and research strategies to enableusers to move away from labor-intensivecase-by-case modelling of individualapplications, and allow them to take fulladvantage of generic adaptive and self-

learning solutions that need minimalsupervision in the field of new genera-tions of digital libraries.

Technologies for multimedia digitallibraries span over many disciplines:information systems, knowledge repre-sentation, computer vision, audio andimage processing, compression andstorage, machine learning and informa-tion retrieval, multimedia data mining,cross-media analysis, user interfaces andinteroperability.

Highly qualified lecturers and interna-tional researchers have been invited toaddress theoretical approaches and prac-tical solutions. Among the invitedresearchers were:• James Wang, Pennsylvania State

University• Fernando Pereira, Instituto Superior

Tècnico Lisboa• Nicu Sebe, University of Amsterdam• Milind Naphade, IBM T J Watson

Research Center• Alex Hauptmann, Carnegie Mellon

University

• Padraig Cunningham, Trinity CollegeDublin

• Jia Li, Pennsylvania State University• Alan Smeaton – Dublin City

University.

This summer school highlighted leadingresearch achievements and improvedgeneral knowledge in these fields. It alsodefined a common vision of futureresearch orientations in multimediadatamining, and provided a significantopportunity to exchange experiences,share views and address common prob-lems. Hopefully, this initiative will stim-ulate the emergence of new research col-laborations.

LLiinnkkss::

Presentations:

http://www-rocq.inria.fr/imedia/

DelosMuscleSummerSchool2006/

MUSCLE: http://www.muscle-noe.org/

DELOS: http://www.delos.info/


Remi Ronchaud, ERCIM office


Joint DELOS – MUSCLE Summer School on Multimedia Digital Libraries by Remi Ronchaud

For the first time, the DELOS and MUSCLE Networks of Excellence organised ajoint summer school. The event was held in San Vincenzo, Italy from 12 to 17June 2006 and the topic covered was Multimedia Digital Libraries.

7

Invited lecturer James Wang fromPennsylvania State University.

Summer schoolparticipants.

ERCIM News No. 66, Otober 2006

88 ERCIM News No. 67, October 2006


The first national chapter of the WorkingGroup, RIMEL (an acronym for Rétro-Ingénierie, Maintenance et Evolutiondes Logiciels), was set up in France inspring 2006. Its main objective is tobring together French teams working onreverse engineering, maintenance andsoftware evolution. In France, RIMELwill become part of a nationalGroupement De Recherche (GDR) onProgramming and SoftwareEngineering, an ongoing initiative.RIMEL brings together more than 20groups from academia and five groupsfrom industry. It promotes software evo-lution as a major topic in the Frenchresearch and education community, aswell as in industry. In order to tackle themany scientific challenges that arise inthis vast and important research domain,RIMEL includes teams from various dis-ciplines, for example database systems,model-driven engineering, software

visualisation, and service-oriented archi-tecture to name but a few.

In 2006, RIMEL organised four impor-tant events in France in collaborationwith the WG on Software Evolution:• In March 2006, a one-day workshop

on software evolution took place,attached to the French-speaking con-ference Langages et Modèles à Objets(LMO) in Nîmes. The workshop'smain goal was to provide PhD studentswith the opportunity to present theirongoing research in their mothertongue. The workshop style gaveplenty of room for lively discussions.

• In April 2006, the main annual work-shop of the WG on Software Evolutionin Lille was organised in close collab-oration with some of the RIMELmembers. As such, it attracted 14French participants, making up morethan one third of the total number ofparticipants.

In June 2006, the official kick-offmeeting took place, as part of theFrench days on Model-DrivenEngineering (Ingénierie Dirigée parles Modèles – IDM).

• In September 2006, during the firstFrench-speaking Conference onSoftware Architectures in Nantes(Conférence sur les ArchitecturesLogicielles - CAL), RIMEL organisedtwo working sessions devoted to archi-tectural evolution and software mainte-nance and re-engineering, respectively.

RIMEL will continue to be very active inthe future, and we hope to exploit thesuccess of this initiative to start similarinitiatives in other European countries.

LLiinnkkss::

Working Group web site:

http://w3.umh.ac.be/evol/

RIMEL website: http://planetmde.org/rimel


Jean-Marie Favre, Laboratoire LSR, Institut

d'Informatique et Mathématiques

Appliquées de Grenoble

Tel: +33 4 76 63 55 71


ERCIM Working Group on Software EvolutionSpawns Local Chaptersby Tom Mens, Jean-Marie Favre and Salah Sadou

The ERCIM Working Group on Software Evolution has begun to establish somelocal chapters in order to achieve a better dissemination of results at a nationallevel and to achieve a better interaction with local companies. This is importantsince the language barrier is still an issue in many countries and the introductionof national activities that are carried out in the native tongue address this problem.

ERCIM Environmental Modelling Workshopby Thomas Lux

The 13th workshop of the ERCIM Working Group Environmental Modelling washeld on 31 August - 1 September 2006 at Fraunhofer FIRST, Berlin, Germany.

Members of the ERCIM Working GroupEnvironmental Modelling participatedalong with researchers from FraunhoferFIRST and their co-operation partners inthe 13th workshop in order to presenttheir recent research work and to discussthe current trends and developments inthe field of environmental modelling.

Research issues that have beenaddressed include: European operational

air quality forecast, satellite data assimi-lation for air quality forecast, modellingand simulation of processes in the soiland groundwater zone, and especiallycurrent aspects of using Grids and Gridworkflows for different environmentalapplications. It was an enjoyablemeeting with interesing scientific sub-jects and fruitful discussions as well as apleasant social event which gave reasonto remember the founding of the working

group which was quite exactly ten yearsago.

LLiinnkk::

Working Group web site:

http://wwwold.first.fhg.de/ercim/


Achim Sydow, ERCIM Environmental

Modelling Working Group coordinator


ERCIM News No. 66, Otober 2006 99


Previous workshops were held in Oxford(March 1996), Cesena (July 1997),Amsterdam (May 1998), Trento (July1999), Berlin (April 2000), Paris (July2000), Malaga (July 2002), Trondheim(July 2003), Linz (September 2004), andLisbon (July 2005). The 2006 workshopwas organized by the MasarykUniversity Brno and the TechnicalUniversity Munich. Forty two registeredparticipants from academia and industryfrom about eleven countries attended theworkshop.

This year the program committeereceived a record number of submis-sions. The 16 accepted regular contribu-tions and 2 accepted tool papers, selectedout of a total of 47 submissions, coverformal methodologies for handling largestate spaces, model based testing, formaldescription and analysis techniques aswell as a range of applications and casestudies.

The workshop program included twoexcellent invited talks, respectively byAnna Slobodova from Intel on'Challenges for Formal Verification inIndustrial Setting' and by Edward A. Leefrom University of California atBerkeley on 'Making ConcurrencyMainstream'.

The award for the best paper was grantedthis year to Michael Weber and MoritzHammer for their excellent paper 'ToStore or Not To Store' Reloaded:Reclaiming Memory on Demand" on thetechniques for handling extremely largestate spaces in model checking of com-puter systems. The award was granted

with the support of the EuropeanAssociation of Software Science andTechnology (EASST).

The final proceedings of the workshopwill be published jointly with the 5thInternational workshop on Parallel andDistributed Methods in Verification(PDMC 2006) as post-proceedings in theSpringer's Lecture Notes in ComputerScience. Selected papers will be invitedfor publication in a special issue of theInternational Journal on Software Toolsfor Technology Transfer. The organizerswish to thank CONCUR for hosting theFMICS 2006 workshop and taking care

of many administrative aspects, andERCIM for its financial support of theworkshop. Additionally, the organizerswould like to thank EASST (EuropeanAssociation of Software Science andTechnology), Faculty of Informatics,Masaryk University Brno and theTechnical University Munich for sup-porting this event.

LLiinnkk::

FMICS Working Group:

http://www.inrialpes.fr/vasy/fmics/


Pedro Merino Gómez, ERCIM FMICS

Working Group coordinator, Universidad de

Málaga / SpaRCIM

Tel: +34 952 132752


http://www.lcc.uma.es/~pedro

by Luboš Brim and Martin Leucker

The 11th International Workshop on Formal Methods for Industrial Critical Systems(FMICS 06) was held in Bonn, Germany, on August 26-27, 2006 as a satellite eventto the 17th International Conference on Concurrency Theory (CONCUR 2006).The workshop, being also an annual meeting of the ERCIM Working Group,continued successfully the aim of FMICS workshop series - to promote the useof formal methods for industrial applications, by supporting research in this areaand its application in industry. The emphasis in these workshops is on theexchange of ideas between researchers and practitioners, in both industry andacademia.

FMICS 2006 - 11th International Workshop on Formal Methods for Industrial Critical Systems

Best paper award winners Michael Weber and Moritz Hammer.

1100

The Internet and other digital networkshave now become an integral part of oureconomy and society. But as we arerapidly adopting more information andcommunication technologies (ICT) inservices and commerce, private infor-mation is at increasing risk and securityand dependability problems become

prevalent. Indeed, today people arebecoming more and more concernedabout the increasing complexity ofinformation and communication sys-tems and the proliferation of privacy-invasive information gathering sourcesand techniques. In their online dailyinteractions, they often find themselvesfaced with high-profile losses of theirpersonal information and with viruses,spam, phishing and other crimes ofgrowing severity and sophistication. As

a result, they find themselves in anundesirable situation in which they mustput ever more trust into environments inwhich there is little or no way of under-standing or assessing them properly.

To build an information society that willdeliver growth and prosperity, we need

to tailor ICTs to business and socialneeds, and ensure that they becomeuseful tools for economic and socialinnovation. The starting point formaking them useful is to foster trust andsafeguard security in a networkedworld. In this respect, Europe's researchframework programmes are committedto the establishment of a solid securityand dependability infrastructure. TheIST-SecurIST project (an ISTCoordination Action) has been charged

with the preparation of a Europeanstrategic research agenda in the field ofICT for Security and Dependability, forthe upcoming 7th research frameworkprogramme (FP7, 2007–2013). In orderto achieve this objective, the SecurISTproject has established two fundamentalbodies: the European Security andDependability Task Force (STF), andthe SecurIST Advisory Board.

The STF is currently comprised of 180members spread across thirteen funda-mental thematic areas (initiatives) ofresearch. It provides a forum for consoli-dation and consensus building. The the-matic initiatives are shown in the figure,which provides a visual interpretation ofhow these initiatives are integrated andwork together.

The SecurIST Advisory Board is com-posed of European experts in informationsecurity and dependability. The charter ofthe board is to oversee, review, enhanceand promote results from the STF (seehttp://www.securitytaskforce.eu).

Recently, based on inputs from the STF,the SecurIST Advisory Board has issueda document presenting its recommenda-tions for a future security and depend-ability research framework in Europe,for the period 2007-2013. Under theheadline From 'Security andDependability by Central Command andControl' to 'Security and Dependabilityby Empowerment', the Advisory Boardis recommending the following nine keyresearch areas: 1.Empowerment of the stakeholders:

Stakeholders of the informationsociety include individual citizens,industry and academia, non-govern-mental organisations and govern-ments. Empowerment of the stake-holder is vital as there is a clear tech-nological trend towards decentraliza-tion of technology, as well as of itsmanagement and control. Respon-sibility, authority and control have tomove more towards the end user.

by Jim Clarke

Europe's research framework programmes are committed to the establishmentof a solid security and dependability infrastructure. The IST-SecurIST project hasbeen charged with the preparation of a European strategic research agenda inthe field of ICT for Security and Dependability, for the upcoming 7th researchframework programme.

A Security and Dependability Strategic Research Agenda for Europe

EEUURROOPPEEAANN SSCCEENNEE

10

Security and Dependability Task Force initiatives integration.


1111

2.Europe-specific security and depend-ability: Europe has a very specific het-erogeneous culture and history and setof attitudes to trust and society thatrequires specific research profiling.

3.Infrastructure robustness and avail-ability: Further research efforts areneeded for the assurance of ICT net-work and service infrastructures, aswell as the robustness and availabilityof critical infrastructure, such ashealth, energy, transport or finance.

4.Interoperability: Research on theinteroperability between security anddependability technologies and stan-dards.

5.Processes for developing secure anddependable systems: Research on thesystematic improvement of secure anddependable system development(including hardware and software)from their design phase.

6.Security and dependability preserva-tion: In an increasingly complex worldof evolving requirements, technolo-gies and systems, maintenance ofeffective system security and depend-ability is critical and is essential forpreserving user confidence.

7.User-centric security and depend-ability standardisation: Strengthen thestructured involvement of end usersand their respective representativesinto relevant standardization activities

involving security and dependabilitytechnologies.

8.Security and dependability of serviceoriented architectures (SOA): Theneed to establish and maintain trustand manage policy regulations andservice level agreements in an SOAcontext, together with commensurateadvances in software engineering todeliver service expectations.

9.Technologies for security: Research isneeded to provide a higher assuranceof trusted communication and han-dling of digital information.Cryptology and trusted functionalityand computing need to be considered.

In addition to these nine key researchareas, four future grand challenges cov-ering a 10-20 year vision are presentedby the Advisory Board. They illustratepossible longer-term possibilities andimplications: 1.Countering vulnerabilities and threats

within digital urbanization: This chal-lenge addresses open problems that wewill face in security and dependabilityfrom the expansion and globalizationof digital convergence by 2010-2015.

2.Duality between digital privacy andcollective security: digital dignity andsovereignty: This challenge deals withfuture privacy issues of all the stake-holders, whether citizens, groups,

enterprises or states. It addresses theproblem of how to override the 'BigBrother' syndrome and 'dark security',ie, the future assurance of digitalsovereignty and dignity for the variousstakeholders.

3.Objective and automated processes:This challenge addresses the problemof how to attain a controllable andmanageable world of complex digitalartefacts by 2015 and how to injectregular, quantitative techniques andengineering to make the field truly sci-entific.

4.Beyond the horizon: a new conver-gence: This last challenge deals withthe preparation of a new convergenceat a horizon of 2020 and beyond,which is the bio-nano-info-quantum'galaxy' and the new security anddependability challenges that willemerge.

LLiinnkkss::

The full report of the SecurIST Advisory

Board recommendations:

http://www.securitytaskforce.eu/


Jim Clarke, Waterford Institute of

Technology, Telecommunications Software

and Systems Group, SecurISTCoordinator


EEUURROOPPEEAANN SSCCEENNEE


The IST Results news service features numerous Europeanresearch organisations and industrial partners whose work onembedded intelligence has been funding under the EuropeanCommission's Information Society Technologies (IST) R&Dpriority.

These research projects focus on many different applicationsof embedded systems, for example:

• the ASAP project team has produced an open source pro-gramming, analysis and optimisation toolkit for pervasivecomputing systems using Constraint Logic Programming(CLP) languages. It has been validated in a series of casestudies and is expected to open the way to more effectiveand efficient embedded systems development in the future.

• The Embedded WiseNts project has brought togethertwelve partners from the top research institutions in wire-less communication, distributed computing and cooper-

ating objects in ten different European countries to focus onthe development of Wireless Sensor Networks (WSN) andtheir applications, especially in the form of CooperatingObjects (CO), to help develop a roadmap for innovativefuture applications.

• A new technology, dubbed Embedded and CommunicatingAgents, has allowed researchers at Sony's ComputerScience Laboratory in France to add a new level of intelli-gence to a robotic dog. Instead of teaching the dog newtricks, the algorithms, design principles and mechanismsdeveloped by the ECAgents project allow the robotic pet tolearn new tricks itself and share its knowledge with others.

For more examples of recent research results related toembedded intelligence and other current ICT themes, visitIST Results at http://istresults.cordis.europa.eu

European Research Advances in Embedded Intelligence

Latest W3C Recommendations• XHTML-Print

20 September 2006, Jim Bigelow, Melinda Grant • Extensible Markup Language (XML) 1.0 (Fourth Edition)

16 August 2006, Jean Paoli, C. M. Sperberg-McQueen, TimBray, François Yergeau, Eve Maler

• Namespaces in XML 1.0 (Second Edition)16 August 2006, Dave Hollander, Tim Bray, Richard Tobin,Andrew Layman

• Namespaces in XML 1.1 (Second Edition)16 August 2006, Dave Hollander, Richard Tobin, AndrewLayman, Tim Bray

• Extensible Markup Language (XML) 1.1 (Second Edition)16 August 2006, Eve Maler, Jean Paoli, C. M. Sperberg-McQueen, Tim Bray, François Yergeau, John Cowan

A complete list of all W3C Technical Reports: http://www.w3.org/TR/

W3C Workshop on Mobile Web in Developing CountriesW3C invites experts to participate in the 'Mobile Web inDeveloping Countries' Workshop in Bangalore, India, on 5-6December 2006.

Participants will discuss the challenges, requirements, and usecases for mobile Web access in developing countries and thushelp bridge the digital divide. The Workshop will bringtogether experts in mobile Web technologies and specialists onemerging countries and the digital divide.

This public Workshop is part of W3C's Mobile Web Initiative,which aims to identify and resolve challenges and issues ofaccessing the Web when on the move.

W3C invites support for this Workshop through a three-tiersponsorship program designed to support participation bypeople or organizations who might otherwise not have thefinancial means to attend the meeting.

Sponsors reach those who are making decisions about thefuture of the Web, as well as those in the public who have cometo rely on the Web as critical infrastructure for development.

To participate in the Workshop, please submit a position paperby email before 1 November 2006.

Links: CfP: http://www.w3.org/2006/07/MWI-EC/cfp.htmlSponsorship Program: http://www.w3.org/2006/07/MWI-EC/sponsors.html

W3C Launched Secure BrowsingInitiativeRecognizing the challenges people face when browsing theWeb, W3C launched an initiative to build a foundation for amore secure Web. The new Web Security Context WorkingGroup will propose standards that will enable browsers to do amuch better job helping people make proper trust decisions.

Part of W3C's Security Activity, the group's mission is three-fold: to build consensus around what information people needfrom browsers in order to understand their "security context,"to find innovative ways to present this information and raiseawareness, and to suggest ways to make browsers less suscep-tible to spoofing of user interfaces that are used to convey crit-ical security information to end users.

W3C chartered this new work after a successful Workshop onUsability and Transparency of Web Authentication in March2006. Per the charter, the group will conduct its technical workin public and will operate under the W3C Royalty-Free PatentPolicy. Organizations interested in security over the Webshould join W3C to participate in this new Web SecurityContext group.

Links: Web Security Context Working Group: http://www.w3.org/2006/WSC/W3C Security Activity: http://www.w3.org/Security/Join W3C: http://www.w3.org/Consortium/membership

W3C Workshop on XSL-FO 2.0RequirementsW3C is holding a Workshop to gather inspiration, needs andtechniques for a future version of XSL-FO, the formattingspecification on 18 October 2006 in Heidelberg, Germany. TheW3C XSL Working Group expects that the enhancements forXSL-FO 2.0 will focus on layout-driven formatting, aug-menting the content-driven layout facilities already defined.The XSL Working Group is looking for experts in the field tobring their knowledge, insight and experience, as well as foruse cases and representatives of user communities.

LLiinnkk::

http://www.w3.org/Style/XSL/2006-Workshop/

NEWS FROM W3C


13ERCIM News No. 67, October 2006

COPRAS 'ICT Research and Standardization' ConferenceBrussels, 17 January 2007

The Cooperation Platform for Research and Standardization(COPRAS), together with its consortium partners CEN, CEN-ELEC, ETSI, W3C and The Open Group, will organize a majorconference on ICT research and standardization, in the per-spective of FP7 on Wednesday 17 January 2007.

The conference will bring together all stakeholders involved inthe research/standards interfacing process, and will feature wellknown speakers from both the research and the ICT standardscommunities, as well as representatives from the EuropeanCommission. Areas to be addressed include:• how passing research output through standards processes

supports innovation, and helps projects bringing their resultsto the market

• what are the major challenges IST projects face when coop-erating with standards organizations, and how these chal-lenges should be addressed

• how should interfacing between research and standardizationbe organized in FP7, and what are the role and challenges forthe European Technology Platforms in this respect

• what are the main areas in ICT standardization where futurecloser cooperation with research projects in FP7 will berequired

• how can contributing to standards-making be made moreattractive to research projects, and how this will contribute toEurope achieving itsoverall goal of furthering theInformation Society and its position in ICT development

• what are the recommendations from FP6 for theresearch/standards interfacing process in FP7, and how canCOPRAS' results support improving cooperation betweenprojects and standards organizations in the new FrameworkProgramme.

The conference will take place at the Bedford Hotel inBrussels, and participation will be free of charge on a first-come/first-served basis. Pre-registration for the conference ispossible at: http://www.w3.org/2004/copras/meetings/open07/registration.html

Links: COPRAS Project: http://www.copras.org/COPRAS contact: [email protected]

W3C Mobile Web Seminar in ParisParis, 16 November 2006

W3C announces a public seminar on the technical solutions forthe special challenges of mobile Web adoption. The event willbe held in central Paris, on Thursday 16 November 2006.

"People could access the Web at all times and in all situations,thanks to mobile devices, but Web sites aren't always ready forthe diverse range of devices," explains Philipp Hoschka,Mobile Web Initiative Leader. "W3C's Mobile Web Initiativehas already produced concrete results to help developers maketheir Web sites accessible to mobile devices. This seminargives a valuable opportunity to discuss those results and state ofthe art techniques."

W3C's Mobile Web seminar is a public event which will focuson current results produced by W3C's Mobile Web Initiative(MWI). The confirmed speakers include representatives fromMWI sponsors such as France Telecom, Jataayu Software,MobileAware, Opera Software, and Vodafone.

This seminar is funded by the Web Beyond 3G (3GWeb) pro-ject, financed by the European Commission's FP6 IST pro-gramme. Attendance to the seminar is free and and open to thepublic, but registration is required.

Links: Mobile Web Seminar: http://www.w3.org/2006/11/mwi-seminar.htmlRegistration: http://www.w3.org/2002/09/wbs/1/mwiparis/

First W3C Webinar on Mobile Web DesignW3C organized a W3C webinar on 26 July 2006, whereCameron Moll, an expert in mobile Web design, spoke aboutmobilizing Web content. This presentation took a look at thecurrent state of the art, and how to best extend an existing orplanned Web site to handheld devices with minimal repur-posing of code, content, and images. It also showed which ele-ments of forward thinking design were needed to maintain con-sistency, regardless of where and how users access your con-tent. As part of the 3GWeb project, W3C plans to organizemore webinars to be announced on the Mobile Web Initiativehome page.

LLiinnkk::

W3C Mobile Web Initiative: http://www.w3.org/Mobile/

14

ARTEMIS(Advanced

Research andTechnology for Embedded

Intelligence and Systems) is a strong,industry-driven European TechnologyPlatform (ETP) which aims to establish acoherent, integrated European researchand development strategy for embeddedsystems (http://www.artemis-office.org).

EPoSS, another ETP launched in July thisyear (see separate article by the author inthis edition) focuses on smart systemsintegration, which is considered animportant emerging area. The key aspectsare building systems from components,interdisciplinarity, a holistic approach topervasive and ubiquitous computing, fastintegration of a variety of technologies,energy autonomy and networking(http://www.smart-systems-integration.org).

Several national research programmes inEurope cover essential aspects of thistheme, for example FIT-IT in Austria(BMVIT, Federal Ministry for Transport,Innovation and Technology), with topicssuch as Embedded Systems, System onChip, Semantic Systems and Security,focusing on radical innovations in theseareas.

'Intelligence' takes account of autonomousreasoning and acting in a co-operativemanner. 'Ambient Intelligence' refers to anexciting new paradigm in informationtechnology, "in which people areempowered through a digital environmentthat is aware of their presence and context

and is sensitive, adaptive and responsiveto their needs, habits, gestures andemotions." (ISTAG scenarios, andSchoitsch, E., Bloomfield, R. et al. (2003),AMSD – Dependable Embedded SystemsRoadmap, IST project 37553) . Thisapplies not only for people-centred tasks,which, of course, seems the most exciting,science-fiction-type, aspect, but also forpurely technical solutions like smartsensors, actuators and control systems,especially in safety related applications.Heterogeneity (of environment,applications, protocols, etc.), autonomy(self-awareness, self healing, self-organizing, etc.), nomadic mobility (adhoc, unreliable, heterogeneous, etc.) andscale-less (number of users, geography,structure, etc.) are the new emergingembedded systems challenges (NeerajSuri, Keynote at DECOS Conference,ME’06 Conference, 2006).

This special theme fits in very well withthe current European framework andstrategic research discussions. The areasaddressed - with the related Europeanresearch projects are referenced inbrackets - include:

Embedded Systems ApplicationsThis part provides an overview of fourdifferent, but nevertheless typical appli-cations of embedded systems, ie, somework on autonomous systems, on intelli-gent road safety through co-operativenetworks (COOPERS); on ambientassisted living support for the ‘agingsociety’ (BelAmI) and on an experimentwith iCat, a personal robot platform withemotional feedback. This wide spectrumof projects is typical for the broad rangeof foci addressed by the engineering ofAmI applications for the future intelli-gent environment of people.

Embedded Systems Design and DevelopmentThis provides an overview of the DECOS(EU-FP6 project) tool chain for thedevelopment of dependable embeddedsystems in a cost-efficient manner; of aproject on reconfigurable hardware for'Embedded Architectures on Demand'(AHEAD); on system-level design offault-tolerant systems (SynDEx); on atool to model the behaviour of embeddedsystem designs (mCRL2); on reactiveprocessing for reactive embedded sys-

by Erwin Schoitsch and Amund Skavhaug

Embedded Systems are a key if Europeis to remain in the forefront of digitaltechnology and as such they have beenclassified as an important research areafor the European Union's SeventhFramework Programme - the mainfinancial tool through which the EUsupports research and developmentactivities. The IST/FET (Future andEmerging Technologies) programme'Beyond the Horizon', coordinated byERCIM, points out that EmbeddedSystems, in combination with pervasive- or ubiquitous - computing, (cognitive)intelligence and software-intensivesystems, which in fact means'embedded intelligence', or 'smartsystems' in the broader context of'smart environments', are the mostimportant challenge for strategic, long-term research, with a huge impact onsociety and the economy. The ITEA2Roadmap (Information Technology forEuropean Advancement) reaches thesame conclusion - that EmbeddedSystems are a crucial technology forEuropean competitiveness.

SPECIAL THEME: Embedded Intelligence

Introduction to the Special Theme

Embedded Intelligence

Illus

tratio

n by

Per

tti J

arla

.


15

tems and on a very specialised aspect ofdesigning concurrent systems (taggedprocedure calls, TPC). This covers abroad range of system architecture,design and development issues in theembedded systems area.

Ambient Intelligence (AmI)Systems, Smart Environments,Platforms and ServicesThe articles provide an overview of oneERCIM member's AmI research facility(FORTH); on a user-centred designapproach to smart environments (EASE);on platforms for ambient services - span-ning the gap between the issue of sensornetworks and adaptive applications -(Ambiance, Construct); on Service-Oriented Architectures for pervasivecomputing environments (ARLES) and,as a special case, intelligent materials forsmart applications. This part should give afeeling for the broad range of perspectivesof 'ambiance' (vs. embedded systems intechnical engineering applications).

Further on, this special theme continueswith articles covering specific keyaspects and technologies of embeddedsystems and ambiance implementations:

Wireless Sensor NetworksFirstly, there is a description of aNetwork of Excellence to overcome thefragmentation of research in this field(CRUISE). Then two key aspects ofwireless sensor networks are high-lighted: Intelligent RFID sensor integra-tion (IntelliSense RFID) and collabora-tive capture, linking information systemsto real work by sensor networks (ACES).

Embedded SystemsA key characteristic of embedded sys-tems is the use of restricted resources, aproblem to be resolved beforehand bymodelling. One article covers autarkicpower generation for networkedembedded systems (FIT-IT, Intermon),another covers guarantees on resourceusage bounds (EmBounded).

Massively Deployed EmbeddedSystems: Trust and SecurityMassively deployed and networkedembedded systems are prone to securitybreaches leading to loss of safety or reli-ability/availability of safety or loss ofbusiness critical services (and loss of

consumer confidence). The ITEA pro-ject Trust4All project has written twoarticles on different aspects of thisproblem - research on context-awaretrust in open environments as a basicresearch issue and a trust managementframework. The special problem of pro-tecting privacy by providing separatedigital identities depending on the con-text of use in ambient environments ishandled in another article (BASIS).

These areas overlap in part, so overlap-ping topics may be addressed as well.This theme should provide a vision ofthe future for anyone interested inembedded systems and address the chal-lenges and risks for research, theeconomy and society rather thanfocusing on a single, isolated subtopic.

In most cases, the articles are based onreports of European research projectsand Networks of Excellence and thusprovide a good overview of research andapplications in the areas addressed.

Several ERCIM working groups andmany teams within ERCIM memberinstitutes are involved in at least some ofthe aspects of this special theme, whichwas suggested by the ERCIM Workinggroup on Dependable EmbeddedSystems. We have tried to provide acomprehensive overview on many dif-ferent aspects of this important theme.


Erwin Schoitsch, ARC Seibersdorf

research/AARIT, Austria



Introduction to the special theme

14 Embedded Intelligenceby Erwin Schoitsch and Amund Skavhaug

Embedded Systems Applications

16 Autonomous Systems - Safety CriticalEmbedded Systems and Intelligenceby Erwin Schoitsch and Wilfried Kubinger

18 Automotive Visions beyond In-CarDriver Assistance: Integrated TrafficManagement with Coopersby Thomas Gruber and Erwin Schoitsch

19 Embedded Intelligence for AmbientAssisted Livingby Frank Bomarius, Martin Becker andThomas Kleinberger

21 iCat - A Friendly Robot that HelpsChildren and Grown-Ups by Bernt Meerbeek, Jettie Hoonhout,Peter Bingley, and Albert van BreemenDesign, Development, Programming

22 The DECOS Tool-Chain: Model-BasedDevelopment of Distributed EmbeddedSafety-Critical Real-Time Systemsby Wolfgang Herzner, Bernhard Huber,György Csertan and Andras Balogh

23 Ambient Hardware: EmbeddedArchitectures on Demandby Kjetil Svarstad

25 System-Level Design of Fault-TolerantEmbedded Systemsby Alain Girault

27 New Tool to Design the Behaviour ofEmbedded Systemsby Jan Friso Groote

28 Reactive Processing for ReactiveSystemsby Reinhard von Hanxleden, Xin Li,Partha Roop, Zoran Salcic and Li HsienYoong

30 TPC: Tagged Procedure Callsby Konstantinos Kapelonis, SvenKarsson, and Angelos Bilas

Ambient Intelligence Systems

31 A European Ambient IntelligenceResearch Facility at FORTH - ICSby Constantine Stephanidis

32 Ecological Approach to SmartEnvironmentsby Veikko Ikonen

33 Ambiance: A Platform forMacroprogramming PersonalisedAmbient Servicesby Reza Razavi

35 Towards a Platform for WidespreadEmbedded Intelligenceby Simon Dobson, Steve Neely, GraemeStevenson, Lorcan Coyle and PaddyNixon

36 Networking Semantic Services forPervasive Computingby Valerie Issarny, Nikolaos Georgantasand Sonia Ben Mokhtar

37 Materials with Intelligenceby Simon Dobson and Kieran DelaneyWireless Sensor Networks

39 The CRUISE Project - NetworkInitiative for Creating UbiquitosIntelligent Sensing Environmentsby Gabriele Kotsis, Anelia Mitseva andNeeli R. Prasad

40 IntelliSense RFID - An RFID Platformfor Ambient Intelligence with SensorsIntegration Capabilitiesby Ovidiu Vermesan, Nadine Pesonen,Cristina Rusu, Aarne Oja, PeterEnoksson and Helge Rustad

42 Collaborative Capture: A NewPerspective for Sensor Networksby Paul CoudercResources

43 Autarkic Power Generation forNetworked Systems by Eberhard Mueller and Ulrich Hofman

45 Formally Bounded Embedded Systemsby Kevin HammondMassively Deployed EmbeddedSystems: Trust and security

46 Adaptable and Context-AwareTrustworthiness Evaluation by Gabriele Lenzini, Santtu Toivonen,and Ilkka Uusitalo

47 Digital Identity in AmbientEnvironmentsby Ben Schouten and Onkar Ambekar

48 Trust4All enhances Trustworthiness ofEmbedded Component-based Systemsby Gabriele Lenzini, Johan Muskens andAndrew Tokmakoff

ARTICLES IN THIS SECTION

SSPPEECCIIAALL TTHHEEMMEE:: EEmmbbeeddddeedd IInntteelllliiggeennccee


Basic characteristics of embedded sys-tems are that distributed, networkedelectronic control units (ECUs) are inte-grated into their technical or naturalenvironment, linked to sensors and actu-ators with properties like cognition andsituation awareness and controlled byrobust algorithms which implies some'intelligence' (smart systems).Applications are to be found in manydomains: safety-critical applicationsinclude aerospace, railways, automotiveapplications, machinery and medicalsystems, as well as entertainment.Challenges relate to the dependability(safety, reliability and security) of hard-ware, software and systems design, com-munication technologies, sensors, actors,materials etc. The Integrated ProjectDECOS is carrying out research into thedesign and development of holisticdependable systems, providing the fun-damental architecture and services in a

domain and platform independentmanner.

The next step on from a collection of'intelligent embedded functions' is thedevelopment of autonomous systems.These are able to perform complex mis-sions in an autonomous manner, copingwith unexpected incidents and interac-tions from the environment. They needto meet two basic criteria in the range oftheir activity, especially in case of inter-action with people:• a dependable, robust infrastructure as

a basis of the system (as developed, forexample, by DECOS)

• robust and reliable algorithms, sensorsand actuators, situation awareness andcognition.

The next step would be co-operative sys-tems consisting of an ad hoc set ofautonomous systems (eg autonomousplatooning of cars on a highway, control-

ling the traffic-flow by car-to-car com-munication).

ARC Seibersdorf research is doingresearch in the area of autonomous sys-tems, especially in the area of automo-tive and other mobile transportation sys-tems, providing vision systems (opticaland laser light systems), advanced driverassistance systems, recognition/identifi-cation of, for example, traffic signs,demonstrations of platooning systems(model cars) and with robots (partner inthe recently accepted robots@home pro-posal, a STREP of FP6).

DARPA Grand Challenge,Automotive ApplicationsA significant milestone in the develop-ment of autonomous vehicles was theDARPA Grand Challenge (DefenseAdvanced Research Projects Agency,USA). The driverless vehicles had tocomplete a course over 132 miles(212 km), through desert land with off-road type terrain in the Mojave Desert,without any external support.

Figure 1 shows SciAutonics/AuburnEngineering's autonomous vehicle,RASCAL (Robust Autonomous SensorControlled All-Terrain Land Vehicle).They invited ARC Seibersdorfresearch's Smart-Systems team to pro-vide their embedded stereo visionmodule SMART EYE SVS (includingall the algorithms for detection and iden-tification of obstacles, lanes, tracks,etc.). The vision system performedbeyond expectations: RASCAL was oneof the fastest vehicles in the qualificationevent, where only 23 out of 40 (of morethan 160 that applied) qualified for thefinal race. RASCAL came 16th in theend, because of a failing communicationhub.

The 2007 DARPA Grand Challenge willtake place in an urban environment. Thismeans that if the mission cannot be ful-filled because of a closed lane, thevehicle has to turn and autonomously

Autonomous Systems - Safety Critical Embedded Systems and Intelligenceby Erwin Schoitsch and Wilfried Kubinger

Embedded systems are everywhere today and will be even more prevalent in thefuture. Most of them perform simple control and support tasks on command ina pre-programmed manner, but that's not the end of the line. Increasingly, theyare designed to carry out autonomous tasks including 'intelligence' (smartsystems). They are also designed to be able to decide, based on complex inputsand situation awareness, what to do under circumstances defined by an unreliableenvironment. Autonomous vehicles are the next step in this direction.

Figure 1: RASCAL on track at DARPA Grand Challenge '05 (with stereo vision sensor).


ERCIM News No. 67, October 2006 1177

decide which alternative route to take,obeying traffic rules.

The embedded stereo vision sensor willbe tested in other environments too (aconcept car for advanced driver assis-tance systems) and in a platooningexperiment with model cars.

Figure 2 shows another demonstration ofhighly-dependable system technology –the TT Vision node application (Time-Triggered Vision node, applying TTtechnology to a demonstrator, avoiding acrash by either stopping or speeding updepending on the situation).

Autonomous vehicles are of primaryimportance in dangerous environmentswhere human drivers would be at risk oftheir lives and will, in future, allow safertransport on roads.

Autonomous RobotsOther short-distance vision systems inindustrial environments for autonomousrobots have been tested with laser lightsystems. Figure 3 shows such a laserlight stereo system.

The advantage of the laser light sectionsystem, which was implemented in a fea-sibility study for a mobile robot systemin an industrial environment, is its sim-plicity and low price.

The smart systems team of ARCSeibersdorf research is a partner in theSTREP initiative 'robots@home' of thelast IST Call 'Advanced Robotics'. Thegoal is to develop an open mobile plat-form that is able to navigate in the dif-ferent home environments encounteredin European and international homes.The home environment poses another set

of challenges yet to be resolved:changing environment, tables and chairswith thin feet, glass objects, humanscrossing the way, different materials andobjects difficult to identify, unreliableenvironment. The robot needs to be ableto 'learn' a new environment in a naturalmanner by accompanying a personaround a room, it needs to be able toclassify objects and obstacles and reactto different stimuli - and all without theowners doing some type of 'program-ming' or 'configuration'. This open plat-form is intended to be a building blockfor service and personal robots for manydifferent home applications and ambientassisted living tasks.

The consortium is composed of teamsfrom TU Vienna, ARC Seibersdorfresearch (Austria), ETH Zurich,BlueBotics, Securitas SA, NestleNespresso SA (Switzerland), LegrangeSA (France), and Otto Bock HealthCareGmbH (Germany).

LLiinnkkss::

http://www.smart-systems.at

http://www.decos.at


Wilfried Kubinger, Erwin Schoitsch, ARC

Seibersdorf research/AARIT, Austria

Tel: +43 50550 4200, 4117

E-mail: {wilfried.kubinger,

Erwin.schoitsch}@arcs.ac.at

Figure 2:TTIP - Time-Triggered Image

Processing Prototype.

Figure 3: Laser Light Section System, example for mobile robot application.


1188

In the Sixth Framework Programme ofthe European Commission, one of themain themes deals with road trafficsafety. Several projects funded by theSixth IST Framework Programmeaddress this topic, but Coopers takes aspecific position with unique ways ofachieving a safety improvement throughan intelligent network which exploitsexisting technologies for co-operativeservices.

Some 38 partners from 15 Europeancountries with industry, university andresearch institution backgrounds formeda consortium under the leadership ofAustria Tech, a company founded by theAustrian Ministry for Transport,Innovation and Technology for drivingtechnological development. The kick-offmeeting took place in early spring 2006

and since then the innovative work hasbeen driven by the partners in the areasof system and service architecture, safetyanalysis and assessment and evaluationof technologies.

Today, there exists a limited set of mea-sures for improving road safety. Thoseresponsible introduce legal regulationsand mainly static roadside installationswhich warn drivers of bends, humps,narrowings and other static danger. Butwhen it comes to accidents or traffic con-gestions, at present there is little guid-ance for the driver. Road traffic authori-ties and road operators rely mainly onthe driver's attention, experience andcapabilities. Radio traffic news is oftenlate and inaccurate in respect of the loca-tion of the event in question. VMS (vari-able message signs), which could give

concrete and more precise advice in caseof situation-related dangers, are expen-sive and therefore rarely used. Currently,the high cost for VMS deployment pre-vents European road operators adoptingthis equipment for major road sections,so missing out on a huge safety gain.

Coopers is preparing the way forimproving road safety on motorways atan affordable cost. Based on existingtechnologies and infrastructure, thedriver is provided with real-time data onthe current traffic situation ahead of him.

Automotive Visions beyond In-Car DriverAssistance: Integrated Traffic Management with Coopersby Thomas Gruber and Erwin Schoitsch

Vehicles connected via continuous wireless communication with roadinfrastructure on motorways, exchanging data and information relevant for thespecific road segment to increase overall road safety and enable co-operativetraffic management – this is the vision of the Coopers project.

Intelligent Infrastructure and Smart Cars plus individual location based services – I2Vand V2I communication.

Coopers Mission:To define, develop and test new safetyrelated services, equipment andapplications using two way communicationbetween road infrastructure and vehiclesfrom a traffic management perspective.COOPERS will build upon existingequipment and infrastructure as far aspossible to incorporate bi-directionalinfrastructure-vehicle links as an openstandardised wireless communicationtechnology. The role of motorway-operators in offering and retrieving safetyrelevant and traffic managementinformation for specific road segments onEuropean motorways based oninfrastructure and in-vehicle data will beinvestigated.

Work Areas in Coopers• Safety related traffic management and

information services• Roadside and vehicle based data

acquisition• Traffic Control Centre – TCC

Applications• Bi-directional I2V / V2I link• Roadside Transmitter• On Board Unit extension and

integration• Standardisation.




Each car is equipped with a receiver forthe I2V (infrastructure to vehicle) com-munication encapsulated in an OBU (on-board unit) and a display gives informa-tion about accidents, traffic jams, roadconstruction sites and other location andtime related events. Only messages rele-vant to the driver on a particular segmentare passed on to him. There is no irrele-vant data about traffic congestions oraccidents in remote areas of the countryas happens with traffic news on the radio.The information is accurate and preciseboth in terms of location and time.

Because the driver will be depending onthe reliability and accuracy of informa-tion, I2V will strongly increase theresponsibility and liability of the infras-tructure operator. The most effective useof I2V communications is expected inareas of dense traffic where the accidentrisk and the probability of traffic jams arevery high. In addition to V2I and in par-

allel to it, but in the reverse direction, ieby a I2V (infrastructure to vehicle) link,the real time communication can also beused for verifying infrastructure sensordata using vehicles as floating sensors.

Coopers has started with an assessmentof existing technologies, of possiblesafety improvements and a selection ofthose services most appropriate fortesting the new approach. The work plancomprises three steps for implementingthe I2V communication:• improve road sensor infrastructure

and traffic control applications formore accurate traffic information witha infrastructure-to-vehicle communi-cation concept utilising a link to roadtolling systems

• develop adequate communication con-cepts and applications ensuring relia-bility, real-time capability and robust-ness considering different technolo-gies like DAB or CALM.

• demonstrate results on importantEuropean motorway sections with hightraffic density (Rotterdam-Antwerp,Berlin/Darmstadt, Nuremberg-Brenner-Verona) and increase devel-opment strategies.

Coopers is compliant with the objectiveof sustainable transport and developmentdefined in the EC White Paper onTransport policy. As a result it willdemonstrate a prototype for future intel-ligent traffic management systems.

LLiinnkkss::


http://www.coopers-ip.eu/


Thomas Gruber, Erwin Schoitsch, ARC

Seibersdorf research, Austria

E-mail: {Thomas.Gruber,

Erwin.Schoitsch}@arcs.ac.at

Demographical and social changes havean enormous effect on health care, emer-gency and welfare services. As theaverage age continues to rise, with a cor-responding increase in chronic diseases,there needs to be a dramatic growth inassistance and care, resulting in evenhigher service costs, a decrease inquality of service, or both. While today'scommercially available emergency andcare equipment and health and fitnessdevices already cover a broad range offeatures, they are mostly stand-alonesystems and are often just too difficultfor the elderly to operate without assis-tance.

Our ambient intelligent care and assis-tance (amiCA) system will increasequality of life by using small, unobtru-

sive sensors which communicate wire-lessly and are energy saving. We employsensors that can be easily installed inelderly people's homes and need not beattached to the body, where possible.amiCA will reason on raw data seriesfrom individual sensors as well as oncombined sensor data to derive moreaccurate, reliable and/or more abstractdata.

By tracking the data, the daily routine ofthe elderly can be learned and moni-tored. Significant deviation from the typ-ical routine is used as a predictor ofpotentially unhealthy behaviour, forexample forgetting to drink or eatingspoiled food due to dementia, or of dan-gerous situations such as being too weakto get out of bed and call for help.

By tracking vital parameters, such asindicators of dehydration, of animpending heart attack, of hypogly-caemia, or a dangerously low body tem-perature, potential health-induced emer-gency situations can be detected earlyon. In this way we should be able toavoid or mitigate chronic health impair-ment as a consequence of delayed treat-ment. This will effectively improvequality of life for the elderly and cut thecosts of health care by shortening oravoiding hospitalisation and subsequentrehabilitation measures.

Following a staged model, amiCA willcall for early assistance from family,friends, or care givers. It will providedoctors with tracks of vital parametersand thus improve the precision of diag-nosis and treatment. In case of an acuteemergency, such as a fall or heart attack,an automatic call to an emergency callcentre will be triggered without interven-tion from the patient, who may not evenbe able to make the call.

Embedded Intelligence for Ambient Assisted Livingby Frank Bomarius, Martin Becker and Thomas Kleinberger

Our aging societies are facing the problem that more and more elderly peoplewant to live longer at home, in their preferred environment, despite diseases andhandicaps. We have a great opportunity to tackle this major demographicalproblem by building embedded intelligence into systems for Ambient AssistedLiving (AAL).



ApproachWe have identified majorproblem areas for the elderlyby talking to care personnel,geriatricians, medical doctorsand emergency centre staff andhave then determined appro-priate indicators that can bemeasured by sensors. Based onthis, we set up our AmbientAssisted Living DemonstratorLaboratory (AAL Lab), whichaims to come up withembedded intelligence solu-tions for elderly people livingalone at home.

The AAL Lab will be continu-ously extended. In the first ver-sion, to be opened to the publicin October 2006, the followingscenarios are just a few of thosewhich will be implemented:

Monitored Drinking –Dementia is a major reason formoving the elderly into nursing homes.Dehydration exacerbates dementia andother health problems. Reminding theelderly to drink enough is an importantfunction. Our computerised cup mea-sures the amount of fluids consumedeach day by the person handling the cup.Potential dehydration can thus bedetected.

Monitoring Food Quality – Eatingspoiled food is another source of healthproblems. An RFID-based system builtinto a refrigerator checks for expiredfood and issues warnings. Food relatedincidents can be prevented or can betraced back to the root cause, allowingfor a better treatment.

Location Tracking – Knowing the exactlocation of a person enables a wholerange of location-aware services in theAAL environment, such as switching thelights on or off or activating those com-munication devices which are closest tothe person. RFID labels embedded in thecarpet of the AAL Lab are read by awalking aid, allowing the location of theperson to be determined.

Fall Detection – A sudden fall is a strongindicator of an emergency situation andis often detected too late, sometimes

hours or days later. Gyro sensors builtinto wrist badges, walking aids, or otherbody-mounted devices can detect suddenfalls by a person and trigger a stagedemergency reaction of the AAL system.First, the person is asked to reset thedevice to rule out false positives. If theperson fails to react, a telephone orvideo-telephone connection will be initi-ated. Based on location-tracking infor-mation, communication devices in theappropriate room will be activated toestablish a communication link betweenthe person in trouble and the emergencycall centre.

Our approach to reasoning on sensordata raises the challenge of coping withinherently unreliable and imprecise data.This is tackled by combining alternativeor redundant sources. For instance, loca-tion tracking can be fed from differentsensor systems (RFID, ultra-sonic, pres-sure, etc.).

Furthermore, the system must evolveover time, so as to adapt to the specificconditions and demands of the assistedperson, as they acquire new diseases orhandicaps or recover from impairment.

The central module of amiCA is MonA(Monitoring and Assistance component).

In MonA, sensor data is col-lected and model-based rea-soning is performed. Themodel base entails models ofthe assisted person, their capa-bilities and needs, and of theperson's daily routine,including recurring medicaltreatments, as well as modelsof the living environment, forexample devices and their loca-tion and capabilities.

ValidationThe impact of our solution willbe evaluated in several steps.The first step is the prototyp-ical development and integra-tion in an apartment-like AALLab (late 2006). The secondstep is integration into anursing home and with theemergency call centre inKaiserslautern (in 2007). Thethird step is validation of emer-gency monitoring within the 6

FP project EMERGE on a pan-Europeanlevel (2006 – 2008).

This work is part of the joint researchproject BelAmI of Fraunhofer IESE, theUniversity of Kaiserslautern, theUniversity of Budapest, the Universityof Szeged, and the Bay Zoltan ResearchFoundation, Budapest. Recent achieve-ments are described in detail in thepaper: Nehmer, J.; Karshmer, A.;Becker, M.; Lamm, R.: "LivingAssistance Systems – An AmbientIntelligence Approach", in Proceedingsof the 28th International Conference onSoftware Engineering (ICSE 2006),Shanghai, China, 2006.

LLiinnkkss::

BelAmi: http://www.belami-project.org

IESE: http://www.iese.fraunhofer.de

University of Kaiserslautern research center

on AmI website:

http://www.eit.uni-kl.de/AmI/frame.html?en


Frank Bomarius, Martin Becker, Thomas

Kleinberger, Fraunhofer IESE, Germany

Tel: +49 631 6800 1200

E-mail: {bomarius, mbecker,

kleinberger}@iese.fraunhofer.de

Ambient Intelligent Care and Assistance (amiCA).

The AmI ParadigmAmbient Intelligence (AmI) refers toelectronic environments that are sensi-tive, adaptive, and responsive to thepresence of people. Its focus is not onlythe physical integration of electronics,but also the creation and generation ofenhanced experiences. Intuitive interac-tion and emotional user-interface robotswill play an important role in the realisa-tion of Ambient Intelligence. PhilipsResearch developed iCat, a research pro-totype of an emotionally intelligent robotthat can provide an easy to use andenjoyable interaction style in AmI envi-ronments.

Developed in the late 1990s, the AmIparadigm presents a vision for digitalsystems from 2010 onwards. Currenttechnological developments enable theintegration of electronics into the envi-ronment, thus providing people with thepossibility to interact with their environ-ment in a seamless, trustworthy and easyto use manner. This implies that embed-ding-through-miniaturisation is the mainsystems design objective from a hard-ware point of view. In software we studycontext awareness, ubiquitous access,and natural interaction. The user benefitsare aimed at improving the quality ofpeoples' lives by creating a desired atmo-sphere and providing appropriate func-tionality by means of intelligent, person-alised and interconnected systems andservices.

The iCat ConceptEmotional robots are generally consid-ered as a new and promising develop-ment for the intuitive interactionbetween users and AmI environments.iCat was developed as an open interac-tive robot platform with emotional feed-back to investigate social interactionaspects between users and domestic

robots with facial expressions; seeFigure 1.

iCat is a 38cm tall cat-like robot char-acter. The head contains 13 servomotorsthat can move the head and differentparts of the face, including eyebrows,eyes, eyelids, and mouth. The servomo-tors can generate facial expressions,which give the robot socially intelligentfeatures; see Figure 2. Through a camerain the nose, iCat can recognise objectsand faces using computer vision tech-niques. Each foot contains a microphone

that can identify sounds, recognisespeech, and determine the direction ofthe sound source. A speaker in thebottom can play sounds (WAV andMIDI files) and connected speech. iCatcan be connected to a home network tocontrol domestic appliances, and to theInternet to obtain information. iCat cansense touch through sensors in its feet. Itcan communicate information encodedby coloured light through multi-colorLEDs in its feet and ears. For instance,the LEDs in the ears can indicate dif-ferent modes of operation such assleeping, awake, busy, and listening.

Psychology meets TechnologyOne of the key research questions foriCat is to find out whether facial expres-

iCat - A Friendly Robot that Helps Children and Grown-Ups by Bernt Meerbeek, Jettie Hoonhout, Peter Bingley and Albert van Breemen

Can a robot cat be a buddy for children and adults? Philips Research in theNetherlands developed iCat, a prototype of an emotionally intelligent user-interfacerobot. It can be used as a game buddy for children or as a TV assistant, or playmany other roles.

Figure 1:iCat, here as agame buddy forchildren.

Figure 2: iCat facial

expressions.

Figure 1: iCat, here as a game buddy for children. Picture: Philips.

Figure 2: iCat facial expressions. Picture: Philips.

Figure 1: iCat, here as a game buddy for children. Picture: Philips.

Figure 2: iCat facial expressions. Picture: Philips.



Pic

ture

s by

cou

rtesy

of P

hilip

s.



sions and a range of behaviours can givethe robot a personality and which type ofpersonality would be most appropriateduring interaction with users for certainapplications. Research with iCat hasfocused on the evaluation of applicationconcepts for iCat, for example as a gamebuddy for children or as a TV assistant.The research questions in these studieswere: “What personality do usersprefer?” and “What level of control dothey prefer?”

The results of the user study with theiCat as game buddy indicated that chil-dren preferred to play games with theiCat rather than to play these samegames with a computer. They were ableto recognise differences in personalitiesbetween differently programmed ver-sions of iCat. Overall, the more extrovertand sociable iCat was preferred to amore neutral personality.

Young and middle-aged adults were alsoable to recognise differences in person-

ality in iCat as a TV assistant. In thisstudy two personalities were combinedwith two levels of control. In the highcontrol condition, iCat used a speech-based command-and-control interactionstyle, whereas in the low control condi-tion it used a speech-based, system-initi-ated natural language dialogue style. Thepreferred combination was an extrovertand friendly personality with low user-control.

One of the most interesting results wasthat the personality of the robot influ-enced the level of control that peopleperceived. This is very relevant in thecontext of intelligent systems that workautonomously to take tedious tasks outof the hands of humans. It suggests thatthe robot's personality can be used as ameans to increase the amount of controlthat users perceive.

ConclusionThe studies with iCat have shown thatmechanically-rendered emotions and

behaviours can have a significant effecton the way users perceive – and interactwith – robots. Moreover, users prefer tointeract with a socially-intelligent robotfor a range of applications, compared tomore conventional interaction means. Arange of further studies is planned. Oneof the questions that will be addressed ishow iCat should behave to inspire trustand compliance in users, important ifone thinks of the robot cat as a personal(health) trainer, for example.

LLiinnkkss::

http://www.hitech-projects.com/icat


Bernt Meerbeek, Philips Research Europe,

The Netherlands

Tel: +31 40 27 47535


Albert van Breemen, Philips Research

Europe, The Netherlands

Tel: +31 40 27 47864


Today, the development of embedded -and in particular safety-critical - systemsin general follows a customized designapproach, resulting in rather isolatedapplications and little reuse of compo-nents and code across different applica-tion domains. For instance, in moderncars sub-systems like power-train con-trol, advanced driver assistance systemsor the body electronic co-exist, eachequipped with its own electronic hard-ware, communication cabling etc. This

approach implies at least increased hard-ware costs, weight, and power consump-tion, last not least due to severely ham-pering the sharing of resources like sen-sors among the different sub-systems.

Therefore, the European project DECOSaims at developing basic enabling tech-nology for moving from federated tointegrated distributed architectures inorder to reduce development, validationand maintenance costs, and increase the

dependability of embedded applicationsin various application domains.'Integrated' means, that several software'IP'-blocks (Intellectual Property) of dif-ferent criticality can be allocated to onenode (ECU – Electronic Control Unit)without interfering with each other, ie,guaranteed encapsulation in space(memory) and time (each job has itsreserved time slot).

DECOS presumes the existence of a corearchitecture providing the core services:• deterministic and timely message

transport• fault tolerant clock synchronisation• strong fault isolation• consistent diagnosis of failing nodes.

Any core architecture providing theseservices (eg TTP/C, FlexRay, or Time-

The DECOS Tool-Chain: Model-BasedDevelopment of Distributed Embedded Safety-Critical Real-Time Systemsby Wolfgang Herzner, Bernhard Huber, György Csertan and Andras Balogh

'Smart systems' applications like adaptive cruise control or brake-by-wire, relyon predictable and reliable embedded system platforms as infrastructure. For thedevelopment of such dependable applications, it is therefore of crucial interestto avoid faults during design and development. Besides intensive testing, animportant way of minimising the risk of faults is by controlling the design anddevelopment process, as well as maximising the coherence of the resulting systemwith the initial requirements. The model-based tool-chain developed by the DECOSproject is described.



Triggered Ethernet) can be a basis forDECOS-based systems. On top of thesecore services, DECOS provides a set ofarchitectural (or high-level) services:• virtual networks (VN) and gateways• an encapsulated execution environ-

ment (EEE)• diagnostics.

To minimize the dependency of applica-tion programming on a certain DECOSimplementation, a Platform Interfacelayer (PIL) provides a techologyinvariant interface of the high level ser-vices for application tasks.

DECOS Tool-ChainA constituent element of such anenabling technology is a tool-chain, cur-rently being developed by DECOS,which encompasses all embedded soft-ware design and development aspects,including configuration and testing. Asillustrated in the figure, the DECOS tool-chain essentially consists of two vertical'lanes': on the left side, the integratedsystem configuration is determined andmiddleware is generated, and on theright side the application functionality isdeveloped. A third lane, containing toolsfor testing and verifying the various(intermediate) results, is not shown.

The specification starts with thePlatform Independent Models (PIMs) ofthe application sub-systems, definingtheir requirements with respect to com-munication (among the applicationtasks), performance, and dependability.

PIMs serve two purposes: firstly,together with the specification of thetarget cluster hardware and resources,the Cluster Resource Description(CRD), they are used to derive thePlatform Specific Model (PSM), whichcontains allocation (of tasks to nodes)and other information relevant for thesuccessive steps. From PSM, configura-tion files and schedules for both taskexecution and message transmission aregenerated, as well as middleware like thePIL.

Secondly, PIMs are used to guide thedevelopment of jobs (ie applicationtasks), by modelling their behaviour withSCADE (a tool set of EsterelTechnologies). If feasible, predefined

Simulink models or modules written inconventional languages like C or Adacan be imported. After application codeis generated from these models, theresults of both activities are integrated toachieve the target executables, whichcan then be downloaded to the applica-tion cluster.

The purpose of the CRD (ClusterResource Description) is to capture thecharacteristics of the platform relevantfor the software-hardware integration.This includes computational resources(CPU, memory), communicationresources and dependability properties.A graphical domain-specific modelingenvironment is developed, based onGME (Generic Modeling Environment).The targeted modeling domain isdescribed formally via the HSM

(Hardware Specification Model), a metamodel which facilitates the validationand reuse of resource models.

Before generating the PSM, it is possibleto add information manually to the PIM(PIM marking), for example informationon specific middleware requirements.

Jobs are assigned to nodes taking intoaccount functional and non-functionalconstraints. In the first phase of assign-ment, constraints are dealt with one byone. Since allocation is an NP-hardproblem, in a second phase a multi-vari-able optimisation approach is proposed.

Scheduling is the next step, where a toolsuite (TTplan, TTbuild) of TTTech (aDECOS partner developing time-trig-gered systems) has been adapted to

DECOS tool-chain basic steps. Boxes denote activities, and disk symbols denote data.Grey elements address specification, yellow address design, green ones implemen-tation, and the blue elements installation.



handle resource restrictions and EEEpartitioning. Then, PIL is generated, pro-viding generic message transfer, globaltime service and membership service(necessary to distribute information onthe state of nodes).

For behaviour modelling, SCADE (byEsterel Technologies) has been chosen asa primary tool for DECOS. SCADE isbased on a formally-defined data flownotation. It offers strong typing, explicitinitialisation, explicit time-managementand simple expression of concurrency.The PIMs are imported via the SCADEUML gateway, yielding empty job skele-tons with correct interfaces. Theirbehaviour is then directly modelled withSCADE, or Simulink models are importedto SCADE via another gateway. SCADE

is used for code generation. For linkingjob code with PIL, so-called 'SCADE-wrappers' are also generated.

The DECOS tool-chain comprises awide variety of tools from model todeployment. To ease handling, a trans-formation tool VIATRA, developed atBudapest University of Technology andEconomics, is used as backbone formodel transformations (from PIM toPSM), PIL generation and domain-spe-cific editors. Four tools are used for theDECOS tool-chain: GME, VIATRA,SCADE and TTplan/build; additionally,commercial and target-platform specifictools are used for deployment (compila-tion, linking, download). This tool-chainis designed for efficient configuration,

development and validation of critical'smart' embedded applications.

LLiinnkkss::


http://www.decos.at


Wolfgang Herzner, ARC Seibersdorf

research/AARIT, Austria

Tel: +43 50550 4132


Bernhard Huber, Vienna University of

Technology, Austria,


György Csertan, Andras Balogh, Budapest

University of Technology and Economics

E-mail: {csertan, abalogh}@mit.bme.hu

Ambient Intelligence (AmI) is consid-ered the next huge leap in computer andinformation technology. One researchproject in this area, 'Ambiesense', usedsmall computational tags serving mobileclients with information and computa-tion via wireless services. Our project,AHEAD, integrates the concept ofreconfigurable electronics with the com-putational tags, enabling mobile clientsto hold actual configuration data andreconfigure the tags on demand througha wireless service.

The main idea of AHEAD - AmbientHardware: Embedded Architectures onDemand - is the combination of threetechnological evolutions:• Hardware or computational machines

becoming smaller and disappearinginto the surroundings.

• Ambient Intelligence, the disap-pearing computers being able to com-municate and in concert solve prob-lems in new ways.

• Reconfigurable hardware, FPGA(Field-Programmable Gate Array, atype of logic chip that can be pro-grammed), closing the gap betweenhardware and software, making func-tionality truly mobile betweenmachine and program.

In the AmI perspective computers willbe tightly embedded in the surroundingenvironment and the computational tasksand services adaptable to any and manydifferent situations and thus architec-tures. Our project hypothesis extendsthis view of embedding the hardware andadapting the software into adaptable andmovable hardware, what we have chosento call 'Ambient Hardware'. This is a

platform where small computers (tags)in the environment are able to recon-figure themselves in order to fulfill thecombined hardware-software require-ments of a given task. The underlyingarchitecture is tag computers equippedwith reconfigurable coprocessors interms of FPGAs. Physically the tag com-puter is also implemented on the FPGA,and it reconfigures itself as the require-ments change as a result of clients andtheir tasks entering and leaving theactivity zone of the tag.

Although research projects and resultsabound in the areas of AmbientIntelligence and ReconfigurableComputing, our objectives in AHEADwere best solved through practicalexperiments to show which of our ideaswere sound, subsequently forming theunderlying theories when we knowspecifically which architectures andsolutions are the most promising.Presently we are still in the experimentalstage of the project.

Our first objective was to choose a goodFPGA platform for our laboratory exper-

Ambient Hardware: Embedded Architectures on Demandby Kjetil Svarstad

Imagine you are at the airport, your flight is indefinitely delayed, you are bored,and your PDA-based chess program offers you no real challenge. Luckily, youhave a machine description of the most powerful chess machine architecture(like IBM's Deep Blue) available on your PDA which you download to the nearestAHEAD-tag server, and then you can play Grand Master-level chess on the tagwith your PDA as client until your flight is finally called.



iments. We tested and evaluated severaldevelopment boards for differentFPGAs. However, the flexibility ofreconfiguration was a limitation on mostof them. In the end we chose a verysimple and available FPGA board, theso-called 'Suzaku system', which wassmall and allowed us the flexibility ofreconfiguration that we required. Thissystem is really a Single BoardComputer in an FPGA with an externalflash memory for reconfiguration dataand some communication controllers forLAN and serial line I/O.

We are now at version 1.0 of our projectroadmap. This is the first major versionwhere the tags are independent of anydevelopment PCs or external computers.The tag is now completely embedded onthe Suzaku board where the XilinxSpartan FPGA is configured with aMicroblaze CPU and devices for runninga Linux operating system variant.Additionally, the tags are independent inthe sense that they can reconfigure them-selves. A simple web server is runningon the tag and, through a specific inter-face, clients may download new configu-rations for the FPGA, then the tag willreconfigure itself and come alive againwith the basic CPU + Linux configura-tion plus the newly configured hardwarearchitecture and the interconnectionsbetween these sub-systems. The webserver interface enables a client to down-

load new configurations either directlyfrom itself or from any networked serverholding properly defined system config-urations.

In order to test version 1.0 of theAHEAD tag, we are currently devel-oping a demonstrator for MPEG re-coding and re-scaling. The objective is toconfigure the tag for receiving massivelyencoded media streams such as theMPEG4 format and to decode thestream, downscale it for a small screenlow-resolution viewer such as a PDA ora Smartphone and recode the stream in amore easily decoded MPEG format (seeillustration). MPEG4-streams are ofcourse well suited to limited bandwidth

distribution over the net, but the com-plexity of decoding them is toodemanding for low performance andlimited power devices such as PDAs andSmartphones. The tag recodes the streamfor less performance and power indecoding, but the format will demand ahigher bandwidth. This demand, how-ever, is not a problem since the tag andthe client are very close and higher band-widths are possible even with lowpower. We expect to test this tag-basedrecoder version of AHEAD by the end ofthis year.


Kjetil Svarstad, NTNU, Norway


Media streaming execution.

Our contribution to research in the fault-tolerant embedded systems consists ofseveral scheduling/distribution heuris-tics. Their common feature is to take asan input two graphs: a data-flow graph

ALG describing the algorithm of theapplication and a graph ARC describingthe target distributed architecture (seefigure).

Also shown is a table giving the worst-case execution time of each operationonto each processor and the worst-casetransmission time of each data-depen-dence onto each communication link.The architecture being a priori heteroge-neous, these need not be identical. Belowis an example of such a table for theoperations of ALG. The infinity signexpresses the fact that the operation Icannot be executed by the processor P3,for instance, to account for the require-ment of certain dedicated hardware.

System-Level Design of Fault-Tolerant Embedded Systemsby Alain Girault

Fault-tolerance is the ability of a system to maintain its functionality, even in thepresence of faults. With the advent of ubiquitous computing and distributedembedded systems, it is becoming an aspect more and more crucial. We haveprovided new functionalities to the SynDEx system-level CAD software. SynDExis ideal for optimising distributed real-time embedded systems and our newfunctionalities allow us to guarantee a specified fault-tolerance level for thegenerated embeddable code.



From these three inputs, the heuristicdistributes the operations of ALG ontothe processors of ARC and schedulesthem statically, together with the com-munications induced by these schedulingdecisions. The output of the heuristic istherefore a static schedule from whichembeddable code can be generated.

Our fault hypothesis is that the hardwarecomponents are fail silent, meaning thata component is either healthy and worksfine, or is faulty and produces no outputat all. Recent studies on modern hard-ware architectures have shown that afail-silent behaviour can be achieved at areasonable cost, so our fault hypothesisis reasonable.

Our contribution consists of the defini-tion of several new scheduling/distribu-tion heuristics in order to generate staticschedules that are also tolerant of a fixednumber of hardware components (pro-cessors and/or communication links)faults. They are implemented insideSynDEx, as an alternative to its owndefault heuristics (called DSH:Distribution Scheduling Heuristic):

• FTBAR (Fault-Tolerant Based ActiveReplication) generates a staticschedule that tolerates Npf processorfaults by replicating actively all theoperations of the algorithm graphALG exactly Npf+1 times. It workswith target architectures having eitherpoint-to-point communication links orbuses, but assumes that all the commu-nication links are reliable. FTBARtries to minimise the critical path of theobtained schedule w.r.t. the knownWCETs of the operations onto the var-ious processors of the architecture.

• RBSA (Reliable Bicriteria SchedulingAlgorithm) also generates a reliableand static schedule by actively repli-cating the operations of the algorithmgraph. The difference with FTBAR isthat the number of times an operationis replicated depends on the individualreliability of the processors it is sched-uled on and on the overall reliabilitylevel required by the user. RBSA triesboth to minimise the critical path ofthe obtained schedule and to maximiseits reliability (these are the two criteriaof this heuristic).

• GRT + eDSH (Graph RedundancyTransformation + extended DistributionScheduling Heuristic) generates astatic schedule that tolerates Npf pro-cessor faults and Nlf communicationlink faults. It first transforms the algo-rithm graph ALG into another data-flow graph ALG* by adding redun-dancy into it such that the requirednumber of hardware component faultswill be tolerated. During this phase, italso generates exclusion relationsbetween subsets of operations thatmust be scheduled onto distinct proces-sors, and subsets of data dependencesthat must be routed through disjointpaths. Then it uses an extended versionof the DSH heuristics to generate astatic schedule of ALG* onto ARC,w.r.t. the exclusion relations generatedduring the first phase.

• FPMH (Fault Patterns MergingHeuristic) is an original approach togenerate a static schedule of ALG ontoARC, tolerant to a given list of faultpatterns. A fault pattern is a subset ofthe architecture's component that can

fail simultaneously. Our methodinvolves two steps. First, for each faultpattern, we generate the correspondingreduced architecture (the architecturefrom which the pattern's componenthas been removed) and we generate astatic schedule of ALG onto thisreduced architecture (we use the basicDSH heuristic of SYnDEx for this).From N fault patterns we thereforeobtain N basic schedules. The secondstep consists of the merging of these Nbasic schedules into one staticschedule that will be, by construction,tolerant to all the specified fault pat-terns.

LLiinnkkss::

Fault-tolerance:

http://pop-art.inrialpes.fr/~girault/Projets/FT/

SynDEx: http://www.syndex.org


Alain Girault, INRIA Rhône-Alpes

Tel: +33 476 61 53 51


To the left is an example of an algorithm graph: it has nine operations (represented bycircles) and 11 data-dependences (represented by green arrows). Among theoperations, one is a sensor operation (I), one is an actuator operation (O), while theseven others are computations (A to G). Below to the right is an example of anarchitecture graph: it has three processors (P1, P2, and P3) and three point-to-pointcommunication links (L1.2, L1.3, and L2.3).

An embedded system can roughly bedescribed as some piece of equipmentwith one or more computer processors init. The processor allows this equipmentto behave with an amazing level of intel-ligence. Copiers order their own paperand cars actively avoid accidents. Theend of this is not in sight. The behaviourof such systems will become far moreadvanced than it is today.

The downside of this is that thebehaviour of systems becomes so com-plex that we cannot easily understand itanymore. As a consequence, most fea-tures of contemporary equipment remainunused. But it is even worse. Because thesystem components are exhibiting moreand more complex behaviour, they donot always understand each other. Theconsequence is that intelligent systemsoften behave erraticly. For modern TVsets a 'misunderstanding' between thecomponents occurs every minute in eachtelevision. Much effort goes intobuilding internal firewalls to restrainthese flaws so they are not noticed bythose watching TV.

Now that the behaviour of systems isbecoming so complex, it has become animportant research topic. Just as in otherengineering disciplines, complexity istackled by making models. Thesebehavioural models describe the poten-tial behaviour of the system, ie whichinteractions can take place when.Typical examples are sending messages,pressing buttons, activating actuatorsand reading sensors. Questions that mustbe answered are: are all messages prop-erly processed; can a car ever beinstructed to drive faster than the speedlimit of the cruise control; does a copieralways order paper in time or is it pos-

sible that it orders twice the amountneeded due to message duplication?

The major challenge to answering suchquestions is the 'state space explosion'.Even relatively simple behavioural

models give rise to millions of states. Inorder to show that a system behaves cor-rectly, all these states must be investi-gated. It goes without saying that anal-ysis tools are essential here. It is impos-sible to carry out the analysis of anindustrial behavioural model by hand,but manual manipulation and humaninsight is also essential to reduce it to amanageable size.

At Eindhoven University of Technology,in close cooperation with CWI inAmsterdam, we are developing a mathe-matical methodology accompanied withtools to model and analyse the (discrete)behaviour of embedded systems. Themethod is called mCRL2, where CRLstands for Common RepresentationLanguage. The methodology is based onprocess algebra, extended with datatypes. The method has been applied toalmost all major embedded systemindustries within the Netherlands.

Furthermore, it is being taught at severaluniversities.

The major idea underlying the method-ology is to transform each model to anormal form, a so called linear process.Due to its simple structure, a linear pro-cess is very amenable to symbolicmanipulation and simplification. Forinstance, by detecting confluent

New Tool to Design the Behaviour of Embedded Systemsby Jan Friso Groote

Can a car drive faster than the speed limit the cruise control indicates? Arecomputerised protocols between planes and air traffic control safe?Communication between software components in cars, planes and other intelligentembedded systems is often very complex and prone to errors. The Design andAnalysis of Systems group at Eindhoven University of Technology is releasinga new toolset, mCRL2, to improve the quality of embedded system design.

State spaces, visualized by the mCRL2 toolset.



Pict

ure:

TU

/e.

Many embedded systems belong to theclass of reactive systems, which continu-ously react to inputs from the environ-ment by generating corresponding out-puts. The programming of reactive sys-tems typically requires the use of non-standard control flow constructs, such asconcurrency and exception handling.Most programming languages, includinglanguages such as C and Java that arecommonly used in embedded systems,either do not support these constructs atall, or their use induces non-determin-istic program behaviour, regarding bothfunctionality and timing. Furthermore,the concurrent behaviour of reactive pro-grams is typically very fine-grained.Measurements indicate that it is notuncommon to have a context switchafter, on average, fewer than five instruc-tions. This makes traditional contextswitch mechanisms, based on an oper-ating system that consumes thousands ofinstructions per context switch, imprac-tical.

To address these difficulties, the syn-chronous language Esterel has beendeveloped to express reactive controlflow patterns in a concise manner, with aclear semantics that imposes determin-istic program behaviour under all cir-cumstances. There are different optionsto synthesise an Esterel program into aconcrete system. In the software syn-thesis approach, the Esterel program istranslated by an Esterel compiler into a Cprogram, which in turn runs on a COTSprocessor. Esterel can also be synthe-sised directly into hardware, via somehardware description language such asVHDL. As a hybrid approach, Esterelcan also be used in HW/SW co-design.However, these classical synthesisapproaches suffer from the limitations oftraditional processors, with their instruc-tion-set architectures geared towards thesequential von-Neumann executionmodel, or they are very inflexible if HWsynthesis is involved.

Recently, another alternative for synthe-sising Esterel has emerged, the reactiveprocessing approach. Here, the Esterelprogram is running on a processor thathas been developed specifically for reac-tive systems. The instruction set of thesereactive processors closely resembles theconstructs found in Esterel, such aswaiting for the occurrence of a signal orabortion. A comparison highlighting themain philosophical differences betweenreactive processors and conventionalprocessors is outlined in Table 1.

The intention of reactive processors is tounite the pros of both the software andhardware synthesis approach whileavoiding most of the cons of both. Weget from the software synthesis approachthe easy programming process withinexpensive testing and short turnaroundtimes. From the hardware side, we gainlow energy consumption and fast pro-gram execution. The synchronous modelof computation inherent in Esterelensures deterministic behaviour.

Two strategies have been proposed todesign reactive processors, whichmainly differ in the way they handle con-currency. Since Esterel is a synchronouslanguage, the operation of all concurrent

Reactive Processing for Reactive Systemsby Reinhard von Hanxleden, Xin Li, Partha Roop, Zoran Salcic and Li Hsien Yoong

Embedded real-time systems must react continuously to stimuli from theirenvironment. Therefore, their control-flow patterns differ from those of traditionalsystems that transform a given input to an output at their own pace. Reactiveprocessors provide direct hardware support for reactive control flow, which keepsexecutables fast and compact and results in lower power consumption comparedto traditional architectures.


behaviour in a linear process, the associ-ated state space is often exponentiallysmaller than without the use of conflu-ence. Experience has shown that allmodels can straightforwardly be trans-lated to linear form. Unfortunately, notevery form of analysis can be carried outon linear processes. For this stateautomata are needed, but these can notalways be generated.

Besides the analysis tools, the toolsetcomes with several visualisation tools. Inparticular, tools that can visualise thestructure of state spaces of up to a mil-lion states are very helpful in obtaining

an insight into the behaviour ofembedded systems. These techniqueshave been used more than once to detectsymmetric or erroneous behaviour. Theyhave also been very helpful in under-standing the extent of problematicbehaviour compared with the overallbehaviour.

The tools and techniques are beingdeveloped in the scientific domain. Thisnot only means that the ideas and tech-nologies being used are published withinthe scientific community (as iscommon), but also that the tools are anopen source, free for both scientific and

commercial use. A preliminary versionof the toolset can be found atwww.mcrl2.org. A full first release isexpected in the autumn of 2006 and willbe available for Linux, Mac OS X andWindows. It will contain all availablemodelling, analysis and visualisationtools.


Jan Friso Groote, Eindhoven University of

Technology, The Netherlands

Tel: +31 40 2475003


http://www.win.tue.nl/~jfg




threads is based on the ticks of a singleglobal clock. The principal challenge isto schedule concurrent threads so thatwithin a logical tick, event producers areexecuted before event consumers. TheEMPEROR processor, developed at theUniversity of Auckland, New Zealand,employs multi-processing and usesdynamic scheduling of events. Thisallows it to handle many threads runningon a cyclic executive on a given pro-cessor or between concurrent threadsrunning on different processors. TheKiel Esterel Processor (KEP) family,

developed at Kiel University, Germany,handles concurrency via multi-threading, which minimises overallresource usage and easily scales up tovery high degrees of concurrency.Unlike the dynamic self-scheduling ofthreads in the EMPEROR approach, theKEP approach schedules all producersbefore consumer threads to uniquely

determine signal statuses. Bothapproaches have their advantages andlimitations. However, overall they per-form much better than conventional pro-cessors while performing reactive com-putations.

Apart from efficiency and determinismconcerns, another advantage of reactiveprocessors is that due to their compara-tively simple structure (no caches, nopipelining) and their direct implementa-tion of reactive control flow constructs,it becomes feasible to precisely charac-

terise their timing behaviour. In conjunc-tion with the synchronous model of com-putation, which discretises time into log-ical ticks, it is possible to derive exact,tight bounds on its Worst Case ReactionTime (WCRT), which tells how muchtime it takes the system to react to theenvironment. The KEP processor seriesis equipped with a Tick Manager that can

provide a constant logical tick rate andalso detects internal timing over-runs.This can serve to detect hardware fail-ures and provides another safeguard, inaddition to static analyses, so that real-time deadlines are met.

FPGA-based implementations of reac-tive processors have proved very com-petitive to classical processor designs.For a standard suite of Esterel bench-marks the code size is typically an orderof magnitude smaller than that of theMicroBlaze, a 32-bit COTS RISC pro-cessor core. The worst case reaction timeis typically improved four-fold andenergy consumption is also typicallyreduced to a quarter.

In the five years since its inception, thereactive processing approach has thusdemonstrated its promise and its practi-cality. However, much remains to bedone. On the theoretical side, a precisecharacterisation of the reactive executionsemantics is still missing and its relation-ship to other semantics needs to beinvestigated, in particular regardingcausality issues. The reactive instructionset architectures also pose interestingcompiler challenges, for example,regarding the efficient mapping of con-currency with minimal contextswitching. Finally, the derivation ofWCRT bounds is so far done very con-servatively - it should be feasible totighten these analyses further.

LLiinnkkss::

http://www.informatik.uni-kiel.de/rtsys/kep

http://www.ece.auckland.ac.nz/~roop/page

Gen.php?pageid=10


Reinhard von Hanxleden, Christian-

Albrechts-Universität zu Kiel, Germany

Tel: +49 431 880 7281


Partha Roop, University of Auckland, New

Zealand

Tel: +64 9 3737599


Table 1: Comparison between reactive and conventional processors.

Features Reactive Processors Conventional Processors

Execution progressionEvolves in discreteinstants separated by “tickdelimiting instructions”.

Evolves continuously.

Preemption

Accomplished throughevent reaction block withimplicit priority resolutionand context switching inhardware.

Accomplished throughinterrupt mechanismrequiring explicit priorityresolution, context savingand restoration insoftware.

View of the environment

Changes at discreteinstants. Inputs are latchedat the beginning andoutputs are sustained tillthe end of a “tick”.

Changes continuously.Inputs can be read at anytime, and outputs can besustained for any duration.

Concurrency

Synchronous parallelexecution and broadcastcommunication betweenthreads.

Asynchronous executionrequiring explicit messagepassing/rendezvous forcommunication betweenthreads.


3300

A huge opportunity and challenge weface today is the design of embeddedsystems that will support demandingapplication domains. Current technologytrends in building such embedded sys-tems advocate the use of parallel systemswith (i) multi-core processors and (ii)tightly-coupled interconnects. However,exploiting parallelism has traditionallyresulted in significant programmereffort. The programming model plays animportant role in reducing this effort.The main challenges in the programmingmodel are to expose mechanisms thatneed to be used directly by the pro-grammer and to hide mechanisms thatcan be used transparently.

It is predicted that multi-core processorswill use the increasing number of transis-tors more efficiently than traditionalsingle-core processors and offer higherperformance. However, multi-core pro-cessors require a large degree of paral-lelism that has traditionally demandedadditional programming effort. There isa need for an intuitive programmingmodel for multi-core processors.

Moreover, future interconnects will playan important role in such systems as theyinterconnect all the cores in a system. Tobe efficient, the programming modelmust facilitate efficient use of the inter-connect.

Tagged procedure calls (TPC), are a newprogramming model that re-draws thebalance between what the programmerneeds to specify and what the architec-ture should provide. Furthermore, TPCaim at unifying intra- and inter-chip pro-gramming abstractions. Informally, themain points behind TPC are:• TPC allow the programmer to aug-

ment procedure declarations and callswith tags that control the way theseprocedures will be executed.

• TPC require the programmer tospecify parallelism through special,tagged, asynchronous procedure calls.TPC also provide primitives that blockexecution until certain procedure callshave completed, see Figure 1.

• TPC discourage the programmer fromaccessing global data within taggedprocedures.

• TPC allow the programmer to usepointers to global data as arguments toprocedures.

Tags placed on procedure calls by theprogrammer define an abstract represen-tation of the required execution seman-tics, see Figure 1. The programmer (a)has to identify the available parallelism,(b) is urged to identify the data usedduring parallel computations and (c) hasto specify how parts of the code will exe-cute, eg as atomic or serializable regions.However, they need not worry abouthow this will be achieved. Finally, TPCaim at unifying intra- and inter-processorprogramming models and dealing withthe inherent heterogeneity of futureembedded systems in a simple and intu-itive manner.

TPC encourage the programmer tospecify parallelism and data used duringparallel execution, but do not expose theunderlying mechanisms for communica-tion, synchronization, etc. We believethat this balance between what the pro-grammer needs to specify and what canbe done transparently will result in bothefficient execution and reduced pro-grammer effort on future embedded sys-tems.

Our current work focuses on a prototypeimplementation of TPC on an embeddedplatform. We are using an FPGA-proto-type (see Figure 2) with multiple cores.Using this implementation platform, weare currently experimenting with thesemantics of TPC. We are porting anumber of existing parallel applications.This effort will allow us to examine thecost introduced by the runtime systemwhen transparently dealing with commu-nication and synchronization issues.

This work has been partially supportedby the European Commission in the con-text of the SARC project.


Angelos Bilas, FORTH-ICS and University

of Crete, Greece


http://www.ics.forth.gr/~bilas/

TPC: Tagged Procedure Callsby Konstantinos Kapelonis, Sven Karsson, and Angelos Bilas

Tagged Procedure Calls (TPC) is a new approach addressing the problem of thehigh programmer effort needed to achieve scalable execution. TPC is targeted atarchitectures ranging from small embedded systems to large-scale multi-coreprocessors and provides an efficient programming model easy to understandand exploit.

Figure 1: TPC example. Calling a functionwith consistency (c) and isolation (i) tags.

Figure 2. Current prototyping platform atFORTH-ICS. The depicted architecture isimplemented using a Xilinx XUP board.

/* Function with OUT argument */void procedure(...){...}void main(){/* Define a handle for function */TPC:DECL(hdl,procedure).../* call asynchronously procedure()* it changes argument s */TPC:CALL(hdl,procedure,s) c,i/* Continue execution */.../* Use handle hdl to wait for * procedure() to complete */TPC:WAIT(hdl)/* s can now be accessed */...}




Ambient Intelligence technologies inte-grate sensing capabilities, processingpower, reasoning mechanisms, net-working facilities, applications and ser-vices, digital content and actuating capa-bilities distributed in the surroundingenvironment. While a wide variety ofdifferent technologies is involved, thegoal of Ambient Intelligence is to eitherentirely hide their presence from users orto smoothly integrate them in their sur-roundings as enhanced environment arti-facts rather than technological gadgets.This way, the computing-oriented con-notation of technology essentially fadesout or even disappears in the environ-ment, providing seamless and unobtru-sive interaction paradigms. Therefore,people and their social situation, rangingfrom individuals to groups, be they workgroups, families or friends and their cor-responding environments (office build-ings, homes, public spaces, etc) are at thecentre of the design considerations.

Ambient Intelligence brings a specialperspective to the on-going researchassociated with technical fields likeubiquitous computing, pervasive andproactive computing, ambient com-puting, embedded computing, and smartobjects. Ambient Intelligence hasbecome well-focused by putting peopleand social contexts at the centre, while atthe same time aiming to distribute,embed, coordinate and interactivelydeliver computing intelligence withinthe surrounding environment. Thenotion of Ambient Intelligence, asdescribed above, is becoming a de factokey dimension of the emergingInformation Society, since many of thenew generation industrial digital prod-ucts and services are clearly shifted

towards an overall intelligent computingenvironment.

From a technological point of view,Ambient Intelligence is a particularlycomplex, multi-faceted and demandingscientific domain, requiring the presenceand seamless integration of most of thekey technologies existing today, thusposing numerous challenges in severalresearch areas and requiring large scaleexperimentation.

The Institute of Computer Science (ICS)of FORTH is in the process of creating alarge-scale, state-of-the-art AmbientIntelligence European Facility, whichwill act as a research nexus for studyingand developing, from a human-centredperspective, related technologies and forassessing their impact on the individual,as well as on society as a whole.

Starting with the creation of an intelli-gent home simulator, the AmI facilitywill be expanded to address specificindoor and outdoor environments, rele-vant application domains (eg housing,work, health, security, education, trans-portation and entertainment), as well astheir related physical, social and culturalcharacteristics.

The AmI Facility will also provide ashowcase for demonstrating the potentialadded-value and benefits of AmI tech-nologies in different aspects of everydaylife and activities. In this way, the AmIFacility will foster the vision of AmbientIntelligence, facilitate multi-disciplinaryinternational collaborations and providea focal point for technology transfer anddissemination of know-how to Europeanindustry, adding to its competitiveadvantage. It will also contribute

towards the European strategic priorityfor "an Information Society that is inclu-sive, provides high-quality public ser-vices and promotes quality of life", whilealso promoting synergies and knowledgediffusion in the context of relevantEuropean research projects and actions.

The creation of this facility builds on thescientific know-how and technicalexpertise of FORTH-ICS in a number ofcontributing critical domains such asHuman-Computer Interaction, UniversalAccess, Artificial Intelligence,Semantic-based Knowledge Systems,Robotics, Computational Vision,Networks and Telecommunications,Information Security, DistributedSystems, Computer architecture,Microelectronics, Sensors andBiomedical Informatics. AllLaboratories of FORTH-ICS are activelyengaged in this effort, and collaborationsare promoted with other FORTHInstitutes, as well as other organisationswithin and beyond ERCIM.

In the context of promoting research inthe domain of AmI, FORTH-ICS hasplayed a key role in the establishment ofthe new ERCIM working group: 'SmartEnvironments and Systems for AmbientIntelligence'. The AmI Facility ofFORTH-ICS will promote and supportactive collaboration and synergiesamong the members of this ERCIMworking group, by offering a technolog-ical platform and an experimentaltestbed for research and developmentactivities.

LLiinnkkss::

SESAMI Working Group:

http://www.ics.forth.gr/sesami/

ICS-FORTH Ambient Intelligence

Programme: http://www.ics.forth.gr/ami


Constantine Stephanidis, FORTH-ICS,

Greece

Tel: +30 2810 391741


A European Ambient Intelligence Research Facility at FORTH - ICSby Constantine Stephanidis

Ambient Intelligence (AmI) presents a vision of tomorrow where 'intelligent'environments react in an attentive, adaptive and active (sometimes proactive) wayto the presence and activities of humans and objects in order to provide appropriateservices to the inhabitants of these environments. It is an emerging field of researchand development that is rapidly gaining wide attention from an increasing numberof researchers and practitioners worldwide, particularly in Europe.


The technological shift of computing,including applications and servicesdirected to two dimensions, embedded-ness and mobility, has already changed alot about our relationship to our environ-ment (both social and technological).

Technology has always been embeddedin our living environment in some way.The technological infrastructure hasbeen fading from our sight as tech-nology has reached a more maturestatus. In the electronic and computerera the wires, base stations and serversare usually hidden and we see only ourpersonal technical appliances. However,the embeddedness of technology is nowshifting from particular, computer-situ-ated spaces, towards a computers-every-where philosophy. The idea that wehave a place where our interaction withour environment is supported by com-puters is changing. Soon we will havetechnological components (communi-cating with each other) everywhere, outof our sight, obtaining and utilisinginformation gathered from the environ-ment.

The increasing mobility of informationand communication technologies hasalso changed our relationship with ourenvironment. With mobile computers wecan carry lots of data with us (eg books,music and photos), but at the same timewe can also create new expressions ofourselves and share this informationwith others. This connection to theglobal network enables continuous infor-mation-sharing and communication invarious ways. When these two dimen-sions work together, when the personwith mobile technology interacts withthe situated smart environment, we areapproaching the area of ubiquitous (orpervasive) computing, also calledambient intelligence.

Even though research in the area of tech-nologically-embedded intelligent envi-ronments is expanding very fast, thedevelopment of design approaches, otherthan technology-driven ones, is still in itsearly stage. The methodologies com-monly used in designing and evaluatinginformation and communication tech-nologies have been quite goal-oriented,but new approaches and methods havebeen introduced and tested along theway. Design principles for intelligentenvironments have been published, butin these statements the technologicalissues are also emphasised to a greaterextent. More profound co-operation

between different designers, developersand researchers is needed in order to putforward new theories and methodologiesthat will help to develop solutions thatnaturally support people in their livingenvironment and take into account boththe complexity of the systems and tech-

nological aspects as well as social, eth-ical and cultural issues.

With EASE, we have brought togetherparticipants from different fields indesigning computerised smart environ-ments and especially those who dealwith human technology interaction inthis research area. The aim of the projectis to identify, model and evaluate theecology of an intelligent environmentfrom the perspectives of human, techno-logical, environmental and differentdesign methodologies. We are speci-fying the attributes and characteristics ofgood smart environments as well asidentifying and developing suitablemethods for the design, implementationand introduction of technologically-enhanced smart environments. The par-ticipants make up an interdisciplinaryteam, which is vital in order to attain

truly holistic knowledge of the phe-nomena. The project pools the knowl-edge that research partners have accu-mulated in other projects. Our aim is tocultivate this know-how and to developand present it in a more general mannerfor designing future smart environments.

Ecological Approach to Smart Environmentsby Veikko Ikonen

The project EASE (Ecological Approach to Smart Environments) aims to offermore profound analysis and research-based evidence to support the design ofsmart environments. Furthermore, the aim is to analyse the design issuesrelated to the smart environments and offer some general guidelines for thedesign of smart environments in different application areas.

32 ERCIM News No. 66, July 2006

Training with smart technologies.

Illus

tratio

n by

Per

tti J

arla

.


ERCIM News No. 66, July 2006 33

Ecology is defined by EASE as researchof the dynamic interaction systembetween environment (eg technology ingeneral) and people, in any relation tothis technologically augmented environ-ment. Users' actions with the smartapplications shape the environment, andthe smart environments produce effectson the users and the usage which mayeven be reflected outside the actualusage environment. The practise of thisdynamic relationship is in continuoustransition although it seeks stability.

Future computerised smart environ-ments are a challenging design target.This issue is especially tricky whendesigning public places and multi-userenvironments. Private or semi-privatespaces (ie, work, car, home) can beadjusted more easily according to indi-vidual users or a certain user group. It isalso easier to compose common rulesand regulations for workplaces than forpublic spaces, for example. One of thegreat challenges and opportunities wouldbe to integrate the designer and the user

again and give control over his comput-erised environment back to the user.

The technical development and consid-eration of technological issues of smartenvironments is going to be a huge jobfor our global society. Equally importantis to consider cultural (including socialand ethical) issues related to the techno-logical progress of future computerisedenvironments. User involvement andcontextual studies of human technology-interaction in general are going to be alargely accepted curriculum of society inthe near future. The great challenge is tomaintain a holistic or ecologicalapproach as a part of this curriculum, dueto its multi-disciplinary nature.

We want to test our approach in multi-faceted environments where contexts areoverlaid, interruptions in simultaneoustasks are more regulation than exceptionand where various user groups, with dif-ferent skills and technologies, want toaccomplish their primary and secondarygoals. Usually, the only valid way to

evaluate the concept is in the real-lifelongitude studies and follow-up studies.

The timeframe of the project is January2005 to December 2006 and it is fundedthrough the Finnish Funding Agency forTechnology and Innovation (Tekes),VTT (Technical Research Centre ofFinland) and TTL (Finnish Institute ofOccupational Health). Other projectpartners are the University of Lapland,Tampere University of Technology,University of Tampere, University ofOulu, Helsinki University ofTechnology and University of Art andDesign Helsinki. The project is coordi-nated by VTT. The steering group of theproject consists of a wide range of enter-prises and organizations.

LLiinnkk::

http://www.vtt.fi/proj/ease/


Veikko Ikonen

VTT - Technical Research Center of Finland

Tel: +358 20 722 3351


A key enabling technology for AmI isnetworks of large numbers of wirelessly-connected small, low-powered com-puters. Such a system is called aWireless Sensor Network (WSN) andeach node, a mote. WSNs can serve as aninfrastructure for the provision of per-sonalised ambient services. However,WSNs face very limited processing,memory, sensing, actuation and commu-nication ability of their motes.Programming such networks meansthose limitations need to be addressed.

Unfortunately, current methods forWSN programming have led developers

to mix serious concerns, such as qualityof service requirements, with low-levelconcerns like resource management,synchronisation and routing. This makesdeveloping software for WSNs a costlyand error-prone endeavour, even forexpert programmers.

Macroprogramming is a new techniquewhich is being developed with the aim ofallowing programmers to capture theoperation of the sensor network as awhole.

In this research, we focus on simplifyingsensor network programming by devel-

oping a platform which supports macro-programming by non-professional pro-grammers. Our aim is to minimise therequired programming knowledge,empowering ordinary users to interactwith the network so that they can intu-itively formulate the expected services.Our architecture also supports an open,concurrent system – requests may comein asynchronously from uncoordinatedend-users. They are formulated using anintuitive and omnipresent World WideWeb interface. They are served ubiqui-tously and in parallel.

Architecture of the AmbiancePlatformTo achieve the above goals, we areworking on extending the AdaptiveObject-Model (AOM) architectural stylewith Actor-based concurrent computa-

Ambiance: A Platform for MacroprogrammingPersonalised Ambient Servicesby Reza Razavi

A key enabling technology for Ambient Intelligence (AmI) is Wireless SensorNetworks (WSNs). Macroprogramming WSNs by non-programmer end-users isbeing studied as a step towards an omnipresent World Wide Web interface forthe provision of personalised ambient services.


3344

tion, communication, sensing and actu-ating. The resulting platform, calledAmbiance, provides a meta-level archi-tecture to automatically translate high-level specifications of global behaviourby end-users into both meta-objects andmeta-actors. These control and cus-tomise the runtime behaviour of passiveand active application objects.

The meta-objects are dynamic. Theyhave the capability to observe the appli-cation objects and the environment(introspection) and to customise theirown behaviour by analysing these obser-vations (intercession). Behaviour modelsand requested services are specified byend-users, in their own terms, using asophisticated interface provided byAmbiance. Our architecture also allowsmeta-objects to modify their behaviourin more fundamental ways if, for

example, the meta-objects are endowedwith learning mechanisms.

The architecture of the Ambiance plat-form comprises four subsystems (seefigure, from left to right):1.Presentation subsystem (APE): the

system's front-end. It is a web-enabledapplication which is responsible forthe presentation and interaction withend-users.

2.Representation subsystem (ARE):offers to ordinary end-users an intu-itive, spreadsheet-like task-specificlanguage for commanding the system.Web pages sent to the APE are dynam-ically generated by ARE, in HTML,thanks to the Seaside framework and

Squeak. ARE exports service modelsin XML to the Transformation sub-system (see below). It also sends exe-cution requests to the Deploymentsubsystem and receives from it theexecution result. On reception, thelatter is communicated to the APE (forformatting and presentation to the end-users).

3.Transformation subsystem (ATE):implemented in VisualWorksSmalltalk, is responsible for modeltransformation and verification. Itallows service requests to be deployedon different runtime environments.

4.Deployment subsystem (ADE): isresponsible for executing models in anadaptive, distributed, and concurrentenvironment.

For implementing this architecture wereuse the Dart meta-level object-oriented

framework for task-specific, artifact andactivity-driven behaviour modeling.This framework offers the reificationsneeded for (1) explicitly representing theambient services and (2) automating theselection and deployment of an appro-priate execution strategy, according tothe service's resource consumptions andthe actual execution environment(Context-aware Computing). Each spec-ification is translated into a group ofmeta-actors which implement protocolsto meet it.

These meta-actors control the concurrentand distributed collection of data byactors in the nodes of a sensor networkand analyse the data. Note that the anal-

ysis may trigger further data collection,or reuse already collected data. Thesystem being developed also provides atype system for a partial semantic check,to ensure that queries are well-formed,so that for example, a query does notresult in a graph of node invocationswhich is cyclic.

Ongoing WorkThis research is conducted within theframework of the Åmbiance project,funded by the University ofLuxembourg, in collaboration with theOpen Systems Laboratory (OSL) and theSoftware Architecture Group (SAG) atthe University of Illinois at UrbanaChampaign (UIUC-USA), as well as theComputer Science Laboratory (LIP6) ofthe Pierre and Marie Curie University(Paris).

It draws on our previous real-isations Dart, AmItalk, andActorNet. A first runningprototype is implementedand we are finalising the sys-tematic transformation ofend-user queries to producemeta-actors and to dynami-cally manage their life-cycle.Such a lifecycle involvesactivation and registration,request management, appli-cation logic and result dis-semination. A type systemfor Dart is being written inMaude (http://maude.cs.uiuc.edu). Once the coresystem is in place, we willwork with experts in domains

to which sensor networks are applicable,such as civil engineering, cooperativetarget identification and tracking, envi-ronment monitoring and security, toexperimentally assess and refine theAmbiance architecture.

LLiinnkkss::

http://ambiance.uni.lu/

http://osl.cs.uiuc.edu/

http://wiki.cs.uiuc.edu/SAG


Reza Razavi, University of Luxembourg /

FNR, Luxembourg


Figure 1: The Ambiance Platform supports macroprogramming WSNs by automated dynamic codegeneration and deployment.


Any large-scale pervasive deploymentwill face a common core of issues.Decentralised control can be used to pro-vide coordinated behaviour in the face oflocalised failures and a dynamic popula-tion of devices. In many cases a knowl-edge-based approach has proved to besuccessful. This is something that isespecially important for sensor-drivensystems, which are inherently noisy andincomplete and so must deal with sub-stantial amounts of uncertain reasoning.The systems must be self-configuring, -managing and -optimising (the 'self-'properties of autonomic computing andcommunications) as well as being exten-sible, standards-based, and upgradableon-the-fly. Finally, systems must scale tohundreds or thousands of devices whileoffering predictable and robust levels ofservice.

Within University College Dublin'sSystems Research Group we have beenlooking at middleware solutions to theseissues, building both on our own experi-ence with middleware and the other suc-cessful pervasive computing frame-works such as the Context Toolkit,Cooltown and Aura. The resultingsystem - Construct - provides a scalableand open platform on which to build per-vasive and adaptive applications.

Construct is unusual in a number ofways. It is entirely data-driven, in that alldata is treated as being sensed either by atraditional physical sensor, a 'virtual'sensor accessing on-line information, ora reasoning strategy, making inferencesfrom other data. All information is repre-sented internally using the ResourceDescription Framework from the

semantic web, allowing easy integrationwith other tools. Applications - bothclients' and sensors' - are written to queryand/or populate this model, which con-siderably simplifies the design anddeployment of new components.

Noise and uncertainty are ubiquitouscharacteristics of any sensor-drivensystem, and Construct applies uncertainreasoning techniques to all its informa-tion. The system can maintain consensusestimates of items of interest, modifyingits results and confidences over time.This means that the results from exces-sively noisy or inaccurate sensors, whichwould tend to disagree with the con-sensus, will be de-emphasised over time,leading to a self-stabilising infrastruc-ture. The disadvantage of this approachis that all inferences are explicitly taggedwith confidence values, which can com-plicate a programmer's task.

Pervasive systems suffer from frequentnode failures, either from genuine faultsor from devices being removed or deacti-vated. This causes problems for manyinfrastructures: where should data bestored so as to maximise availability?Construct answers this by having eachnode store a copy of the public state ofthe system, and gossiping this state ran-domly between nodes. Any informationsensed at a node, for example, may beplaced into the public store and will begossiped to another node. If the sourcethen fails, its observations will havealready been transferred to other nodesand so need not be lost. Careful handlingis required to ensure both the informa-tion is not lost and that the public storedoes not grow infinitely. Gossiping is

extremely robust as it relies on very littleknowledge of the network's topology,and makes very fair use of availablecommunications resources without gen-erating hot-spots at popular producers orconsumers.

The goal of Construct is to provide asimple framework for the community ofpervasive and adaptive systems devel-opers. The platform addresses decentral-isation, robustness, fusion and otherproblems at an infrastructural level,allowing the developer of an individualsensor or service to focus on their ownnovel contribution. We have developed anumber of sensors, including interfacesto the PlaceLab and Ubisense locationsystems and virtual sensors for a numberof web sites, services and standard for-mats such as iCal. We are actively solic-iting the help of others to improve theplatform, its services and algorithms.

Construct is being licensed under LGPL.Our hope is that it will serve as a com-munity basis for rapid exploration andintegration of new pervasive, adaptiveand sensor-driven embedded intelli-gence.

This work is partially supported byScience Foundation Ireland under grantnumbers 04/RPI/1544, "Secure andPredictable Pervasive Computing" and05/RFP/CMS0062, "Towards aSemantics of Pervasive Computing".

LLiinnkkss::

http://www.construct-infrastructure.org

http://www.ucd.ie/csi


Steve Neely, University College Dublin /

IUA, Ireland


Towards a Platform for Widespread Embedded Intelligenceby Simon Dobson, Steve Neely, Graeme Stevenson, Lorcan Coyle and Paddy Nixon

The vision of pervasive computing is that objects, buildings and environmentsmay be endowed with software intelligence to improve human interactions bothwith the individual objects and with the system as a whole. Realising this dreamis posing significant challenges for designers - allowing individual applicationsto co-exist in a common space without interfering, making the capabilities of newsensors and services available to other applications as they appear and providingthe necessary decentralised control to obtain robust behaviour.





Various pervasive applications may beprovided in today's networked environ-ments. Figure 1 illustrates exploitation ofthe networked devices to create smartspaces for users. In such an environment,the nomadic user may, in particular, ben-efit from a large display nearby to watchmultimedia content provided by a localmedia server. The content will automati-cally be switched to the user's smartphone display as soon as other users getclose, to preserve the user's privacy. Thisapplication may further be extended toshare experiences with others. Forinstance, assume that the user is actuallyin a holiday resort, watching an adver-tisement for available excursions. Theuser may wish to get feedback about theexcursion, which may be obtainedthrough blogs supplied by other nearbyvisitors on their handhelds and/or theInternet. The user may even wish to lookat alternative excursions, whether adver-tised at nearby resorts, at tourist informa-tion offices or even on the Internet.Actuator resources such as displays andsoftware resources such as blogs, multi-media streams and advertisements on theWeb, all constitute networked capabili-ties that may conveniently be combinedto dynamically create a pervasive appli-cation providing a holiday planner fornomadic users.

Pervasive computing environments, likethe one shown above, embed networkeddevices, possibly wireless, from variousapplication domains, eg, mobile and per-sonal computing, consumer electronics,and home automation domains. Thisdiversity of devices, along with those ofnetworks and of software infrastructures,

call for computing paradigms that enableloose, dynamic coupling among hetero-geneous resources. Service-orientedarchitectures (SOA) have emerged as anappropriate computing paradigmresolving a number of heterogeneityissues. However, pervasive environ-ments have further introduced new chal-lenges for SOA both at middleware andat application layer. Regarding theformer, devices need to dynamicallydetect services available in the open net-worked environment and to interact with

them. This requires enabling interoper-ability of both discovery and communi-cation protocols, as services are imple-mented on top of diverse middlewareplatforms (eg, UPnP used in the homeand Jini in the mobile domain).Regarding the latter, applications need tocompose, in an ad hoc fashion, servicesavailable at the specific time and place,which have been developed indepen-dently without a priori mutual knowl-edge. The major challenges are enablingcommon understanding and adaptationof the semantics of the services beingassociated and interoperability between

service behaviour protocols. A signifi-cant enabler for dealing with these issuesis the Semantic Web and in particular theSemantic Web Services paradigms,which allow description of servicesemantics and behaviour and rigorousreasoning about the former. Tacklingboth semantics and behaviour hetero-geneity in a combined way is still anopen, challenging issue, further exacer-bated by the fact that semantic reasoningis a compute-intensive task, toodemanding for the wireless, resource-constrained devices of the pervasiveenvironment.

As part of the work of the INRIAARLES team, on the development ofdistributed systems enabling the perva-sive computing vision, we are devisingenablers for pervasive services. Wehave, in particular, investigated interop-erability enablers and their integration

within the base SOA architectural style.The resulting SOA style for pervasivecomputing offers two key functionali-ties, each one vertically integratingrelated advanced features in the middle-ware and in the application layer, asdepicted in Figure 2. First, semantic-richinteroperable service discovery enablesthe locating of networked servicesaccording to their semantic descriptions.In the middleware layer, discovery isachieved, despite the possible coexis-tence of various discovery protocols, byinterworking between them, which fur-ther includes middleware data type

Networking Semantic Services for Pervasive Computingby Valerie Issarny, Nikolaos Georgantas and Sonia Ben Mokhtar

With computing and communication capabilities now embedded in more andmore physical objects in the surrounding environment and with most userscarrying wireless computing devices, the pervasive computing vision pioneeredby Mark Weiser is being realised more than ever. Nevertheless, such an abundanceof ambient capabilities in the open, dynamic pervasive environment naturallycomes with high heterogeneity, which hinders their ad hoc integration into richpervasive applications. Within the INRIA ARLES team, we are investigating efficientsolutions to pervasive heterogeneity building upon semantic service-orientedarchitectures.

Figure 1: Integrating resources in user-centric smart spaces.

Networking Semantic Services for Pervasive Computing

By Valerie Issarny, Nikolaos Georgantas and Sonia Ben Mokhtar

With computing and communication capabilities now embedded in more and more physical objects of the surrounding environment and most users carrying wireless computing devices, the pervasive computing vision pioneered by Mark Weiser is more than ever becoming possible.Nevertheless, such abundance of ambient capabilities in the open, dynamic pervasive environment naturally comes with high heterogeneity, which hinders their ad hoc integrationinto rich pervasive applications. Within the INRIA ARLES team, we are investigating efficientsolutions to pervasive heterogeneity building upon semantic service-oriented architectures.

Various pervasive applications may be provisioned in today’s networked environments. Figure 1illustrates exploitation of the devices in the surroundings to create smart spaces for users. In such anenvironment, the nomadic user may in particular benefit from a large display that is nearby to watch multimedia content provided by a local media server, while the content will automatically be switchedto the user’s smart phone display as soon as other users get close to preserve the user’s privacy. This application may further be extended to share experience with others. For instance, assume that the useris actually in a holiday resort, watching an advertisement for available excursions. Then, the user maywill to get feedback about the excursion, which may be obtained through blogs maintained by other nearby visitors on their handhelds and/or the Internet. The user may even will to look at alternative excursions, whether advertised at nearby resorts, at tourist information offices or even on the Internet. Actuator resources like displays, and software resources like blogs, multimedia streams andadvertisements on the Web, all constitute networked capabilities that may conveniently be composedto realize a pervasive application enabling a holiday planner for nomadic users.

User detection

Figure 1. Integrating resources in user-centric smart spaces

Pervasive computing environments as the one exemplified above embed networked devices,possibly wireless, from various application domains, e.g., mobile and personal computing, consumerelectronics, and home automation domains. This diversity of devices along with the ones of networks and of software infrastructures call for computing paradigms that enable loose, dynamic coupling among heterogeneous resources. Service-oriented architectures (SOA) have emerged as a well-serving such paradigm resolving a number of heterogeneity issues. However, pervasive environments have further introduced new challenges for SOA both at middleware and at application layer. Regarding the former, devices need to dynamically detect services available in the open networked environment andto interact with them. This requires enabling interoperability of both discovery and communicationprotocols, as services are implemented on top of diverse middleware platforms (e.g., UPnP used in the home and Jini in the mobile domain). Regarding the latter, applications need to compose in an ad hoc fashion services available at the specific time and place, which have been developed independentlywithout a priori mutual knowledge. Major challenges concern enabling common understanding andadaptation of the semantics of the services being associated, and interoperability between service behavior protocols. A significant enabler for dealing with these issues is the Semantic Web and in particular the Semantic Web Services paradigms, which allow description of service semantics andbehavior and rigorous reasoning about the former. Tackling both semantics and behavior



system interworking. In the applicationlayer, semantics and behaviours of ser-vices are reasoned upon and matchedwith the requested ones in order todecide their adequacy for a given userrequest. Second, semantic-rich interop-erable service interaction enables ser-vices to cooperate and compose. In themiddleware layer, support is providedfor communication protocol and datatype system interworking. In the applica-tion layer, semantics and behaviours ofservices are adapted and interworked toenable their seamless interaction.Prototypes of the related architecturalcomponents have been implemented.

Performance evaluation shows that oursolutions comply with the requirementsof the pervasive computing environ-ment. We are now in the process of inte-grating them to offer an instance of theoverall architecture. We are furtherstudying extensions to the architecture toeffectively exploit beyond third genera-tion networks and related multi-radiowireless handheld devices; this willexpand the scope of our interoperabilitysolution to further cover radio and net-work heterogeneity.

Enabling the pervasive computing visionhas given rise to tremendous research

over the last decade. In particular, var-ious supporting software infrastructureshave been proposed, some enablingsemantic awareness for services, othersdealing with middleware interoper-ability. Our contribution lies in the inte-grated treatment of service interoper-ability, from the application to the mid-dleware layer. Our solution has furtherbeen designed to be lightweight and thuscan be supported in most pervasive envi-ronments, not requiring the presence of adedicated server. As a result, true open,pervasive computing environments areenabled - integrating and composing inan ad hoc fashion the most appropriatenetworked services in order to deliverrich applications to users, independentlyof heterogeneous underlying softwaretechnologies and semantic representa-tions.

Our research is conducted as part of theEuropean IST Amigo and Plastic pro-jects and industrial collaborations withAlcatel and Thales.

LLiinnkkss::

ARLES site: http://www-rocq.inria.fr/arles/


Valerie Issarny, Nikolaos Georgantas,

INRIA, France

Tel: +33 1 39 63 57 89

E-mail: [email protected],

[email protected]

Figure 2: Extending SOA with interoperability enablers.

Embedding sensing into a physical sub-strate has a number of attractions. Eachsensor package can sense a number oflocal variables such as the stress on thematerial, its orientation in space, itsproximity to other materials etc.

Combine these sensors into a networkand we can construct a global view of thematerial and its relationships to the realworld. Add processing and we have thepotential to build materials that "knowthemselves", in some sense, and which

can react in ways that are far moresophisticated than are possible with sim-pler, 'smart' materials.

If this all sounds a bit abstract, imagine aperson with a broken leg who is wearinga plaster cast. For a physiotherapist, thechallenge is to make the person take ade-quate exercise to stimulate the break,while at the same time stopping themfrom attempting too much and riskingfurther damage. As the physiotherapyprogramme changes over time and inconjunction with on-going assessment of

Materials with Intelligenceby Simon Dobson and Kieran Delaney

Sensor networks are the key enabling technology for building systems that adaptautonomously to their environment, without direct human intervention. Mostsensor networks operate in free air, but research being conducted in Ireland,between the School of Computer Science and Informatics at UCD Dublin and theCentre for Adaptive Wireless Systems at Cork IT, is starting to explore the toolsand techniques we need in order to build 'augmented materials' which combinesensing, actuation and processing into the fabric of built objects.


3388

the injury, the exercise required of thepatient and the optimal levels of rigidityand support required of the cast, will alsochange. If we embed sensing and pro-cessing into the cast, the network cansense the load being placed on the cast asthe person walks around. This can thenbe compared with a downloaded therapyprogramme and react, for example, byglowing green when things are fine, butflashing red lights if the person is over-doing their exercise. It is even possible tobuild materials with variable rigidity sothat the cast adapts the support it pro-vides over the course of treatment.

The individual elements of an aug-mented material can be based around any'mote' technology. We are using the plat-form developed by Ireland's TyndallNational Institute, made available viathat institute's National AccessProgramme. While current mote systemsare too large for use in practice, Tyndall's2.5cm-on-a-side motes (see figure) arebeing reduced to a 1cm form andbeyond, making them realistic forembedded use.

Since a single object might contain hun-dreds of elements, the elements them-selves need to be substantially config-ured, by making connections to neigh-bouring elements during curing, forexample. Changes to this network maycome from node failures, but may alsocome from physically significant eventssuch as cutting which manifests itself asa (rather structured) partitioning of thenetwork. Dealing with these changes in a

tractable way means developing a pro-gramming model that operates at thelevel of the complete material rather thanat the level of the individual elementswithin it. It should also be able to handlefailures and errors in sensing in the indi-vidual elements. This is not somethingthat is easy to do in traditional languagesand we are investigating some tech-niques pioneered in high-performancecomputing (skeletons and categoricaldata types) as a possible basis forbuilding self-managing applications ontop of the underlying unreliable sensorsand communications.

Augmented materials are in many waysthe ideal co-design challenge. The prop-erties of the material determine directlywhat can be sensed and processed, whilesoftware provides capabilities toenhance and complement the material'sunderlying physics. A physical phe-nomenon, such as placing one aug-mented object on top of another, givesrise to individual sensor readingsaffecting pressure, orientation and theestablishing of new wireless communi-cations links etc. These in turn give riseto a semantic inference that can be usedin software to drive high-level responsesbased on the intention inferred from per-forming this particular action with theseparticular objects.

While this work is still at a very earlystage, we hope that the programme willlead to useful advances in embeddedsensor networks, to new forms ofmicrosensing and actuation associated

with explicit software control, toimproved autonomic control of commu-nications and routing, and to generallyuseful programming models for sensor-rich multi-agent networks and environ-ments.

LLiinnkkss::

http://www.ucd.ie/csi

http://www.aws.cit.ie

http://www.tyndall.ie/research/mai-group/

25cube_mai.html


Simon Dobson, University College Dublin /

IUA, Ireland


A 2.5cm-on-a-side'mote', developedby Ireland'sTyndall NationalInstitute.



3399

CRUISE's NoE partners are combiningtheir efforts to make a significant contri-bution to both the coordination and theeffectiveness of research into the com-munication and application aspects ofWSN in Europe, promoting not onlyopen cooperation between individualsand organisations from Europeanacademia, but also research into cooper-ation with industry. The main aim forCRUISE is to push for a broad integra-tion of actions and to become one of the

focal points for these actions, with thestrategic goal of greatly strengtheningthe European position in this area ofresearch. CRUISE is to amalgamate themajority of European research entities inthe field of Wireless and SensorNetworks. It is doing this on the under-standing that this critical mass will putthe European Wireless and SensorNetworks research community in a posi-tion to successfully compete with the USand Asia and to provide support for

European industries in this field. Thesehigh-level objectives are pursued inactivities such as: • evaluating state-of-the-art wireless

sensor networking and collatingresearch topics covered in CRUISEand in Europe into a knowledge model

• depicting a path from current techno-logical status to a successful near-termand long-term vision, by creating thevision and roadmap on necessaryresearch in wireless sensor networkingin Europe

• implementing a set of well-definedjoint research activities and publishingwhite papers and a joint book pro-viding an integrated research view

• sharing and integrating methods, soft-ware tools and test-beds and makingan inventory of the existing platformsand evaluating of the experiencesusing them

• facilitating the mobility of researchersand PhD students as a key factor forintegration and rationalisation ofresearch efforts

• creating distributed PhD teaching cur-ricula and training curricula in sensornetworking and related topics;Organising regular CRUISE Schoolsas a tool for training young researchersand communicating with the industrystandardisation community and inter-national research community

• organising open collaboration eventswith industry, SMEs, academia, otherprojects with participation of expertsfrom Europe, the USA and Asia, ie,cluster meetings, workshops,roundtable discussions, web-basedtutorials, discussion forums, etc., toexchange and discuss ideas and to fur-ther elaborate thoughts exploitingCRUISE results.

The joint programme of research activi-ties (see figure), is structured in four self-contained but still inter-related clusters:Cluster A - Architecture and Topology;Cluster B - Protocols and data fusion;

The CRUISE Project - Network Initiative for Creating UbiquitosIntelligent Sensing Environmentsby Gabriele Kotsis, Anelia Mitseva and Neeli R. Prasad

Recent developments in research and technology have drawn attention to wirelesssensor networks as one of the key enabling technologies of the next ten years.ubiquitous intelligent sensing environments have a promising future in supportingthe everyday life of European citizens, bringing important social benefits for eachperson and for society as a whole. Taking into account the current fragmentationof European research in this field, the CRUISE Network of Excellence (NoE) aimsto be a focal point in the coordination of research on communication andapplication aspects of wireless sensor networking in Europe.

Network integration and research activities.




Cluster C - Security and Mobility;Cluster D – Transmission.

CRUISE Services to Industries,SMEs and the PublicOne of the main goals of CRUISE is toestablish efficient and durable links withindustry, using existing relationships andcollaboration frameworks as a lever.Each CRUISE partner is to furtherdevelop existing liaisons with industry.In order to provide mechanisms ofknowledge transfer, the project consor-tium is to offer Business Deck@CRUISEvia the project website. Members of'Business Deck' are offered access tosome of the internal documents preparedby the network. They could influence thechoice of topics for tutorials prepared bythe network and, to some extent, overtopics presented at training schoolsorganised by the partners. They could bealso involved in the tools' integrationactivities.

Cooperation with industry is also to becarried out through links with other IST

R&D projects working on close topics,such as eSENSE IP, MAGNETBEYOND IP, GoodFood IP, andthrough national projects. Special atten-tion will be paid to promoting collabora-tion with SMEs, which are expected toplay a particularly important role in thedeployment of sensor networks, butwhich often lack specific researchrecourses in order to complement theirown skills. Via the CRUISE Portal,SMEs will be able to reach a workinggroup or an appropriate expert from theproject with a specific problem to besolved. Adding to the above, CRUISE isto create links with other related IST pro-jects, such as NoE, e-NEXT, NoENEWCOM, NoE ACE, Coordinatedaction Embedded WiSeNets.

Last but not least, as part of the plan forusing and disseminating the knowledgeand to raise public awareness about theimmense social benefits of the sensornetwork applications, active collabora-tion will be pursued with professionalcommunities and organisations whose

interests span different aspects and sce-narios of sensing environments apartfrom the purely technical ones. This willserve the European community beyondacademia and science.

This work has been carried out withinthe framework of the IST-4-027738 NoECRUISE, which is partly funded by theEuropean Union. The project started onJanuary 1 2006 and has a duration of 24months. Thirty two internationallyrecognised groups from Europe,including the ERCIM members VTT andFraunhofer-FOKUS, are coordinated bythe Center for Telenfrastruktur (CTIF) atAalborg University, Denmark.

LLiinnkk::

http://www.ist-cruise.eu


Neeli R. Prasad

Aalborg University, Denmark


Embedded intelligence, or ambient intel-ligence, is the vision of the future whereinformation will be available to peoplewithout their being aware of the sur-rounding information technology itself.In the ambient intelligent world and inthe ubiquitous information society, com-puting is distributed all around in theenvironment. Users are mobile and com-munication is set up between individ-uals, between individuals and objectsand between objects. This communica-tion is obtained via 'invisible intelligentdevices' that will tag almost everythingin our society in an easy and friendly

way. These devices will receive, sense,carry, and distribute context informationin our environment.

Radio Frequency IDentification (RFID)is the key technology for local connec-tivity and an early vehicle for readyingthese future proactive computing sys-tems.

Nanoelectronics allows the mass produc-tion of such new silicon RFID deviceswith sensing capabilities and this makesmassive and low-cost tagging of objectsfeasible.

In this context, the aim of theIntelliSense RFID project is to developmulti-protocol RFID devices with built-in sensing capabilities, operating atmulti-frequency bands. The device canbe associated with an object, a person ora location through a simple ID and iscapable of measuring and acquiring dataabout the user's behaviour and his envi-ronment (such as temperature T, pres-sure P, humidity H, and pH), thus cre-ating a smart environment based on sur-rounding 'invisible intelligent devices'.

It is a two-year project (2006-2007) andis part of the Nordic research pro-gramme NORDITE. It involvesSINTEF (Norway) as the institute

IntelliSense RFID - An RFID Platform for AmbientIntelligence with Sensors Integration Capabilitiesby Ovidiu Vermesan, Nadine Pesonen, Cristina Rusu, Aarne Oja, Peter Enoksson and Helge Rustad

IntelliSense RFID project expands the technical and market potential of RFIDtechnology by developing multi-protocol devices with sensing capabilities thatare able to sense the environment and communicate at different frequency bands.



leading the project, VTT (Finland),Chalmers University of Technology(Sweden), and IMEGO AB (Sweden).The NORDITE programme aims at sup-porting research institutes and universi-ties from Sweden, Norway and Finlandto enhance state-of-the art research inthe fields of SW radio, wireless sensors,short-range wireless networks and RFIDor MEMS utilising RF technology. Theprogramme goal is to stimulate collabo-ration among national and Nordic coun-tries, create synergism, and consolidatetechnological and economic develop-ment opportunities for companies andindustries. These technology develop-ment research projects are funded byTEKES (Finland), VINNOVA(Sweden) and the Norwegian ResearchCouncil (Norway).

One functional area of great relevance tomany supply chain applications is theability to monitor environmental orstatus parameters using a RFID devicewith built-in sensor capabilities. Ageneric sensor interface is needed in theRFID platform to enable various appli-cations that can be realised by inte-grating the proper, low-cost sensor ele-ment to the RFID platform. TheIntelliSense RFID project focuses ondeveloping a generic sensor interface,integrated with the RFID circuitry, forexternal capacitive sensor elements. At

this stage, capacitive sensors will beused to measure temperature, pressure,humidity, and pH.

One of the major issues with large-scaleRFID networks is the lack of standardis-ation across many fronts. This rangesfrom the different data formats used, viareduced interoperability between RFIDreaders and tag/transponder devicesfrom different vendors, to interferenceproblems between RFID products fromdifferent manufacturers and the differentfrequency bands used in different coun-tries.

To overcome such problems, the 'globalsensing RFID devices' developed in theIntelliSense RFID project are designedto be able to communicate at differentfrequency bands with different commu-nication protocols. They can thus com-municate with the existing infrastructure(ie, existing RFID interrogators) in var-ious environments. Of special impor-tance is the combination of high fre-quency (HF) 13.56MHz and ultra highfrequency (UHF) 860-960MHz RFIDstandards. Indeed, multi-band tags areneeded to penetrate from logistics tomobile phone-based consumer markets.For example, the same dual-band tag canbe exploited for logistics, by using theUHF RFID interrogators, and by con-sumers, to retrieve the content of the

device memory by utilising the inte-grated HF RFID reader in a mobilephone. Applications that can benefitfrom such a dual-band RFID tag tech-nology include object authentication andinformation, personalisation, protectionagainst theft, maintenance, and recy-cling.

The IntelliSense RFID technology plat-form will address the HF, UHF andmicrowave (MW) 2.45GHz frequencybands focusing, in the first phase, on theintegration of the HF and UHF RFIDspecific protocols.

The whole infrastructure 'readers-RFIDdevices' will form the future 'Internet ofThings' that will be one of the main ele-ment of the 'Ambient Intelligence' and'Ubiquitous Information' society.


Ovidiu Vermesan, SINTEF ICT, Norway

Tel: +47 22 06 75 06


Multi-protocol RFID devices withbuilt-in sensing capabilitiesused for logistics and mobilephone-based consumerapplications.



An important aspect of ambient intelli-gence is to link information systems withreal work. To this end, sensors and net-works of sensors play an important rolein providing data and events from thereal world. Sensors and networks of sen-sors are thus gaining more and moreattention from the research community.Enabling cooperation between these tinydevices, in an extremely dynamic con-text, poses real challenges. In particular,we have to cope with very limitedresources (energy, processing power,memory), high failure rate and nodemobility. Another important issue is tooffer programming abstractions andsystem architectures to virtualise a col-lection of sensors such that it can be usedas a single virtual resource. The mainapplication domain for these devices ismonitoring at large (environmental,medical, industrial, etc.)

Surprisingly, much less attention hasbeen paid to what is becoming the mostimportant sensor network ever deployed:the multimedia phone. While these arewell-known as personal messagingdevices, their role as sensors has beenlargely ignored, in spite of their powerfulcapabilities in this respect. These devicescan capture sound, picture, video, cancooperate with other sensors such asGPS positionning receivers and, ofcourse, can communicate with global aswell as local networks.

An important challenge for theseexisting and promising sensing infras-tructures is to explore their potentialapplications and to address the problemsraised by their emergence as an unex-pected sensor network.

The ACES (ambient computing andembedded systems) research team at

INRIA has a strong background in thedomain of cooperating objects, whichmeans groups of devices which sponta-neously team up to form a short-livedinformation system (called spontaneousinformation system) in order to accom-plish a task. It has demonstrated severalapplications of this concept. Recently,the team proposed the concept of sponta-neous hotspots. This is where swarms ofmobile phones are used to increasedownload-efficiency of context-relatedresources in a GPRS network using blue-

tooth-enabled cooperation (SpontaneousHotspots: Sharing Context-DependantResources Using Bluetooth, Xavier LeBourdon, Paul Couderc, Michel BanâtreInternational Conference on Networkingand Services (ICNS'2006) July 16-19,2006 - Santa Clara, CA, USA). Thisapproach both balances the networkdownload cost between users and

reduces the network load, as well asdecreasing download times. A typicalusage scenario would be, for example, ata sports event, where many people in astadium would like to see an instantreplay of the recent action. In this con-text, the same resources (video of thereplay) are requested by many users. Thespontaneous hotspot allows the networkto balance the download and to flood thereplay into the stadium through blue-tooth.

We are now investigating a related, butmore innovative concept: collaborativecapture, where the shared data is nolonger resources hosted in the network,as in the previous case, but structuredinformation composed of elements con-

tributed dynamically by a collection ofnodes. Consider, for example, someonearriving late in a meeting. Many phones,treated as audio capture devices, couldhave recorded the beginning of themeeting to allow the newcomer to catchup with the missing part. Similar sce-narios with photos can be envisaged,allowing for example, a group of people

Collaborative Capture: A New Perspective for Sensor Networksby Paul Couderc

Ubiquitous and collaborative capture unveils new perspectives in ambientcomputing. In the same way that distributed computing proposed approaches tothe distribution of data and processing on a set of computing nodes, collaborativecapture will have to propose efficient and practical solutions to recomposestructured data from items collected asynchronously from contributing nodes.

Many phones, treated as audio capture devices, could have recorded the beginning ofthe meeting to allow the newcomer to catch up with the missing part.



taking pictures at an event to merge theircaptures and provide a complete collec-tion. Another interesting use of collabo-rative capture is to use several nodes toincrease the quality or the accuracy of agiven data, by combining their indi-vidual data.

These applications, which can appearsimple, raised difficult issues. Contextdescriptions of available resources andon-going capture processes should beaccurate enough to allow sharing,without overloading the local communi-cation medium to allow their efficientdistribution on-the-field. Obviously, pri-

vacy is a big issue. For example, youmay agree to share the pictures you aretaking at a given event with only a few ofthe people present, but not the picturestaken an hour ago in another context.Traditional security, with authorisations,is difficult to apply in practice in the con-text of spontaneous interactions withnew and potentially unknown nodes.Innovative approaches, based forexample on more dynamic notions suchas trust and behaviour, have to be inves-tigated before practical security mecha-nism for these systems can be found. Inrelation to trust, we must be able to rankthe quality of the contributions, which

may include poor data, or even falsedata. Worse, the very notion of ubiqui-tous capture can be frightening: thepotential capture activity of anyone, any-where may change social relationsbetween people. It's too late to do any-thing about that, this distributed infras-tructure is already in our pockets!

LLiinnkk::

http://www.irisa.fr/aces/


Paul Couderc, INRIA-IRISA, France


The economic relevance of this researchactivity can easily be seen by examiningthe efforts companies take to optimisethe energy consumption of embeddeddevices. The use of alternative energysupply, like Piezo-Elements or Thermo-Elements, will thus be a great chance toachieve this goal so as to minimise costsand to make devices independent fromexternal energy supply.

Energy management is not only consid-ered in terms of minimising the energyconsumption, but also as finding addi-tional ways for energy to be supplied.Furthermore, the energy-consumingactivities are bound to the current statusof energy production and the current costfunctions. For example, energy con-suming communication tasks are carriedout not continuously, but at optimal

opportunities by episodic transmissionprotocols. This principle is verified usingtwo very different application scenarios:skis and bicycles.

The scientific aim of the project is todevelop an innovative integrated frame-work and tool for modelling the energyand communication processes in net-worked embedded systems. Today'smodelling approaches and tools arebased on a combination of discreet mod-elling of the communication processesand an underlying model describing theenergy consumption. The modelling andsimulation of the energy processes insinks (sensors, actuators, physical com-munication) and sources (generator, bat-tery) are realised by means of continuousmodels.

This two-paradigm approach has its dis-advantages:• Handling two modelling approaches

and isolated tools is difficult for theresearchers/engineers and increasesthe duration of the development cycle.

• Interfaces between different modellingapproaches/tools have a potential formisinterpretation by the experts intheir respective domain.

Autarkic Power Generation for Networked Systems by Eberhard Mueller and Ulrich Hofman

A research project funded by the FIT-IT programme of the Austrian ministry fortraffic, innovation and technology (BMVIT) investigates the potential of differentenergy sources for communication purposes, especially for out-door activities.

Figure 1: Logical components of Unified Model.


4444

To overcome these drawbacks we modelthe data communication process by fol-lowing a new approach that we haveinvestigated in the European IST projectINTERMON. There, data traffic is mod-elled as a continuous (in time and value),fluid process. We bring this paradigminto the modelling domain of networkedembedded systems. Using the fluidabstraction, which was successfullyemployed in large scale IP networks,will open the way to a unified continuousmodelling framework. This frameworkwill accommodate the energy and com-munication models that will be devel-oped within the project. In a secondstage, the models will be validated andintegrated into a complex system. As aproof of concept this new integratedmodelling approach will be employedfor the design and the development of aprototype.

The main challenge is to optimise datacommunication, sensor data generationand storage of relevant data in the proto-type. This includes an autonomouspower supply which makes use of theforces, temperatures or other physicalphenomena that occur during the use ofsuch devices. State-of-the-art sensortechnologies, power sources and com-munication systems will be investigatedand their process models will be derived,measured and integrated into the contin-uous unified modelling framework.Measurement data of field trials will be

used for the validation of the scientificand technical approaches.

Figure 1 shows the logical componentsand models which are developed. The

send model includes continuous andepisodic transmission protocols abovethe Layer-2 WLAN-protocol (Figure 2).Figure 3 shows transfer intervals of amobile station to different base stations.

LLiinnkkss::

Salzburg Research:

http://www.salzburgresearch.at/company/

ESYCS - Embedded Systems Cluster

Salzburg: http://esycs.salzburgresearch.at/

INTERMON: Advanced architecture for

INTER-domain quality of service

MONitoring, modelling and visualisation:

http://www.ist-intermon.org/


Eberhard Mueller and Ulrich Hofmann

Salzburg Research GmbH, Austria

Tel: +43 662 2288 445


[email protected]

Figure 3: Data transfer to different base stations.

Figure 2: Optimal data

transfer to a basestation.

Figure 4: The author working with Piezo-Elements and Embedded System.


A major difficulty with constructingembedded systems is that the availablesystem resources (processor, computermemory, power) are necessarilyrestricted due to cost and other consider-ations. If it were possible to determinestrong bounds on the use of suchresources, then potentially there wouldbe significant benefits in terms of manu-facturing cost, reliability and perfor-mance. However, determining suchbounds automatically is a difficultproblem and doing so manually isbecoming impracticable as embeddedsoftware increases in complexity.

The EmBounded project, led by DrKevin Hammond at the University of StAndrews, aims to research this problem.Our vision is one where certificates ofthe bounds on resource usage can beobtained from a source program, throughautomatic program analysis, indepen-dently of the usual software compilationprocess. These certificates may then beverified using formal proof techniquesbased on a program logic that capturesthe meaning and behaviour of programs.

The EmBounded project builds onEuropean strengths in programming lan-guage design, program analysis andembedded applications. Our work is

based around a new domain-specificprogramming language notation, Hume,that aims to find an optimum pointbetween good programming abstractionand the ability to derive good cost infor-mation. Hume embeds a rule-basednotation for sequential computationswithin a finite-state-machine-based nota-tion for concurrency control. The pro-cess notation is designed to be easilytractable to standard analyses such asmodel-checking for deadlock detection;while the rule-based notation is designedso that we can easily expose informationabout cost. Hume is designed to be prac-tical as well as having interesting lan-guage properties. Prototype implementa-tions of Hume produced at Heriot-WattUniversity can run in less than 30KB ona standalone Renesas M32C develop-ment board, with minimal dynamic

memory requirements,depending on the end-applica-tion.

To date, project researchers atLudwig-Maximilians-Universität, München, and StAndrews University have devel-oped new theoretical models ofbounded resource usage fordynamic memory and time met-rics based on Hume programsource. They have related theseformally to the underlyingmachine operations using anabstract-machine approach. Weare now constructing automaticanalyses based on these modelsthat will expose resource con-straints and other program prop-erties in our program logic. Thiswill give a good model of pro-gram cost based on the programstructure.

In order to obtain accurate infor-mation about time usage on aspecific computer architecture,we are using the aiT tool devel-oped by AbsInt GmbH to provideprecise and guaranteed timingcosts for fragments of machinecode programs. This low-level

Formally Bounded Embedded Systemsby Kevin Hammond

The EU-funded EmBounded project aims to develop new techniques for providingformal guarantees on resource usage bounds. It is a collaborative project betweenAbsInt GmbH, a high-tech company specialising in high integrity real-time systems,and academic partners in the UK, France and Germany.

The 'EmBounded' vision.

embound, v.poet. arch.

trans. To set bounds to; to confine, contain, hem in.Hence embounded ppl. a.1595 SHAKESPEARE: The Life and Death of King John IV. iii 137

That sweete breath which was embounded in this beauteous clay.




information can then be combined withthe high-level information obtained fromour program source to provide a highlyaccurate analysis that will give guaran-teed bounds on program execution time.

In order to produce convincing results, itis necessary to test our approach usingsophisticated and realistic applications.Researchers at the Laboratoire desSciences et Matériaux d'Électronique —LASMEA (Clermont-Ferrand, France),

have been developing a number of real-time computer vision algorithms inHume. These algorithms can beexploited to detect, for example, roadfeatures from a moving vehicle, anessential component of any self-con-trolled (autonomous) road vehicle. Indue course, we intend to apply thesealgorithms to the sensor and control sys-tems used in the CyCab autonomousvehicle (http://www.robosoft.fr), a self-controlled electric car resembling a golf

buggy, and capable of speeds up to arather alarming 30 km/h.

LLiinnkkss::

Project: http://www.embounded.org

Hume: http://www.hume-lang.org.


Kevin Hammond, University of St Andrews,

Scotland, UK



Trust is an increasingly important phe-nomenon to grasp and support in openenvironments, such as where networkeddevices operate. A common scenariosees a Trustor, the subject of trust, searchfor a Trustee, the object of trust. TheTrustor's trustworthiness evaluationdirected to the Trustee should be facili-tated. Semi-automatic trustworthinessevaluation is of special relevance inembedded systems as well as in thesemantic web, or in the ubiquitous envi-ronments, where the Trustors andTrustees can be computer programmesin addition to human beings. To performan appropriate evaluation, Trustorsrequest various information character-ising Trustees. Typical solutions ask forusers' trust credentials, often expressedin terms of user profile, reputation, andrecommendations. The differencebetween reputation and recommendationhere is that reputation is based on theexperiences of the Trustor, whereas rec-ommendations are communicated expe-riences of others.

Contextual data, that is information con-sidered relevant to the interactionbetween a Trustor and a Trustee,including the environment in which this

interaction occurs, can influence theresult of the trust establishment. Forexample, acting in a safe environmentcan affect the trust that an applicationhas in a component. Context-awarenesshas been recognised in many researchareas of information technology, such asinformation filtering and retrieval, ser-vice provisioning and communication.However, the relationship between con-text and trust has not received very muchattention, apart from some occasionalwork. This is unfortunate, since suchrelationship can easily be recognised andits existence justified.

In the first phase of our research, we con-textualise trust-evaluation by consid-ering individual context attributes andassign them with values influencing thetrustworthiness evaluation process.Depending on the importance of a givencontext attribute, determined by what wecall a trust purpose, weights can beassigned to amplify or weaken therespective attribute's influence on theoverall trust. For example, the Trustee'slocation can have significant importancewhen deciding if to grant access to thelocal wireless network. A policy can beintroduced so that nobody (regardless of

possessing the network key or not) isallowed to use it if residing outside thecompany boundaries. We apply thenotion of context also to reputations andrecommendations. When dealing withreputations and recommendations, con-textual data can be used to emphasiseother experiences that have taken placeunder “similar enough” conditions tothose in which the Trustor currentlyfinds himself. Here “similarity” betweencontexts can be computed, for examplebased on reasoning or on context ontolo-gies, and expressed as minimal contex-tual distance.

The above-mentioned trust evaluationworks on the assumption that reputationdata and recommendations are directlyavailable to the Trustor. In the secondphase of our research, we extended theadaptability of context-aware trustwor-thiness evaluation by considering sce-narios where this direct information isnot necessarily available to the Trustor.In such cases, we highlight three sub-cases to be studied. First, the Trustee'sbehaviour in other contexts may beunknown. Secondly, the Trustee'sbehaviour in the current context may beunknown, for example due to absence ofknowledge referring to the current con-text. Thirdly, reliable recommendersand/or recommendations about theTrustee might not be available or theycould refer to the Trustee acting in dif-ferent context. We address each of these

Adaptable and Context-Aware Trustworthiness Evaluation by Gabriele Lenzini, Santtu Toivonen and Ilkka Uusitalo

Conducted as part of the EU-ITEA Trust4All project, which aims to enhancedependability in networked and component-based systems, VTT TechnicalResearch Centre of Finland and Dutch Telematica Instituut began investigatingtrustworthiness evaluation in ubiquitous environments. So far, the research hasconcerned context-aware trust and trust evaluation with indirect information.


4477

cases with separate proposals, all ofwhich are based on similarities betweenthe entities - Trustor, Trustee, recom-mender and other entities connected toeach other by a (social) network - or con-texts attached to the entities. Forexample, the lack of direct informationabout a new scientific conference doesnot stop the Trustor inferring trustwor-thiness by considering the reputation ofthe conference chair and favourable rec-ommendations about the programmecommittee members and the conferenceproceedings publisher.

Trust4All is a two-year EU-ITEA pro-ject. Trust4All aims to enhance depend-ability and trustworthiness of networkedand component-based embedded sys-tems. Trust4All project partners comefrom industry, research institutes, andacademia: CWI, Océ –Technologies,Philips Research (the project coordi-nator), Telematica Instituut, Univ.Eindhoven (TU/e), Univ. Leiden (theNetherlands); Nokia, Solid InformationTechnologies, VTT (Finland); ESI,FAGOR, IKERLAN-Electrónica,Robotiker, Visual Tools (Spain).

LLiinnkkss::

Telematica Instituut: http://www.telin.nl/

VTT: http://www.vtt.fi

EU-ITEA project Trust4All:

http://www.win.tue.nl/trust4all/


Gabriele Lenzini, Telematica Instituut,

The Netherlands


Santtu Toivonen, VTT Technical Research

Centre of Finland


Historically, personal relationships,face-to-face encounters, notaries, andthird party counsel verified our identity.Nowadays, reliance on paper has shiftedto electronic documents and in the sameway, reliance on traditional trust factorshas had to shift to electronic authentica-tion for exchanging information, goods,and services. As territories and commu-nities are newly defined in the digitalage, there is a need for a new definitionof digital identity. The question is: Dowe possess only one 'federal identity' asin our passport, or do we allow the use ofpartial identities, depending on the appli-cation we are using or on the relationshipor community of which we are a part?(See Figure 1.)

The research project BiometricAuthentication Supporting InvisibleSecurity (BASIS) tries to answer thesequestions. It aims to develop authentica-tion protocols in networked and dis-tributed systems and, in particular,studies robust face recognition in imagesfrom multiple cameras. Users will havean aware, adaptive and responsive per-sonal space, in the home environment.The aim is to investigate the possibilitiesof biometric authentication for securing

access to information and services in thispersonal environment, with a focus onuser convenience and privacy protection.

The research project addresses threetopics in the following work packages(WPs). The WP Transparent Biometricsis about transparent biometric authenti-cation (read without taking any specificaction) as a means to enhance user con-venience - for instance, to automaticallyhear your favourite music. The WPTemplate Protection is about anonymousbiometric authentication as a means of

protecting the user's privacy and the WPHome Biometrics is about the specificproblems of biometric authentication inthe home environment. Research atCWI, WP leader for Home Biometrics,started in January 2006.

In the home environment, it is unlikelythat biometrics based on the recognitionof one modality - such as speech orvision - will suffice. If we want to pro-vide the user with easy and transparentuse we should be able to identify sub-jects in changing and uncontrolled con-ditions. In the project, identity is notdefined as a unique set of personal data,like name or age, but more generally.Height, weight, or anything that can beused to classify user profiles is used. Forexample, we are not interested in the

Digital Identity in Ambient Environmentsby Ben Schouten and Onkar Ambekar

Embedded systems and ambient technology enable users to interact at any timeand anywhere. In the BASIS project for identity management, CWI investigatestransparent biometrics in home environments. Possible application areas areuser profiling for shopping , listening to one's favourite music and operatinggadgets and appliances in the home.

Figure 1: People share different identity data with different organizations and people.

Healthcare Government Music Shopping Friend

AddressAgeBlood groupDate of birthDiaryDietDrivers licenseGenderInsuranceLikes and dislikesNameNr of ChildrenPlace of birthTax Income

Application or Relationship

Par

tial i

dent

ities


Component-based Software Engineeringis an emerging development paradigmthat promises to accelerate softwaredevelopment and to reduce costs byassembling systems from pre-fabricatedsoftware components. It has also been

adopted for the development of softwarefor embedded devices.

There are several trends in component-based embedded devices. First, systemsare becoming more open in the sense that

more third party components (ie, softwarecomponents that have not been developedby the device vendor) are used for theirdevelopment. So, component-baseddevices run various components thatinteract with each other to provide thehigh-level system functionalities.Furthermore, systems are becoming moredynamic in the sense that the softwareconfiguration of a device changes whenthe device is owned and used by the enduser. These trends enable an increasingnumber of features to be offered by

Trust4All enhances Trustworthiness of Embedded Component-based Systemsby Gabriele Lenzini, Johan Muskens and Andrew Tokmakoff

As part of the EU-ITEA research programme, Trust4All project partners fromindustry, research institutes and universities teamed up to design and developan embedded systems framework that helps component-based applications tomaintain a high level of dependability and security. Philips research and the DutchTelematica Instituut designed and developed the framework's trustworthinessmanagement features.



name of the toddler who is approaching adangerous stairway in order to automati-cally close the stair gate.

With BASIS, information from differentsensors (speech, video etc) will be fusedand analysed to produce a coherent inter-pretation. The main complexity herestems from the heterogeneity of the datasources. As a consequence of small usergroups in the home environment itmakes sense to include so-calledlightweight or 'soft' biometrics likeweight, height or the location of a subjectto improve reliability.

Up to now, many ways of fusing datahave been investigated. In the case of arecognition system, they can roughly bedivided into four groups according to thelevel on which fusion takes place: pre-processing level, feature level, likeli-hood score level and decision level. Aswe will cope with uncertain inputs andchanging modularity, the best candidateis the likelihood score level. A verypromising candidate for solving thefusion task is to use a Bayesian network(BN) or an extension. Using BN, one cancope with uncertain inputs and priorknowledge (in the form of beliefs andstatistics), which could be incorporatedin a BN. Also Kalman-filters and HMMapproaches, used widely and success-fully for coping with noisy inputs and

tracking, could be expressed in terms ofBayesian networks.

The project outcomes can support futureapplications in smart environments,adaptive physical environments, pro-filing applications, and parental restric-tions (see Figure 2). In the future weforesee people carrying wireless identitytokens, for example in a handheld phone,RFID or implanted chip, constitutingpartial identities by which people presentthemselves, enabling them to authenti-cate transparently by using differentapplications. To avoid privacy con-straints and repudiation, leaving controlin the hand of the user, negotiation andinteraction in the authentication process,are essential. Just as in the 'old days'when police would take down your name

and address as a consequence of jumpinga traffic light.

The research consortium consists ofUniversity of Twente, University ofEindhoven, CWI and Philips HomeLaband is co-financed by IOP GenCom ofthe Dutch Ministery of EconomicAffairs. In the course of the four-yearproject the results will be demonstratedin the Philips Homelab. In the short timethe project has been in existence, resultshave already led to a patent application.

LLiinnkk::

http://www.cwi.nl/pna4


Ben Schouten

CWI and FONTYS, The Netherlands

Tel: +31 20 5924050


Figure 2:By ambienttechnology,

users canauthenticate

and profilevarious

applications.

Representation

Healthcare

Safety

Music Collection

Telecommunication

Shopping

EmailIntelligent Sensing

Environment

UsersFeedback

Sensors

Authentication

Applications

embedded devices, but threaten thedevices' dependability and security prop-erties. In the context of an open compo-nent-based system that supports unman-aged upgrading and extension (changingthe software configuration), it is possiblethat component-based devices may hostunstable components that could nega-tively affect the devices' dependability.Here, the most risky scenario is the down-load of new components to extend thesystem's capabilities; this exposes the sys-tem's internals to potentially maliciouscomponents that may compromise thesystem's overall security.

Recent research indicates that the use ofTrust Management can be a useful toolwhen addressing the dependability andsecurity concerns encountered in dis-tributed and embedded systems.

The Trust4All project has embraced thisemerging topic and investigated its appli-cation in the context of component-basedembedded systems. Our approach hasbeen to extend the component-basedmiddleware, first developed within theEU-ITEA Robocop and Space4U pro-jects, to include a TrustworthinessManagement Framework (TMF), asshown in the figure. This framework pro-vides low-level mechanisms that can beused to control and enforce components'behaviour, based upon establishedTrustor-Trustee relationship. The goal ofTMF is to support easy and late (possiblyruntime) integration of components andstill have dependability and securityproperties that are satisfactory to the user.

The TMF design addresses the followingchallenges: • The TMF acts on behalf of Trustors

and explicitly takes their trust require-ments into account.

• The TMF makes use of 'metricsources' to monitor and report on thesystem's overall behaviour.

• The TMF makes use of 'actuators' tocontrol a Trustee's mode of operation,thereby influencing the system'sbehavioural characteristics.

• The TMF makes use of the Space4UResource Management Framework todecide whether a component'srequested mode of operation should beallowed from a resource consumptionpoint of view.

A core part of the TMF is theTrustworthiness Evaluation Function(TEF), which is responsible for calcu-lating the trustworthiness of a compo-nent/Trustee. It can be parameterisedwith Trustor-specific dependability andsecurity requirements. By analysing aTrustee's 'estimated quality attributes'and recommendations, the TEF can eval-uate a component's 'present trustworthi-ness' (or compliance) in terms of thedegree to which it satisfies the Trustor'sstated dependability and securityrequirements. Through analysis of com-ponent's behaviour, the TEF calculates(utilising Subjective Logic) the 'believedtrustworthiness' of a component in termsof the following aspects:• Benignity: the belief that the compo-

nent will continue to satisfy theTrustor's requirements,

• Stability: the belief that the Trustee'sbehavioural qualities of will remainwithin a given neighbourhood.

The triple of compliance, benignity, andstability is used to make control deci-sions that regulate the system's overalldependability and security characteris-tics. For example, when a componentwith positive compliance begins to showa reducing benignity, the TMF maydecide to 'wrap' the component in anattempt to minimise its potential impacton the system, since it is displayingincreasingly less-compliant behaviour.Similarly, when there is a strong disbe-lief in stability, this may indicate a need

to re-evaluate the component's estimatedquality attributes and, as a consequence,re-evaluate its compliance.

The project began in July 2005 and willrun for two years. Its efforts are nowmoving from research and designtowards implementation and construc-tion of demonstrators in the domains ofconsumer electronics, mobile devices,and domotics.

Many companies and researchers collab-orate in Trust4All: CWI, Océ–Technologies, Philips Research (pro-ject coordinator), Telematica Instituut,Eindhoven University of Technology(TU/e), Univ. Leiden (the Netherlands).Nokia, Solid Information Technologies,VTT (Finland), ESI, FAGOR,IKERLAN-Electrónica, Robotiker, andVisual Tools (Spain).

LLiinnkkss::

http://www.research.philips.com

http://www.telin.nl

http://www.win.tue.nl/trust4all


Gabriele Lenzini, Telematica Instituut,

The Netherlands

Tel: + 31 53 4850463


Johan Muskens, Philips Research Europe,

The Netherlands

Tel: + 31 40 2742491


The Trust Management Framework, as part of the middleware of the embeddedcomponent-based devices, ensures the dependability of the system, when a newcomponent is added.



Pict

ure:

Tel

emat

ica

Inst

ituut

and

Phi

lips.


Articles in this Section50 Nanotechnology

Single Molecule Switch andMemory Elementby Emanuel Lörtscher and HeikeRiel, IBM Zurich Research Lab,Switzerland

52 Mobile Hybrid Wireless Networks

HyMN: A Self-Organizing Systemfor Interest-based DataDistribution in Mobile HybridWireless Networksby Steffen Rothkugel, Matthias R.Brust and Adrian Andronache,University of Luxembourg

53 Visualization

Visualization of ColourInformation on Highly Detailed 3DModelsby Matteo Dellepiane and MarcoCallieri, ISTI-CNR, Italy

55 Databases

Modular Access to Large Socio-Economic Databases - the SedoProject in Luxembourgby Uwe Warner, Centre d'Etude dePopulation, de Pauvreté et dePolitiques Socio-Economiques, andThomas Tamisier and Fernand FeltzCentre de Recherche Public -Gabriel Lippmann, Luxembourg

56 Digital Libraries

BELIEF: Bringing Europe'sElectronic Infrastructures toExpanding Frontiersby Federico Biagini, PasqualePagano and Franco Zoppi, ISTI-CNR, Italy

58 Electromagnetic Radiation

Cell Phone Dangerby Martin Röösli, Unversity of Bern,and Harry Rudin, Consultant,Switzerland

60 Satellite Communication Services

SAFE: Satellites for Epidemiologyand Health Early Warningby Catherine Chronaki, ICS-FORTHand Laurent Braak,

61 Satellite Devices

Planck Mission 70 GHz Receiversby Jussi Varis, MilliLab-MillimetreWave Laboratory of Finland

62 Detector Technology

Data Acquisition in ScientificApplicationsby Rob Halsall, CCLC, UK

63 Security

Network Anomaly Detection byMeans of Machine Learningby Roland Kwitt, Salzburg ResearchGmbH, Austria

64 Web 2.0 and Novel Interaction

From Solitary Podcasting to Ambient Pondcasting by Emmanuel Frécon and PärHansson, SICS


In the August 4 issue of SMALL1, wereport on a single-molecule switch andmemory element. Using a sophisticatedmechanical method, we were able toestablish electrical contact with an indi-vidual molecule to demonstratereversible and controllable switchingbetween two distinct conductive states.This investigation is part of our work toexplore and characterize molecules tobecome possible building blocks forfuture memory and logic applications.With dimensions of a single molecule onthe order of one nanometer (one mil-lionth of a millimeter), molecular elec-tronics redefines the ultimate limit ofminiaturization far beyond that oftoday's silicon-based technology.

The results show that these moleculesexhibit properties that can be utilized toperform the same logic operations asused in today's information technology.Namely, by applying voltage pulses tothe molecule, it can be controllablyswitched between two distinct 'on' and'off' states. These correspond to the '0'and '1' states on which data storage isbased. Moreover, both conductive statesare stable and enable non-destructiveread-out of the bit state - a prerequisitefor nonvolatile memory operation -which we demonstrated by performingrepeated write-read-erase-read cycles.With this single-molecule memory ele-ment, we have documented more than500 switching cycles and switchingtimes in the microsecond range.

Crucial for investigating the inherentproperties of molecules is the ability todeal with them individually. To do this,we extended a method called themechanically controllable break-junc-

tion (MCBJ). With this technique, ametallic bridge on an insulating substrateis carefully stretched by mechanicalbending. Ultimately the bridge breaks,creating two separate electrodes thatpossess atomic-sized tips. The gapbetween the electrodes can be controlledwith picometer (one thousandth of ananometer) accuracy due to the veryhigh transmission ratio of the bendingmechanism. In a next step, a solution ofthe organic molecules is deposited on topof the electrodes. As the junction closes,a molecule capable of chemicallybonding to both metallic electrodes canbridge the gap. In this way, an individualmolecule is 'caught' between the elec-trodes, and measurements can be per-formed.

The molecules investigated are speciallydesigned organic molecules measuringonly about 1.5 nanometers in length,approximately one hundredth of a state-of-the-art CMOS element. The moleculewas designed and synthesized byProfessor James M. Tour and co-workersat Rice University, Houston, USA.

The main advantage of exploiting trans-port capabilities at the molecular scale isthat the fundamental building blocks aremuch smaller than today's CMOS ele-ments. Furthermore, chemical synthesisproduces completely identicalmolecules, which, in principle, arebuilding blocks with no variability. Thisallows us to avoid a known problem thatCMOS devices face as they are scaled toever smaller dimensions. In addition, wehope to discover possibly novel, yetunknown properties that silicon andrelated materials do not have.

Nanotechnology

Single-Molecule Switch and Memory Elementby Emanuel Lörtscher and Heike Riel

Scientists at the IBM Zurich Research Laboratory have demonstrated how a singlemolecule can be switched between two distinct conductive states, which allowsit to store data. As published in SMALL, these experiments show that certaintypes of molecules reveal intrinsic molecular functionalities that are comparableto devices used in today's semiconductor technology. This finding is a resultcoming from IBM's Research Labs' effort to explore and develop noveltechnologies for beyond the CMOS (Complementary Metal Oxide Semi-conductor)era.

1 (E. Lörtscher, J. W. Ciszek, J. Tour, and H. Riel, "Reversible and Controllable Switching of a Single-Molecule Junction",Small, Volume 2, Issue 8-9 , pp. 973-977, 4 August 2006)

51


Promising Nanotechnology for Beyond the CMOS Era The single-molecule switch is part of aseries of results achieved by IBMresearchers in their efforts to explore anddevelop novel technologies that will sur-pass conventional CMOS technology.Miniaturizing the basic building blocksof semiconductor chips, therebyachieving more functionality on thesame area, is also referred to as scaling,which is the main principle driving thesemiconductor industry. Known as'Moore's Law', which states that the tran-sistor density of semiconductor chips

will double roughly every 18 months,this principle has governed the chipindustry for the past 40 years. The resulthas been the most dramatic andunequaled increase in performance everknown.

However, CMOS technology will reachits ultimate limits in 10 to 15 years. Aschip structures, which currently havedimensions of about 40 nm, continue toshrink below the 20 nm mark, ever morecomplex challenges arise and scalingappears to be no longer economicallyfeasible. Below 10 nm, the fundamental

physical limits of CMOS technologywill be reached. Therefore, novel con-cepts are needed.

LLiinnkk::

http://www3.interscience.wiley.com/

cgi-bin/abstract/112713559/ABSTRACT


Emanuel Lörtscher and Heike Riel, IBM

Zurich Research Laboratory, Switzerland

E-mail: {eml,hei}@zurich.ibm.com

Figure 1: Contacting a single molecule: Scanning electronmicroscope image of a metallic bridge, used forestablishing contact to an individual molecule.

Atomic-sized tips, serving as electrodes, arecreated by stretching and finally breaking this

'nano' bridge. The method is called themechanically controllable break junction (MCBJ).

The sketch at the top shows the switchingmolecule "caught" between the electrodes by

closing the bridge gradually with picometeraccuracy until one molecule "reaches" both

electrodes. The switching molecule is a speciallydesigned organic molecule (bipyridyl-dinitro

oligophenylene-ethynylene) with dimensions oftypically 1.5 nm.

Figure 2: Memory operation of the single-moleculesystem. The blue line shows the write, read, anderase pulse pattern applied and the red linedemonstrates the resulting switching between"off" and "on" states of the molecular system.As can be seen in the picture, the system isinitially in the "off" state. Then a write pulse of+1.6 V is applied and the molecule switches tothe "on" state. This state can be read out usinga voltage of +1.1 V. The molecule can beswitched back to the "off" state by anotherpulse (erase pulse) of -1.6 V.



RR&&DD AANNDD TTEECCHHNNOOLLOOGGYY TTRRAANNSSFFEERR

Ad-hoc networks emphasise flexibilityand survivability of the whole system.However, centralised approaches, forexample for group management andinformation provisioning, do not workwell in such settings. Moreover, due tofrequent topology changes, connec-tivity of devices cannot be generallyguaranteed. In particular, this makes ithard to disseminate information in areliable way. We overcome these limi-tations inherent to pure ad-hoc net-works by (a) establishing local groupsof communicating devices in a self-organising manner and (b) introducingdedicated uplinks to a backbone infras-tructure. Such uplinks are used foraccessing resources available on theInternet. Additionally, they areemployed to directly interconnect dis-tant devices, either within a single parti-tion and particularly also across dif-ferent partitions. In practice, uplinks arerealised for instance using cellular net-works, satellites, or via Wi-Fi hotspots.Consequently, ad-hoc networks withdevices that provide uplinks are calledhybrid wireless networks throughoutthis paper.

One main goal of the HyWercs project isto develop a middleware that enablesmobile devices to organise themselves inorder to optimise the data access andprovisioning process. The devices run-ning the middleware keep track of theuser's interests, which can be single dataitems or topics that comprise several dataitems. Devices sharing common interestswill join interest groups. Whenever thereis the need to retrieve information from abackbone network - like the Internet - or

to establish a communication link to adevice in another network partition,uplinks are employed. The mobiledevices elect a particular member of theinterest group, a so-called injectionpoint, which maintains the uplinks.

As one proof of concept, the HyMNsystem (Hybrid Multimedia Network)has been implemented prototypically ontop of the JANE1 simulation environ-ment. JANE is designed to support appli-

cation and protocol design in the realmof ad-hoc networks. One distinct featureof JANE is that applications can be exe-cuted on top of simulated devices (seefigure) as well as in so-called platformmode on real hardware.

HyMN is designed for users interested inlive multimedia news, for example fromcertain sports events. Mobile devices offootball fans, for instance, create aninterest group in a local ad-hoc networkpartition. Multiples of such groups co-exist, such as football fans in pubs, thosewatching another match, traveling fansand more. In each of these cases, a con-siderable number of people have ashared interest and might join forces in alocal setting.

Mobile Hybrid Wireless Networks

HyMN: A Self-Organizing System for Interest-based Data Distribution in Mobile Hybrid Wireless Networksby Steffen Rothkugel, Matthias R. Brust, Adrian Andronache

The concept of self-organisation already exists in the realm of ad-hoc networks.The inherent limitations of pure ad-hoc networks can be overcome by introducinguplinks to the Internet backbone. Extending the principles of self-organisationby transparently including those additional network links is the focus of theHyWercs project. HyMN, in turn, is a system aimed at fostering user experienceby employing these concepts in the domain of multimedia content distribution.

The HyMN System in simulation mode.

5533


The devices in the ad-hoc network run-ning HyMN organise themselves inclusters, where the clusterheads act asinjection points. For each group, interestin certain football matches is registeredwith the injection point. The injectionpoint maintains an uplink to the back-bone network in order to receive multi-media news related to the interests of thead-hoc members. Thus, the football fanswill receive injected information such assmall videos, pictures or text messageseach time something interesting is hap-pening during the match. The multi-media files received remain stored onthe mobile devices and will be providedto further devices joining the interestgroup later on.

In order to optimise the injection pro-cess, the backbone multimedia files aresplit into chunks. These chunks areinjected concurrently to different injec-

tion points within a single ad-hoc net-work partition. Thus, the bandwidth forinjecting data into a single partition isincreased and the files reach the ad-hocnetwork faster. After reaching the net-work, the chunks are exchanged amongthe interested devices until all devicesreceive the complete file.

To realise the architecture presented, wefocus on mobile devices establishingclusters in a self-organising way. Forthis, we have developed a clusteringalgorithm - namely WACA - thatemploys a heuristic weight function. TheWACA algorithm (WeightedApplication aware ClusteringAlgorithm) is designed to build an ad-hoc network topology that fits the needsof the application running on top of it. Toachieve this, several parameters can beset in the weight function. The HyMNapplication focuses on multimedia con-

tent distribution from a backbone net-work to ad-hoc networks. To optimisethis process, parameters like signalstrength to the backbone network, longbattery lifetime, dissemination degreeand clustering coefficient are used forclusterhead election.

On-going research focuses on investi-gating how the mobile devices deal withdifferent mobility scenarios. The HyMNsystem will be further developed and amiddleware for wireless hybrid networkdevices will be developed from it.

LLiinnkk::

http://mocca.uni.lu

http://syssoft.uni-trier.de/jane


Steffen Rothkugel,

University of Luxembourg

Tel: +352 466644 5259



The evolution of technology and impor-tant advances in the field of visualizationof huge 3D datasets mean that it is nowpossible to acquire and display detailed3D models. However, in order to achievea completely realistic result, high qualitycolour information must be added to thegeometric structure.

A coloured model can be created startingfrom: a 3D dataset, a set of photos (suchas those shown in Figure 1) and calibra-tion data, ie, the values of the parametersof the camera that took the photos. Theycan be divided into two groups: extrinsicparameters (translation vector and rota-tion matrix), related to the position of thecamera in the space, and intrinsic param-eters, the 'internal' settings (focal length,lens distortion) of the camera.Unfortunately, in most cases, calibration

data are not available and have to be esti-mated.

Camera parameters are calculated byaligning each photo to the model: sincethis can only be done automatically inparticular cases (eg, 'shape from silhou-ette'), user intervention is necessary.

Algorithms which align a photo on amodel need some correspondencesbetween them: the user must indicatesome corresponding points on bothobjects.

Our software, called TexAlign, allowsusers to load both the model and all thephotos, creating an Alignment Processwhose data (correspondences coordi-nates, parameters of aligned images) aresaved in an xml file. It is also possible to

set correspondences not only between aphoto and the model, but also betweenphotos, using overlapping sections. The'image to image' correspondence is usedby the application to infer new corre-spondences with the model. A 'workloadminimizer' analyses the graph of corre-spondences between the elements of theAlignment Process and helps the user tocomplete all alignments very quickly.

Once calibration data have been esti-mated, we have to 'project' the colourinformation onto the model. In 3Dgraphics, there are two ways to displaycolour on a model: texture mapping,where a single texture image is 'mapped'to the model, and 'per vertex' colour,where a RGB colour value is assigned toeach vertex. The first technique givesvery good results; the main limitation isthat the photos have to be 'packed' in animage of maximum 4096x4096 resolu-tion. Since we normally deal with tens ofphotos, we would need to subsample

Visualization

Visualization of Colour Information on Highly Detailed 3D Modelsby Matteo Dellepiane and Marco Callieri

It is now possible to display very detailed colour information on big 3D models.Research at the Visual Computing Lab, ISTI-CNR, Pisa, Italy, aims at creating amostly automatic, easy-to-use way of mapping colour on geometric data.

5544


them in order to put all of them in asingle image. Hence, since we can pro-vide very detailed 3D models, colour pervertex is the best choice to preserve bothgeometric and colour detail. We assignto each vertex of the model the colour ofthe images which project on it. Sincemore than one image could project ontothe same vertex, a major issue rises: howcan we automatically 'weigh' the contri-bution of each photo in order to achieve arealistic result?

To solve this problem, we developed anapplication called TexTailor. Startingfrom the set of photos, the model and thecalibration data, TexTailor automaticallycreates a mask for each photo, assigninga weight to each pixel. The weight is cal-culated by considering three mainvalues: the angle between the normal ofthe vertex associated with the pixel andthe direction of view, the distancebetween the point of view and the vertex(depth) and the distance of the pixel from

a discontinuity in the 'depth map' (that isa map where each pixel has the value ofthe depth of the associated vertex).These values are combined to calculate aweight. The colour value of each vertexis a 'weighted sum' of the contributionsof all images. A result achieved usingthis approach is shown in Figure 2.

Even though the results so far areencouraging, more work is needed toimprove the technique. The mostpromising area for future investigation isrelated to the estimation of the lightingenvironment. This can be done usingknown techniques (ie mirrored spheres)during the photographic campaign, or bycombining image processing and anal-ysis of geometric information during theprojection operation. A good estimationof lighting would prevent the projectionof dark areas in photos and could alsoprovide information about thereflectance properties of the material.

LLiinnkk::

Visual Computing Lab at ISTI-CNR:

http://vcg.isti.cnr.it/joomla/index.php

PPlleeaassee ccoonnttaacctt::::

Matteo Dellepiane, Marco Callieri,

ISTI-CNR, Italy

Tel: +39 0503152925

E-mail: dellepiane|[email protected]

Figure 1: An example of a 3D model (centreimage) with a set of photos. Theresult will be the 3D model with thecolour information extracted fromthe photos.

Figure 2: Details of the model of a woodenstatue in Benabbio (Lucca): 36 photos projected on a 2.5 million face model.


5555


The use of the Net for querying socio-economic databases has recently caughtthe attention of both experts and thepublic at large, with a number of onlineservices provided by statistical organisa-tions of various countries, such as thosein Europe that belong to the CESSDA.Not only does the networked publicationof statistics allow instantaneous updateand correction, but it is also suited tobuilding a relationship between the dataprovider and the data consumer. Theconsumer can deliver his feedback andformalise his requirements and theprovider has the means to control the useof the data, so that the figures are betterand more properly used. Other advan-tages are easing the retrieval of largedata sets, allowing them to be cus-tomised, and enabling end-users to exe-cute statistical functions for processingthe data on the server.

The CEPS/INSTEAD is a publicresearch institute for statistics and socio-economic studies dedicated to popula-tion, poverty and socio-economic policy.Through the Sedo (Socio-EconomicDatabase Online) platform, set up in co-operation with the Centre de RecherchePublic - Gabriel Lippmann, it offers ageneralised and convenient access to thevoluminous database on the nationalpopulation. The information, collectedthrough socio-economic householdpanels, consists of longitudinal data,which means that the same questions areperiodically checked with the samerespondents. Relevant domains are:demography, education, post-natalchild-care, employment, equal opportu-nities, resources, housing, equipmentand consumption and opinions about themonetary situation of the household and

its evolution. Such exhaustive and dis-parate information is difficult to exploitdue to the complexity and volume of thedata. Up to now, only researchers with asolid background in longitudinal anal-ysis could effectively benefit from thiswork, whereas it is relevant to all kindsof professional activities outside the areaof the statistician experts. In addition,thanks to its precision and its constant

updating, the socio-economic picturedrawn by the combination of these mis-cellaneous themes is apt to significantlyease communication between the popu-lation and public servants and to supportthe decisions made by the politicalauthorities.

Since autumn 2005, Sedo has beenavailable online, featuring three dif-ferent types of user access. A monitoredfree access for everybody to statisticaltables and graphs on the themes of thestudy. A controlled access on demandusing a user login and password thatallows registered users to explore thebanks of data without having to installstatistical software on their own com-puters. An access of registered users,making a confidentiality pledge who aretrained in statistical analysis to maintainindividual and sophisticated statisticsand research on the micro data. Finally,an Internet based discussion forumallows all users to exchange their expe-

rience and allows communication withthose conducting the surveys and theresearch community. Compared to otherinformation tools in the community, theSedo project has great flexibility, a uni-versal portability and the possibility ofadapting to different databases. Themost obvious benefit of the system is

Databases

Modular Access to Large Socio-EconomicDatabases - the Sedo Project in Luxembourgby Uwe Warner, Thomas Tamisier and Fernand Feltz

Now online and regularly accessed, the Sedo homepage is the entry point to amine of useful information about citizens' lives in Luxembourg. Underneath, thesite is powered by a customisable platform for maintaining huge banks of data.Ergonomic search methods quickly deliver mains trends or detailed figures.Advanced users and people without statistical training can perform their owncalculations with the guarantee of clear and accurate results.

Sedo's results for the public at large.



The BELIEF digital library (DL) offersuniform access to multimedia documen-tation (eg technical reports, presenta-tions, videos, manuals, on-line tutorials)providing continuously updated infor-mation on eInfrastructure-related pro-jects, initiatives and events. The contentsare harvested from different sources,such as projects web sites, repositoriesand databases. The DL provides servicesto support the submission, description,searching, browsing, retrieval, access,preservation and visualization of multi-media documents. Although designed tomeet the needs of a specific community,the technology adopted by BELIEF can

be easily adapted to meet the informationand collaborative needs of other scien-tific communities.

In the BELIEF DL, users can define theinformation space they wish to searchor browse in terms of the collections (ie,sets of documents) selected from thosemanaged by the DL. Differentsearch/browse options are offered:Google-like or fielded (with fieldsselected from a variety of known meta-data formats), with or without relevancefeedback. Users can search/browse anyinformation associated with digital doc-uments and their parts.

Supporting e-Infrastructure ProjectsProjects can submit their documentsdirectly to the digital library: BELIEFoffers a Submission Interface and aDocument Hosting Service to host pro-jects' documents directly on the DL. Toensure that all documentation can beeasily acquired, the DL stipulates a stan-dard description format.

Projects using the Document HostingService do not need to bear any extracost for the development of tools or ser-vices for their documents' submission,collection and retrieval. Projects thatprefer to maintain their own documentscan make them available to the commu-nity by simply providing a standardaccess to their documents and metadataso that they can be transparently 'har-vested' by the DL.

Digital Libraries

BELIEF: Bringing Europe's ElectronicInfrastructures to Expanding Frontiers by Federico Biagini, Pasquale Pagano and Franco Zoppi

BELIEF aims at creating a platform where e-infrastructure providers and userscan collaborate and exchange knowledge, thus bridging the gap separating theresearch infrastructure providers from the users. A key role is played by theBELIEF digital library, specifically designed to meet the needs of this communityand to guarantee wide dissemination of its results.

being able to obtain customised presen-tations of statistics with guaranteedaccuracy and without the need for spe-cialised statistical tools or an advancedexpertise in statistical techniques.

For the sake of the reusability, it hasbeen preferred to build an integratedplatform rather than developing featureson top of existing tools. This approachgives room for the integration of pow-erful statistical tools and lets the userperform online calculations according tothe data. With every single variable orresult are associated information typessuch as the statistical measurement(metric, ordinal, nominal), the popula-tion concerned (individuals, house-holds), or the expected results for thevariable. Based on these types, thesystem defines the statistical analysisthat can be performed in order to ensurethe coherence and soundness of the oper-ations delivered to the user. To guarantee

the coherence of the results and to ensurethat they are correct with respect to theinterpretation criteria of the user, Sedoalso features an automatic calculation ofthe actual sampling size after a sequenceof statistical operations.

From the programmer's point of view,full latitude is allowed for the integrationof thesauruses and the customisation ofthe search-engine. The system is alsocharacterised by a total freedom in thechoice and the management of thedatabases.

With Sedo, the database server can bedistinct from the server where the clientrequests are processed and the statisticaloperations are performed. Thanks to thisindependence of data storage and dataprocessing, Sedo is portable to any kindof database. This separation is alsouseful for protecting the confidentialityof the data. In this case we must consider

the different kind of users who will notbe entitled to the same access to the data.It also minimises the volume of dataused during the calculations, becauseonly the data relevant to the requests isextracted and passed on to the processingphase.

LLiinnkkss::

http://sedo.ceps.lu

http://www.ceps.lu

http://www.lippmann.lu


Uwe Warner, Centre d'Etude de

Population, de Pauvreté et de Politiques

Socio-Economiques, Luxembourg

Tel: +352 58 58 55 554


Thomas Tamisier, Centre de Recherche

Public - Gabriel Lippmann, Luxembourg

Tel: +352 47 02 61 622.


56


57


Supporting Users' Community NeedsThe BELIEF DL facilitates access to anddissemination of eInfrastructure projectdocuments as it provides the user com-munity with a “one-stop-shop” where allrelevant documentation can be found.The DL provides a wide range of infor-mation on eInfrastructures, how theywork, what projects exist and how to usethem. Even documents stored in innersections of the websites – and thus notindexed by general purpose searchengines - are accessible via BELIEF.The user can search documents in a uni-form way, and access documents pro-duced by different providers via a uniqueinterface. The DL knows the user'sinterest, as all users submit user profileinformation on initial registration. TheDL can thus provide users with the mostrelevant documents according to theirrequirements and profile, thus providingan intelligent bridge between informa-tion provider and seeker. The figureshows the display of results for a searchfor documents containing information onthe Globus and Glite systems plus theCondor project.

Technical IssuesWe have concentrated on implementingharvesting tools that support the use ofstandard formats and conventions.BELIEF currently supports the fol-lowing standard protocols and codingformats:

• Metadata encoding protocols: DC,DCQ (recommended), MARC, UNI-MARC, MARC21.

• Metadata harvesting protocols: OAI-PMH - Open Archives InitiativeProtocol for Metadata Harvesting (rec-ommended) or any API call returningan XML file containing metadataencoded in one of the above men-tioned encoding.

• File formats: XML (recommended),RSS.

Although information sources can beaccessed via multiple protocols and for-mats, BELIEF has chosen to use DublinCore Qualified encoding for metadatadefinition as it facilitates sharing ofinformation between sources using dif-ferent encoding systems with no loss ofsemantics.

A web-based interface allows easy sub-mission of metadata and documents tothe DL and 'document submissionmodels' have been provided to cope withthe specific characteristics of the docu-ments managed by the DL (presenta-tions, articles, books, tutorials, etc.). Thismodel is easily customizable, even byend-users, and new models can be addedby DL administrators at any time.

For developers interacting with the DL aset of APIs has been implemented. TheBELIEF DL was created as an instance

of OpenDLib (see Links), some of theseAPIs were already available as part ofthe OpenDLib Toolkit. However, inorder to cope with emerging require-ments a number of new APIs have beenimplemented; these are now included inthe OpenDLib Toolkit (Browse,BrowseInfo, BrowseStandard,GetDocumentsPerTerm, Submit,ListCollections).

Future WorkFurther activity is planned to add addi-tional administration functionality toensure smooth management of wideruser communities nad more friendly helpfacilities for end-user. The SubmissionService will be enhanced and the AccessService will be extended to support newprotocols and formats to assist metadatacompatibility over heterogeneoussources.

LLiinnkkss::

Networked Multimedia Information Systems

Laboratory website:

http://www.isti.cnr.it/ResearchUnits/Labs/

nmis-lab/

OpenDLib website:

http://www.opendlib.com

BELIEF project website:

http://www.beliefproject.org


Pasquale Pagano, ISTI-CNR, Italy


BELIEF digital library webinterface.




Our dependence on and investment incell phones rise steadily. The annual saleof cell phones is approaching 1 billionunits. Is the result of all the accompa-nying electromagnetic radiation going tohave an effect on our health? Almosteveryone has heard some 'anecdotal evi-dence' where there have been someundesirable effects on people: headache,concentration problems, irritability,insomnia, etc. But, can anything be saidon a scientific basis? Some recent

research from Switzerland has shedsome light on the subject. More light isneeded and some is coming.

The topic is difficult in itself and mademore difficult by the enormous financialinterests involved. Ideally, in a cold, sci-entific experiment, one should find agroup of humans, some of whom wouldbe subjected to the electromagnetic radi-ation produced by cell phones and theirbase stations for several years and someof whom in the group would not beexposed. All of them would be blind toexposure and they would not differ interms of other factors influencing their

health. One could carefully measuretheir health status and their well being,distinguishing between those exposedand those not exposed. Obviously thishas not been done nor will it be done inthe future.

While much material has been publishedwith the objective of shedding light onthe problem, many of the publishedstudies have serious methodological lim-itations - leading to doubts concerning

their validity. Reference [1] summarizesstudies on mobile phone radiation andhealth-related quality of life, concluding,"The results are contradictory and thegreater part of these studies is not able toaddress the issue of causality betweenexposure and outcome."

One of the first studies dealing withradio frequency electromagnetic fieldexposure was about the shutdown of theSwiss Schwarzenburg short-wave radiotransmitter. While this took place back in1998, a recently published paper [2]reports of measurements in the change inthe melatonin produced by 54 volunteers

living in the vicinity. The measurementsshowed that in the week after the shut-down, sleep quality improved and mela-tonin excretion increased by 15 percent,on average. (Melatonin plays an impor-tant role in regulating sleep-wake cyclesbecause its production is inhibited bylight and permitted by darkness.Melatonin is sometimes prescribed whensuffering from jetlag.) However, the shutdown date was known to the study par-ticipants. Thus, the authors state that"blinding of exposure was not possiblein this observational study and this mayhave affected the outcome measure-ments in a direct or indirect (psycholog-ical) way."

In September, 2003, the Dutch TNOPhysics and Electronic Laboratoryannounced the results of a study. Theywrote: "The present study contributes tothe research on finding a relationbetween electromagnetic fields and brainfunctions." Further, "Exposure to GSM-900 or GSM-1800 electromagnetic fieldshad no effect on well-being in eitherexperimental group. However, exposureto a UMTS-like signal resulted in asmall, but statistically significant impair-ment of well-being. Interestingly thiseffect was not only observed in individ-uals with self-reported health complaintsattributed to daily life EMF exposuresbut also in a reference group withoutsuch complaints. This effect was foundafter only about half an hour's exposureto what, by everyday standards, was arelatively high environmental fieldstrength (1 V/m). In practice, while indi-viduals in the vicinity of operationalUMTS antennas will be subject to con-tinuous exposure, the field strengths inquestion will be lower." Thus, the trans-ferability of the 'TNO'-results to normalday-to-day environmental exposure isquestionable. [3]

There was considerable criticism of thecircumstances of this study includingthat by its authors themselves. Theseauthors and others recognized the needto make a more careful repeat of the

Electromagnetic Radiation

Cell Phone Dangerby Martin Röösli and Harry Rudin

As we become more and more keen on cell phones there remains the openquestion of what effect these cell phones have on our health and well being. Somerecent work in Switzerland sheds some light on these issues.

What effect havecell phones onour health andwell being?

5599


study. However, one result of the TNOstudy was a temporary and partial mora-torium on the construction of new basestations in Switzerland, pending a morethorough study.

A Swiss scientific consortium wasformed to carry out such a more thor-ough study. It was performed by a groupof scientists from the University ofZurich, the University of Bern, and fromthe Swiss Federal Institute ofTechnology in Zurich. Several method-ological improvements were imple-mented, including investigating a largersample size and applying two differentUMTS exposure levels, all yieldingmore robust results. The results werepublished on June 6, 2006. [4]

The effect of UMTS radiation found inthe TNO study was not confirmed in thenew Swiss study. It should be noted thatthe experiment had to do only with 45-

minute UMTS-radiation exposure inter-vals (four of them separated by a periodof one week) and radiation absorption inbrain tissue was considerably smallerthan during the use of a mobile phone.Thus, the authors concluded, "No con-clusions can be drawn regarding short-term effects of cell phone exposure orthe effects of long-term, base-station-like exposure on human health."[4]Nevertheless, several of the temporaryand partial moratoria for the construc-tion of new UMTS base stations inSwitzerland were withdrawn, eventhough the question of long-term effectsof UMTS radiation on humans remainsopen.

Meanwhile, the REFLEX study underProfessor F. Adlkofer, was making mea-surements on isolated cells to see if thesewere effected by a high or low-frequency(power-line frequency) EMF (electro-magnetic field). Quoting from the report:

"The main goal of the REFLEX projectwas to investigate the effects of EMF onsingle cells in vitro at the molecular levelbelow the energy density reflected by thepresent safety levels."[5] The studyshowed that in certain human cell typesthere was indeed a significant increase inthe number of single- and double-strandbreaks in the DNA as a result of high [6]and extremely low-frequency electro-magnetic fields [5].

These results fascinated Professor PrimoSchaer at the Center for Biomedicine atthe University of Basel. In a talk at ameeting on June 29 of this year, orga-nized by the Swiss Research Foundationfor Mobile Communication, ProfessorSchaer gave a preliminary report on hisown experiments which confirm thework in Vienna, showing that intermit-tent extremely low frequency fields canresult in damage to DNA. At this point intime these research results are unpub-lished; but, when published, we willreport on them here in ERCIM News.Professor Schaer emphasized that someDNA damages are repaired by the DNArepair mechanism. Thus, the observedgenotoxic effects do not necessarilymean that EMF is carcinogenic for thehuman.

In conclusion, we now know that EMFdoes have some effects on humans andhuman cells. Since we still do not knowwhat the effect is on our health and well-being, some degree of caution wouldseem to be called for.

LLiinnkkss::

TNO follow up study: http://www.mobile-

research.ethz.ch/projekte.htm#18

EMF projects at the University of Bern:

http://www.ispm.ch/index.php?id=814

Primos research group at the University of

Basel:

http://pages.unibas.ch/dbmw/biochemie/


Martin Röösli, University of Bern,

Switzerland


Harry Rudin, Consultant, Switzerland


[1] Seitz H., Stinner D., Eikmann T., Herr C., Röösli M.; ElectromagneticHypersensitivity (EHS) and Subjective Health Complaints Associated withElectromagnetic Fields of Mobile Phone Communication: A literature review [ofpapers] published between 2000 and 2004. Science of the Total Environment, 2005; 349 (1-3): 45-55.

[2] Altpeter, E.-S, Martin Röösli, M., Battaglia, M., Pfluger, D., Minder, C. and Abelin, T.;Effect of Short-Wave (6-22 MHz) Magnetic Fields on Sleep Quality and MelatoninCycle in Humans: The Schwarzenburg Shut-Down Study, Bioelectromagnetics, Vo. 27, pp 142-150, 2006.

[3] Zwamborn, A.P.M., Vossen, S.H.J.A., van Leersum, B.J.A.M., Ouwens, M.A.,Makel, W.N.; Effects of Global Communication System Radio-Frequency Fields onWell Being and Cognitive Functions of Human Subjects with and without SubjectiveComplaints. Netherlands Organization for Applied Scientific Research (TNO). FEL-03-C148, 2003.

[4] Regel S.J., Negovetic S., Röösli M., Berdiñas V., Schuderer J., Huss A., Lott U.,Kuster N., Achermann P. ; UMTS Base Station Like Exposure, Well Being andCognitive Performance. Environmental Health Perspectives 2006, 114 (8): 1270-1275(http://www.ehponline.org/members/2006/8934/8934.html)

[5] Winker, R, Ivancsits, S., Pilger, A., Adlkofer, F. and Rudiger, H. W.Chromosomal Damage in Human Diploid Fibroblasts by Intermittent Exposure toExtremely Low-Frequency Electromagnetic Fields, Mutation Research, Vol. 585,Issues 1-2, pp. 43-49, 1 August 2005, http://www.verum-foundation.de/cgi-bin/content.cgi?id=euprojekte01

[6] Diem E., Schwarz C., Adlkofer F., Jahn O., Rüdiger H.; Non-Thermal DNABreakage by Mobile Phone Radiation (1800 MHz) in Human Fibroblasts and inTransformed GFSH-R17 Rat Granulosa Cells in Bitro. Mutation Research, Vol. 583,(2), pp. 178-183, 2005.

RReeffeerreenncceess


6600


Detection and monitoring of potentialrisks has become an important part of theso-called 'Epidemic Intelligence' as areaction of surveillance systems to therisk of epidemics arriving from less-developed areas outside Europe. Healthearly warning systems in areas remote,inaccessible, or prone to natural or man-made disasters enabled by satellite com-munications can significantly limit therisk of onset and the effects of epidemicsand contribute to settling major publichealth issues. Economically speaking,satellite communications can also savemoney by enabling rapid and coordi-nated response and optimal adjustmentsof resources when deploying an emer-gency plan.

The SAFE project, co-funded by theEuropean Space Agency (ESA), willdevelop and demonstrate the addedvalue of satellite communication ser-vices, including low and high bandwidthaccess to Internet, co-operative workingand geolocalisation for all phases of bio-logical crisis including prevention, early

warning, and crisis management. SAFEaims to establish a roadmap for ESA todetermine how satellite services, byenabling or restoring access to informa-tion, can be integrated in Europeanhealthcare systems and civil protectionauthorities.

The SAFE project is part of the ESAtelemedicine via satellite transitionphase, set up to pave the way for the cre-ation of a European user-driventelemedicine via satellite programme.ESA together with the World HealthOrganization (WHO) are joining forcesfor reaching this goal.

ESA through SAFE and with the assis-tance of WHO will specify, develop, anddemonstrate a European outbreak earlywarning system that will become a keytool for the national and European bodiesin charge of epidemiological surveil-lance and especially the EuropeanCenter for Disease Control (ECDC).SAFE aims towards a better assessmentof the epidemiological risks based on

real-time objective data, which is astrategic objective not only for publichealth policy, but also for the eHealthindustry. The SAFE solution that will bemarketed to the different actors of epi-demiological surveillance comprises thefollowing components: • a network of expert centers and hospi-

tals able to communicate by satellite,and interoperable with other networks(eg gateways to Galileo and GMES-RESPOND);

• a coordination van equipped with aDVB-RCS capability and Internetaccess for communication with deci-sion makers and expert centers. It hasthe role of the local coordinationcentre. Communication between thelocal coordinator and the mobile teamson the field will be ensured via a localWi-Fi network and Satellite phonesoutside the Wi-Fi coverage. Mobileteams will be equipped with handheldterminals for data collection;

• a field laboratory or lab kit equippedwith biological and biomedical equip-ment will identify microbial agentsresponsible for the threats to enable in-situ analysis;

• a SAFE information system includinga data collection sub-system and aGIS. Besides geolocalisation, the GISwill enable visualizing the evolution ofthe epidemic and monitoring the alarmlevels associated with the differentgeographical regions. This system willbe based on relevant standards andwill be interoperable with existinginformation systems, ie national healthinformation system, deployable emer-gency response information system,epidemiological information systemsand others. It will facilitate monitoringof indicators and other measurespotentially useful in modeling thetransmission patterns. Disseminationof this data can be done with theInternet capability via satellite.

An end-to-end case of early warning atthe onset of an epidemic will be demon-

Satellite Communication Services

SAFE: Satellites for Epidemiology and Health Early Warningby Catherine Chronaki and Laurent Braak

The risk of epidemics and emerging or re-emerging diseases such as avian fluor chikungunya fever is rising. Luckily, these risks can be contained withprevention, early warning, and prompt management. In this context, current earlywarning and response systems, although well-developed in Europe, could benefitfrom satellite services.

08/08/2006 SAFE 1

SAFE Concept (1)

Operation field

Local application server

Sat comm. Providerterrestrial gateway

Network equipment

InternetInternet

Data consulting on remote siteVideo conference users

Field agents

Data handling (consulting,meeting, reporting)

End users / Expert centers

SAFEremote

applicationserverField

laboratory

SAFE Coordination van

Globalstar Gateway

GlobalstarEutelsat

GPS

DVB-RCS

Components of the SAFE demonstrator.


6611


strated at the Heraklion prefecture, in theisland of Crete, Greece. The scenariobased on epidemiological monitoringafter an earthquake, namely the onset oftyphoid fever epidemic, will validate theSAFE approach within the user commu-nity and if successful it could becomepart of regular earthquake readinessexercises. Furthermore, other demon-strations will be envisaged and imple-mented, namely related to avian flu andbiological and radiological threats.These itinerant and on-demand demon-strations will be implemented all overEurope upon user request, promoting theadded value of SAFE and allowingpotential end-users and stakeholders toassess the demonstration according torelevant criteria, such as availability,interoperability, and security.

The SAFE consortium of nine partnersfrom four countries is coordinated byMEDES (Institute for Space Medicineand Psychology, France). End users arerepresented by two epidemiology insti-

tutes - ISCIII (CNE) Instituto de SaludCarlos III 'Centro National deEpidemiologia' (Spain) and InVS'Institut de Veille Sanitaire' (France) - toensure compliance with policies andpractices of epidemiological monitoring.ICS-FORTH (Greece) coordinates thedemonstration of post-disaster biologicalrisk monitoring in Crete. MICRO-UNIVr (University of Verona, Italy) isresponsible for the mobile lab equipmentand definitive identification of themicrobial agent causing the biologicalcrisis. MEDIVAN (Italy) providesindustrial expertise on marketing mobileunits for medical emergencies andrescue. GMV (Spain) ensures the overallsystem design, integration and the devel-opment of the GIS. TTSA (France)brings expertise in satellite communica-tions and marketing of satellite servicesin health. REMIFOR (France) providesthe coordination van and links with theprofessional bodies of civil protectionand firemen. MEDES (France) coordi-nates the economical analysis and pro-

vides recommendations relying on itsongoing activities in epidemiology andhumanitarian medicine involving satel-lite communications.

LLiinnkkss::

European Space Agency: http://www.esa.int

European Center for Disease Control:

http://www.ecdc.europa.eu/

World Health Organization:

http://www.who.int


Audrey Berthier,

MEDES, France (project coordinator):


Pierluigi Mancini, ESA, Directorate of

European Union and Industrial programs

(Program manager)

E-mail [email protected]

Catherine Chronaki, ICS-FORTH, Greece


Finland has been responsible forbuilding the crucial 70 GHz instrumentsfor Planck. MilliLab, VTT has beenleading the instrument development.Together with Elektrobit MicrowaveLtd. and Metsähovi Radio Observatory,

MilliLab delivered the six Planck lowfrequency instrument (LFI) 70 GHzfront end modules (FEMs) and six backend modules (BEMs) for the mission(see figure) in 2006.

The scientific performance of Planck isultimately limited by the receiver noiselevel. To ensure very low noise perfor-mance, state-of-the-art IndiumPhosphide (InP) semiconductorsdesigned by the Finnish team were usedin the 70 GHz FEMs and BEMs. Thesemiconductor devices were tested inMilliLab's unique cryogenic test systemat the temperature of 20 K, which is theoperating temperature of the FEMs in thePlanck satellite. The FEMs and BEMswere manufactured and integrated inElektrobit Microwave, and further testedat their cryogenic facility at the satelliteoperating conditions. The receivers builtby the Finnish Planck team have demon-strated world record sensitivity at 70GHz.


Jussi Varis, MilliLab-Millimetre Wave

Laboratory of Finland

Tel: +358405694089


Satellite Devices

Planck Mission 70 GHz Receivers by Jussi Varis

Planck is a European Space Agency - ESA mission to map the Cosmic MicrowaveBackground (CMB) to an unprecedented sensitivity covering the whole sky. Anaccurate map of the CMB will allow the cosmologists to determine definitivelythe properties of our universe, like the age, the geometry, and the matter andenergy composition. The Planck mission will be launched in 2008.

Planck LFI 70 GHz Front-EndModule (FEM, in front)) andBack End Module (BEM, in back)developed by the Finnish Planckteam. The FEM componenthouses four InP semiconductorlow noise amplifiers and twophase shifters, and the BEMadditional four low noiseamplifiers.


6622


In other science areas (eg SynchrotronRadiation (SR), Neutron beams) thereare growing requirements to move tosimilar detector technology. Althoughon a smaller scale this is resulting in highsample rates and channel counts with adramatic increase in the data rates overprevious systems from around 100MByte/s to 10 GByte/s.

Recent advances in FPGA’s have seenthese devices evolve into programmableSystem On Chip (SOC) ASICs makingthese requirements feasible at reasonablecost and size. Typical FPGA compo-nents now contain multiple microproces-sors, multiple multi-gigabit transceiverscapable of driving common networkstandards, as well as multi million gates

of programmable logic with DSPenhancements allowing performancerates up to 250 GMACS/s for the largestdevices.

This provides the opportunity forshrinking our systems and producingembedded high performance DAQ withstandard networked readout such as GBitEthernet – Large Scale Science DAQsystems shrink wrapped into DAQSensor appliances.

In CCLRC Technology we have demon-strated the technical feasibility ofembedded hardware based on high-speed serial optical links, standard net-work protocols, and programmableSystem On Chip (SOC) FPGA devicesto deliver across a wide range of require-ments and projects.

Our designs have implementedembedded hardware support featuressuch as electronic fuses, supply moni-toring, thermal monitoring, shutdown onfault, Wake On LAN type recovery fromshutdown and low stand-by power.Remote reconfiguration of FPGAFLASH memory is possible togetherwith encryption of the bit stream andautomatic reversion to a fallbackFLASH memory in the event of failureduring re-programming. A typical smallsystem might consist of a pixel sensorchip readout by an FPGA which is inturn readout by a PC server over a GBitEthernet connection. A commoditySDRAM memory module connected tothe FPGA provides data buffering andoptical transceivers allow transmissionof the GBit Ethernet over longer dis-tances.

Figure 3 shows an embedded DAQmodule planned for use on SR applica-tions reading out 16 CMOS Pixel imagesensors capable of a multi KHz framerate. The module can log up to 800MByte/s into SDRAM for a few secondsand then is readout to a PC server on twoGBit Ethernet cables each sustaining 80MByte/s.

Detector Technology

Data Acquisition in Scientific Applicationsby Rob Halsall

Large scale scientific instrumentation systems for particle physics research atCERN are currently being deployed with many millions of sensor channels andData Acquisition (DAQ) rates of the order of 100s of GByte/s. These systems,consisting of many racks of large format electronics modules, are typicallyimplemented using custom digital logic in Field Programmable Gate Array (FPGA)devices. Installed in under ground caverns, closely coupled to massive detectorsand networked to PC processing farms, these systems could be regarded as verylarge examples of embedded systems.

Figure 1:One of 500 digitiser cards for a LargeScale Science Project.

Figure 2: High performance Embedded DAQ forCustom CMOS Sensor (not shown).

Figure 3: Multiple Custom CMOS Pixel Sensor with high performance Embedded DAQ.


6633


One problem for these more widelydeployed smaller projects based on highperformance embedded modules is thedesign cost and turnaround time.Unfortunately there is no all-encom-passing design tool for DAQ systems.The design process relies on an increas-ingly large and diverse portfolio of com-plex tools that require a high level ofexpertise to operate. Thus there is agrowing cost of keeping up with the toolflows required to access more powerfulhardware.

Currently we have an R&D programfunded internally by our Centre forInstrumentation (CFI) which will inves-tigate the embedded processor designtools. This project aims to reduce thecost of the firmware and software designfor DAQ projects by developing a uni-fied approach based around FPGAvendor embedded development tools.

The strategy we propose is to exploit thehardware IP modules and facilitiesexisting within the embedded designtools by adding instrumentation-specificDAQ IP for integration of the sensor datastream. Customising the vendor’s frame-work minimises the cost by reducing theamount of work that needs to be done,and minimises the risk by using standardinterfaces.

A similar approach is taken for the soft-ware libraries - by using standard net-work protocols (Ethernet) and operatingsystem APIs the software side of theDAQ library is more like ‘DAQ middle-ware’.

The addition of embedded networkedmicroprocessors within our designsgives us the ability to add higher levelsof intelligence to our systems whichmight otherwise be difficult to include.

Even allowing for the availability of C tohardware tools and an abundant gatecount in the FPGA using embeddedmicroprocessors would still be an effi-cient method.

Examples of the type of features whichcould be added include real-time tuningof the data flow and processing algo-rithms, logging of environmental infor-mation, safety critical monitoring of sen-sors (eg to prevent radiation damage),interaction with higher level control sys-tems using standard protocols and provi-sion of information to the e-science sys-tems to aid in the management of theintegrity of the data set.


Rob Halsall, CCLRC Technology

Rutherford Appleton Laboratory UK

Tel: +44 1235 445140


Since the number of reported securityincidents caused by network attacks isdramatically increasing every year, cor-responding attack detection systemshave become a necessity in every com-pany's network security system.

Generally, there are two approaches totackle the detection problem. Most com-mercial intrusion detection systemsemploy some kind of signature-matchingalgorithms to detect malicious activities.In intrusion detection terminology this iscalled “misuse detection”. Generally,such systems have very low false alarmrates and work very well in the event thatthe corresponding attack signatures arepresent. However, the last point leadsdirectly to the potential drawback ofmisuse detection: missing signaturesinevitably lead to undetected attacks!

Here, the second approach, termedanomaly detection, is in evidence.Anomaly detection maps normalbehaviour to a baseline profile and triesto detect deviations. Thus, no a prioriknowledge about attacks is needed anylonger and the problem of missing attacksignatures does not exist.

Our work deals with the applicability ofmachine-learning approaches, specifi-cally Self-Organising Maps and neuralnetworks in the field of network anomalydetection. By presenting a set of normal(problem-specific) feature vectors to aSelf-Organising Map, it can learn thespecific characteristics of these vectorsand provide a distance measure of howwell a new input vector fits into the classof normal data. In contrast to othermachine-learning approaches, Self-

Organizing Maps do not require ateacher who determines the favouredoutput.

Considering the aforementioned featurevectors, the vector elements stronglydepend on the preferred layer of detec-tion. That means, if anomaly detection iscarried out at the connection level forinstance, connection specific features,for example packet counts, connectionduration, etc., have to be collected.However, if we require detection at ahigher level, for example at applicationlayer, another set of features will be nec-essary. Generally we can state, that themore discriminative power a feature pos-sesses, the better detection results weget.

In case we use a neural network architec-ture for anomaly detection, such as aFeed-Forward Multi-Layer Perceptron(MLP), the training procedure fornormal behaviour might not be apparentat first sight, since such networks gener-

Security

Network Anomaly Detection by Means of Machine Learningby Roland Kwitt

Anomaly detection should be an integral part of every computer security system,since it is the only way to tackle the problem of identifying novel and modifiedattacks. Our work focuses on machine-learning approaches for anomaly detectionand tries to deal with the problems that come along with it.


Podcast was elected Word of the Year2005 by the editors of the New OxfordAmerican Dictionary and is now offi-cially defined as “a digital recording of aradio broadcast or similar programme,made available on the internet for down-loading to a personal audio player”. In itscurrent form, podcasting is a way forindividuals and radio stations to broad-cast shows by making them accessible asdigital audio (and now also video) filesover the internet. Listeners decide thetype of programme they wish to hear and

automatically have it downloaded sothey can listen later, off-line, on portablemedia players such as the Apple iPod,which gave its name to the phenomenon.

Both the process of selecting shows tosubscribe to for automatic synchronisa-tion and that of listening at a later timeare solitary experiences. While the var-ious communication media offered bythe internet can help in discovering andselecting shows, there are few ways fora co-located group to interactively share

content and experience from a given setof already-loaded media players. This isa problem, especially as the venue ofmedia players has been criticised as aninvention that increases personal isola-tion in traditionally public or semi-public spaces such as streets, transportsystems or even homes.

The SICS Pondcaster project addressesthese problems and proposes a systemthat radically steps away from currentpodcast selection and consumption bysupporting shared experiences. ThePondcaster system comprises a tabletopinteractive artefact, which uses a pondenvironment as its metaphor, and anynumber of ambient listening stationsplaced in the environment to allow grouplistening, thus, bringing back the social

Web 2.0 and Novel Interaction

From Solitary Podcasting to Ambient Pondcasting by Emmanuel Frécon and Pär Hansson

Podcasting has taken the world by storm and become a mass phenomenon inless than a year. The core activities of podcast consumers, namely selectingbroadcast sources and listening to portable media players, are solitaryexperiences. The Pondcaster project seeks to take podcasting a step further byencouraging group discovery and shared listening through a combination ofinteractive tabletop displays and personalised ambient listening stations.

ally require instances of both normal andabnormal data in the training phase.However, anomaly detection assumesthat we only have instances of benign,normal data. This problem can be suc-cessfully solved by training the neuralnetwork with a set of random vectors, inthe first step, pretending that allinstances are anomalous. In the secondstep, the neural network is trained withthe set of normal instances. In the idealcase, the net is then able to recogniserecurrent benign behaviour as normaland classify potential attacks asabnormal. It is clear that we have to

make the assumption that maliciousbehaviour is potentially abnormal.

One big problem that comes along withanomaly detectors is the lack of normaldata in the training phase of a system.Most anomaly detection systems willmiss an attack if the training data con-tains instances of that attack. However,no system administrator will waste histime, cleansing network traces fromembedded attacks in order to produceadequate training data. What we needare detectors that are tolerant withregards to contaminated training datainstances and still produce accurate

results. Again, machine-learningwith Self-Organising Maps canremedy that problem to a certainextent. Unless the amount ofcontamination is too high (whichdepends on the SOM parametersand the amount of normal data),no map regions will evolvewhich accept anomalous data.

Future work on this topic willinclude research on the featureconstruction process, which wefound to be one of the most crit-

ical parts of an anomaly detector.Furthermore, we plan to evaluatemachine-learning approaches that incor-porate time as an important factor. Lastbut not least, we have to find a solutionto the open problem of “concept drift”,which denotes the fact, that normalbehaviour changes over time, leading tohigh false positives (normal behaviourclassified as anomalous) in most detec-tion systems.


Roland Kwitt, Salzburg Research, Austria

Tel: +43 664 1266900




Simplified anomaly detection process.

6655


Figure 1: The Pondcaster interactive artefactand some mobile devices.

Figure 2: Upon interaction, creatures representing shows surface and the artefactstarts playing their content.

aspects of radio listening from the earlydays of radio. The tabletop artefactbuilds on devices that can be found inevery listener's pocket, such as mediaplayers and possibly mobile phones, andfeatures unencumbered interaction via atouch screen.

The tabletop artefact acts as a sink for thepodcast content of the users' mediaplayers, providing a snapshot of theircurrent choices and podcast tastes. Onphysical connection, the content of anymedia player is automatically shown,using an animated aquatic metaphor. Thepond shows groups of similar content asanimated shoals of creatures, where eachcreature represents a podcast show. Toavoid clutter, the metaphor is self-cleaning, whereby old, unused shoalsslowly sink to the bottom of the virtualpond. By interacting with the creaturesrepresenting podcast shows, users canaccess show details and start listening,which favours collaboration and dis-covery of new sources. The artefact isalso linked to a phone number and anyuser can send an SMS containing thewords of a query. On reception, acting asan internet podcast search engine, theartefact will display matching podcastshows as a new shoal. These shows aretypically not present on any connected

media player, which again favours dis-covery of new shows. Finally, the arte-fact keeps track of bluetooth-enabledmobile phones in its close vicinity. Userscan express their intention to listen fur-ther to newly discovered shows by drag-ging specific creatures to a representa-tion of their phone.

The associations between shows andmobile phones will be remembered bythe system. At a later time, users are ableto listen to the chosen shows at publicradio-like terminals. Users within thebluetooth vicinity of such a terminal willbe recognised and their selectionsretrieved and played using a queuesystem. The setting encourages social-ising, discussion and mobility since itkeeps track of listening state across a setof listening stations for a given user.

The Pondcaster is based on existinginternet technologies and because thesystem uses bluetooth tagging, no data isever moved to mobile phones and mediaplayers. Subscriptions and preferencesare kept on the Pondcaster and madeavailable for further access by mediaplayers and Web browsers. As such, theartefact forms the premises of a person-alised podcast aggregator.

LLiinnkk::

http://www.sics.se/ice/projects/pondcaster


Emmanuel Frécon, SICS, Sweden

Tel: +46 8 6331500


Figure 3: Two users listening to podcastshows at one listening station. The TVscreen shows the waiting list (some usersoff-picture) and sound comes from thedirected speaker hanging from the roof.


6666

EEVVEENNTTSS


EPoSS brings together European privateand public stakeholders in order to coor-dinate and to bundle efforts and to set-upsustainable structures for improving the

competitiveness of European R&D onSmart Systems Technologies and theirintegration. Smart system integration isan emerging area with applications inmany industrial sectors. A particular fea-ture of the technology is its interdisci-plinarity and the holistic viewpointrequired for successful implementation.Systems combining sensing, processingand actuating functions are increasinglycomplex, involving a variety of disci-

plines and principles from physics,chemistry, engineering and biology. Thesystems may also integrate mechanical,electrical and biological functions in

order to create new features like uninter-rupted access to information, securityand ease of use.

New business opportunities - for examplefor automotive industries, a strong pillarof European economy (road safety, envi-ronment) - are also in the focus of EPoSS.Europe has a good position in micro/nanosystems, advanced technologies andembedded systems. Further progress,

however, requires a change in the leveland nature of these activities: regroupingthe scattered industry and research com-munities and developing new instru-ments for interdisciplinary R&D activi-ties as well as new models and formats ofpolicy support for a fast transfer ofresearch into innovative products.

Areas covered by EPoSS include:• Automotive • Aerospace• Telecommunications• MedTech Systems• Logistics/RFID• Commonalities (Cross Cutting Issues).

Rosalie Zobel, EC Director forComponents and Systems welcomed theexpected co-operation with the ETPARTEMIS (embedded systems) andENIAC (European NanoelectronicsInitiative Advisory Council).

Important industrial players are sup-porting and actively participating inEPoSS, such as EADS, Airbus, Thales,VW, Fiat, Siemens, Bosch, Infineon, toname just a few.

ERCIM members participating inEPoSS are Fraunhofer Gesellschaft andVTT. ARC Seibersdorf research (amember of the Austrian ERCIM memberAARIT) will also join EPoSS.

The author, as the chairman of theERCIM Working Group on DependableEmbedded Systems and member of theDECOS coordination team, wasattending this event. Information onERCIM and DECOS was distributedamong the about 200 participants.

LLiinnkk::

http://www.smart-systems-integration.org/


Erwin Schoitsch

ARC Seibersdorf research, Austria


EPoSS – A new European Technology Platform on Smart Systems Integrationby Erwin Schoitsch

The European Technology Platform on Smart Systems Integration (EPoSS)was launched in Brussels on 7th of July 2006. Information Society andMedia Commissioner Viviane Reding officially opened the event andoutlined in an encouraging manner the major issues to be tackled byEPoSS.

Information Society and Media Commissioner Viviane Reding.

Pho

to b

y co

urte

sy o

f EP

oSS

.


EVENTS

The theme of this year's conference wasTowards the European Digital Library.The aim was to emphasise the contribu-tion made by the European DigitalLibrary research community (and itsliaisons with the international researchcommunity in general) to the currentefforts of the European Commission inthis direction. Speakers were invited toelaborate on this topic and the confer-ence opened with a commemorative talkby Dr. Yannis Ioannidis looking back onthe history of the ECDL. A CD-ROM isnow available to the DL research com-munity containing the proceedings of allthe ECDL Conferences.

Although strongly international, the con-ference retained a distinctive Europeanflavour. The 459 participants came from45 countries: 66 percent were fromEurope (26 countries), 23 percent fromAmerica, nine percent from Australasiaand two percent from Africa.

The event followed the usual format ofECDL conferences consisting of a mainconference, a doctoral consortium work-shop, five tutorials, posters and demon-strations (this year there were 33), sixworkshops and the Cross-LanguageEvaluation Forum (CLEF), a majorevent on its own, with 140 participants.

The three guest speakers were HorstFoster, Director of Content in theInformation Society and Media DG ofthe European Commission, who pre-sented the Digital Libraries Initiative,one of the flagships of the i2010 strategicframework; Ricardo Baeza, director ofYahoo! Research Barcelona (Spain) andYahoo! Research Latin America inSantiago (Chile), who focused on thepotential of exploiting users' behaviourin search processes and Michael Keller,librarian at Ida M. Green University,

Stanford, USA, who presented GoogleBook Search, a book-indexing project,and its benefits to readers and publishers,as well as a catalyst for other initiativessuch as The European Digital Library.

A panel discussing the topic SustainedDigital Libraries for Universal Use waschaired by Ching-chih Chen of Simmons

College, Boston, USA and JoséBorbinha of INESC-ID, Portugal. Theywere joined by Abdelaziz Abid fromUNESCO, Vittore Casarosa fromDELOS and Eric van der Meulen fromthe European Library.

Some 36 papers were selected for themain conference, an acceptance rate of28 percent. Papers were organised in 12sessions around the topics: Architecture(I, II), Preservation, Retrieval,Applications, Methodology, Metadata,Evaluation, User Studies, Modelling,

Audiovisual content and LanguagesTechnologies. The event was broadcastlive on the Internet.

The doctoral consortium was scheduledthe day before the conference at the sametime as the tutorials in order to allowboth seniors and students to fully partici-pate in the conference itself. Twoawards were made. The Best PaperAward, supported by the IEEE Technicalcommittee on Digital Libraries (IEEE-TCDL) went to Carl Lagoze, DeanKrafft, Tim Cornwell, Dean Eckstrom,

Susan Jesuroga and Chris Wilper for thepaper Representing ContextualisedInformation in the NSDL, and the BestYoung Researcher Award, supported byDELOS Network of Excellence onDigital Libraries, went to Daniel CoelhoGomes, Universidade de Lisboa, ascoauthor of the paper Design andSelection Criteria for a National WebArchive.

As well as the scientific programme,ECDL 2006 included a full range ofsocial activities allowing participants to

ECDL 2006 - 10th European Conference on Digital Libraries by M.Felisa Verdejo

The tenth European Conference on Digital Libraries was held at the University ofAlicante, Spain from September 17 to 22, 2006. The event was jointly organisedby the University of Alicante, Miguel de Cervantes Virtual Library and UniversidadNacional de Educación a Distancia (UNED).

Ching-chih Chen of Simmons College, Boston, USA.

6688

EEVVEENNTTSS

The main objectives of the Cross-Language Evaluation Forum (CLEF) areto stimulate the development of mono-and multilingual information retrievalsystems for European languages and tocontribute to the building of a researchcommunity in the multidisciplinary areaof multilingual information access(MLIA). These objectives are realisedthrough the organisation of annual eval-uation campaigns and workshops. Thescope of CLEF has gradually expandedover the years. While in the early years,the main interest was in textual docu-ment retrieval, the focus is now diversi-fied to include different kinds of textretrieval across languages and on dif-ferent kinds of media (ie not just plaintext but collections containing imagesand speech as well). In addition, atten-tion is given to issues that regard systemusability and user satisfaction with tasksto measure the effectiveness of interac-tive systems.

Evaluation TracksIn CLEF 2006 eight tracks were offeredto evaluate the performance of systemsfor:• mono-, bi- and multilingual document

retrieval on news collections (Ad-hoc) • mono- and cross-language structured

scientific data (Domain-specific) • interactive cross-language retrieval

(iCLEF) • multiple language question answering

(QA@CLEF) • cross-language retrieval on image col-

lections (ImageCLEF) • cross-language speech retrieval (CL-

SR) • multilingual web retrieval

(WebCLEF) • cross-language geographic retrieval

(GeoCLEF).

Test SuitesMost of the tracks adopt a corpus-basedautomatic scoring method for the assess-ment of system performance. The testcollections consist of sets of statements

representing information needs (queries)and collections of documents (corpora).System performance is evaluated byjudging the documents retrieved inresponse to a topic with respect to theirrelevance (relevance assessments) andcomputing recall and precision mea-sures.

The following document collectionswere used in CLEF 2006:• CLEF multilingual comparable corpus

of more than 2 million news docu-ments in 12 European languages

• CLEF domain-specific corpora:English/German and Russian socialscience databases

• Malach collection of spontaneousspeech in English and Czech, derivedfrom the Shoah archives

• EuroGOV, ca 3.5 M webpagescrawled from European governmentalsites.

The ImageCLEF track used collectionsfor both general photographic and med-ical image retrieval: • IAPR TC-12 photo database; LTU

photo collection for image annotation; • ImageCLEFmed radiological

database; IRMA collection for auto-matic image annotation.

Cross-Language Evaluation Forum - CLEF 2006by Carol Peters

The results of the seventh campaign of the Cross-Language Evaluation Forumwere presented at a two-and-a-half day workshop held in Alicante, Spain, 20-22September, immediately following the tenth European Conference on DigitalLibraries. The workshop was attended by over 130 researchers and systemdevelopers from academia and industry.

get together, and enjoy the local culturalheritage. On Sunday evening there was aboat trip around Alicante bay and the fol-lowing evening there was a reception atthe Santa Barbara fortress. The fortresslies on top of mount Benacantil and hassplendid views over the harbour, the cityand the coastline. Built by the Muslimsin the ninth century, it was redesignedaround 1580. On Tuesday evening therewas a visit to the archaeologicalmuseum. This museum, which wasEuropean Museum of the Year in 2004,tells the history of the Mediterraneanusing installations, audiovisual projec-tions and computer graphics. The confer-ence closed with a dinner and jazz band.Delegates received a USB memory stick

with photos of the conference events andfurther information about participants,sponsors and cooperating institutions.

This tenth conference highlightedresearch achievements in the field andprovided a significant opportunity todemonstrate how the ECDL communityis contributing towards the EuropeanDigital Library.

We are looking forward to the nextECDL Conference from September 16 to21, 2007, in Budapest. For more infor-mation see http://www.ecdl2007.org/

The proceedings of ECDL 2006: JulioGonzalo, Costantino Thanos, M.Felisa

Verdejo , Rafael Carrasco (Eds). LectureNotes in Computer Science, Vol 4172,569 p.

The organisers of ECDL 2006 appreciatethe financial support given to theConference by Universidad de Alicante,Ministerio de Educación y Ciencia,Grupo Santander, GeneralitatValenciana, Patronato de turismo deAlicante and Universidad Nacional deEducación a Distancia (UNED).

LLiinnkk::

http://www.ecdl2006.org


69

EVENTS

ParticipationParticipation was up again this year with90 groups submitting results for one ormore of the different tracks: 60 fromEurope, 14 from North America, 10from Asia, 4 from South America and 2from Australia.

WorkshopThe campaign culminated in the work-shop held in Alicante, 20-22 September.The workshop was divided between ple-nary track overviews, parallel, posterand breakout sessions. In her opening

talk, Carol Peters, the CLEFCoordinator, stressed the need for moretechnical transfer activities. She com-mented that although many advanceshad been made in the multilingual infor-mation access research field there werestill few real-world operational cross-language systems. In her opinion, CLEFshould be paying more attention to issuesthat directly regard the user and theneeds of the application communitiesrather than focusing most attention onsystem performance only in terms of pre-cision and recall. In fact, one of the mostinteresting activities this year was thereal-time question answering exercise,organised on-site by Fernando Llopisand Elisa Noguera, U. Alicante (seefigure). Here the aim was to examine theability of question answering systems torespond within a time constraint. Theneed for more technical transfer wastaken up again in the final session in twotalks. Martin Braschler, U. Applied

Sciences Winterthur, Switzerland, gavean insightful talk on “What MLIAApplications can learn from EvaluationCampaigns” while Fredric Gey fromU.C Berkeley, USA, summarised someof the main conclusions of the MLIAworkshop at SIGIR 2006 in Seattle,where much of the discussion was con-centrated on problems involved inbuilding and marketing commercialMLIA systems. There was also aninvited talk by Noriko Kando, NationalInstitute of Informatics, Tokyo, Japan,on new evaluation activities at the

NTCIR evaluation initiative for Asianlanguages.

The presentations given at the CLEFWorkshops and detailed reports on theexperiments of CLEF 2006 and previousyears can be found on the CLEF websiteat http://www.clef-campaign.org/. Thepreliminary agenda for CLEF 2007 willbe available from mid-November. CLEFis an activity of the DELOS Network ofExcellence for Digital Libraries.

LLiinnkk::

http://www.clef-campaign.org


Carol Peters, ISTI-CNR, Italy, CLEF

Coordinator

Tel: +39 050 3152897


The real-time exercise: demonstrating the interface to the participants.

Embedded systems are regarded as themost important computer applicationfield in the years to come, and a businessdriver for European industry. Thereforethey are a targeted research area forEuropean Research Programmes inFramework 7, promoted by ARTEMIS,the 'Technology Platform for AdvancedResearch and Technology for EmbeddedIntelligence'.

The workshop, entiteled 'DependabilityIssues of Networked Embedded Systems:Research, Industrial Experience andEducation', covered topics such as: • Hardware/Software co-design,

System-on-Chip integration• Software processes for efficient develop-

ment of dependable embedded systems• Architectures, Methods and Tools for

Design and Development of criticalsystems

• Interconnected, networked embeddedsystems

• Time triggered systems• Functional Safety (system assessment,

evaluation and standards)• Verification, validation and certifica-

tion of hardware and software, compo-nents and systems.

• Use/integration of COTS hardware and/orsoftware as well as of legacy systems.

Attendance was free thanks to the spon-sorship of the DECOS project.

LLiinnkk::

ERCIM/DECOS Workshop:

http://www.itk.ntnu.no/misc/ercim/


Erwin Schoitsch, Austrian Research

Centers/AARIT, Austria


ERCIM-DECOSWorkshopby Erwin Schoitsch

The ERCIM Working Group 'DependableEmbedded Systems' held a workshopin cooperation with the EuropeanIntegrated Project DECOS - DependableEmbedded Components and Systemson 29 August 2006 at Euromicro 2006in Cavtat, Dubrovnik, Croatia.


7700

EEVVEENNTTSS

ARTEMIS (Advanced Research andTechnology for EMbedded Intelligenceand Systems) is a strong, industry-drivenEuropean Technology Platform (ETP)whose goal is to establish a coherent,integrated European Strategic ResearchAgenda (SRA) to maintain the Europeanlead in the field of embedded systems.From the scientific-technical point ofview, ARTEMIS tries to establish acommon approach and common tech-nology across application domains. Itdoes this by creating a public-privatepartnership bringing together peoplefrom industry, small and medium-sizedbusinesses, universities, research centresand European public authorities, in thefield of Embedded Systems. Commontechnology is defined by three workinggroups: • Reference Designs and Architectures:

to offer standard architecturalapproaches for major application con-texts and to address complexity chal-lenges, dependability issues and buildsynergies between application sectors(eg aeroplane safety at automotivecosts).

• Seamless Connectivity andMiddleware: to offer wide-scale inter-operability and connectivity, to pro-vide new services, functionality and tobuild the ambient intelligent (smart)environment

• System Design Methods and Tools: toenable dependable, rapid design anddevelopment.

Additionally, a strategic group works ongeneric enabling technologies based onfoundational sciences' results.

Application contexts are wide-spread,according to the principle thatEmbedded Systems are Everywhere.

The target application areas are: • Industrial (automotive, aerospace, rail-

ways, machinery, process industry,

biomedical, manufacturing, medicalsystems, etc.)

• Nomadic Environments (enablingdevices such as PDAs, on-body sys-tems) for the mobile environment

• Private Spaces (homes, cars, offices,entertainment, comfort, well-being,safety)

• Public Infrastructures (airports, high-ways, railways, communication net-works, security systems, energy, etc.).

ARC Seibersdorf research was repre-sented at the exhibition by a booth dis-playing animated videos on the applica-tion of dependable embedded systems atNASA (time-triggered systems for theMars mission), DARPA GrandChallenge (Embedded Vision Systemsbeing part of one of the finalists in thisrace of autonomous vehicles through thedesert in Nevada) and the DECOS pro-ject. The EU-funded Integrated Project

DECOS - Dependable EmbeddedComponents and Systems - was pre-sented via videos explaining the DECOStechnology in the context of automotiveapplications and by a demonstrator TT-Car, a model car demonstrating theadvantages of time-triggered tech-nology.

ERCIM and AARIT, the Austrianmember of ERCIM, was represented atthe DECOS project booth. The ERCIMWorking Group, Dependable Embedded

Systems (DES-WG), is cooperatingclosely with DECOS. The DES-WG is adissemination partner of DECOS andDES-WG events are sponsored byDECOS.

LLiinnkkss::

ARTEMIS: http://www.artemis-office.org/

DECOS: https://www.decos.at/


Erwin Schoitsch, ARC Seibersdorf research/

AARIT, Austria


by Erwin Schoitsch

The third ARTEMIS annual conference took place in Graz, Austria on 22-24 May.The event was attended by nearly 250 delegates from European companies,representatives of SMEs, governments, academia and research Institutes.

ARTEMIS Annual Conference 2006

ERCIM at the Smart Systems/ TTTech Booth at the ARTEMIS Conference Exhibition,Graz 2006.


7711

EEVVEENNTTSS


As the largest agile conference everorganized in Europe, the five day pro-gram included altogether 18 tutorials,five keynote speeches, 22 research andexperience paper presentations, 20 dif-ferent workshops and activities, threecutting edge panel discussions andposter demonstrations, as well as 20round-the-clock open-space events self-organized by the enthusiastic agile

crowd. All together, 283 industrial andacademic professionals from 25 coun-tries around the world, and from over100 different companies, gatheredtogether to discuss their needs and ideasfor incorporating agile methodologiesinto the production models. In thisrespect, the conference can undoubtedlybe regarded as a huge success in terms ofboth quality and quantity!

The conference was also honoured withthe presence of two legends in softwareindustry; Barry Boehm (Professor ofSoftware Engineering, ComputerScience Department Director at USCCenter for Software Engineering andcreator of, eg, COCOMO, Spiral Model,and the Theory W approach), and KentBeck (Founder and Director of ThreeRivers Institute, the father of ExtremeProgramming methodology and one ofthe founders of the Agile Manifesto).

In all, the success of XP2006 is one indi-cator that agile methodologies and theirvariants are here to stay. The enormousand ongoing interest of both the industryand academia shows no signs of sub-siding. However, it is also evident thateven though the agile ideologies andmethodologies have emerged since themid 1990's, there is still a great need tokeep the debate and evolution movingupwards and onwards. In fact, a majorityof software organizations functioning incomplex human and technical ecosys-tems are still contemplating whether tomake the transition towards agile soft-ware development and how they shouldactually proceed in pursuit of increasedquality and customer satisfaction, anddecreased lead-time and costs.

XP2007 will be organized in June, 2007in Como, Italy. Welcome to join theagile crowd!

LLiinnkkss::

http://virtual.vtt.fi/inf/pdf/symposiums/2006/

S241.pdf (Collection of Tutorials,

Workshops, Activities and Key Note

Speeches of XP2006)

http://www.xp2006.org

http://agile.vtt.fi


Outi Salo, Pekka Abrahamsson

VTT Technical Research Center of Finland,


[email protected]

Fellowships available in GRID ResearchThe CoreGRID Network of Excellence currently offers

Fellowships for postgraduate students in the field of GRID Research

The CoreGRID web site also offers the possibility to post job announcementsrelated to GRID research. Job postings are free of charge for academic

institutions and organisations.

For available positions and job postings, see http://www.coregrid.net/jobs

Contact: Céline Bitoune, ERCIM; [email protected]

CoreGRID is a Network of Excellence administrated by ERCIM

XP2006 - Extreme Programming and Agile Processes in Software Engineeringby Outi Salo and Pekka Abrahamsson

The 7th International Conference on eXtreme Programming and Agile Processesin Software Engineering was organized in Oulu, Finland in June 2006 by VTTTechnical Research Centre of Finland and the University of Oulu.

Barry Boehm (left) and Kent Beck at XP2006.

7722

EEVVEENNTTSS

CCaallll ffoorr PPaarrttiicciippaattiioonn

GRIDS@work: CoreGRIDConference, Grid Plugtests andContest

Sophia-Antipolis, France, 27 November - 1st December 2006

Following the successful second GRIDPlugtests in 2005 with 230 participantsfrom many different countries, ETSI,INRIA and CoreGRID decided to orga-nize this year an enhancedGRIDs@work event, composed of:• a series of conferences and tutorials:

view the programme• a 3rd GRID Plugtests, during which

two competitions will take place: theN-Queens Contest and the FlowShopContest.

This event will take place on a Gridinfrastructure. You may providemachines to this effect.

MMoorree iinnffoorrmmaattiioonn::

http://www.etsi.org/plugtests/Upcoming/

GRID2006/GRID2006.htm

Information & registration: http://www.etsi.org/plugtests/Upcoming/GRID2006/GRID2006.htmJoin our sponsorship programme! http://www.etsi.org/plugtests/Upcoming/GRID2006/SponsorGRID2006.htm

OO

O WITH THE SUPPORT OF:

O EUROPEAN PARTNERS:

O CO-ORGANIZERS

Object WebOpen Source Middleware

FLYER GRID 2006 13/09/06 18:10 Page 1

CCaallll ffoorr PPaarrttiicciippaattiioonn

SOFSEM '07 - Current Trends in Theory andPractice of Computer Science

Harrachov, Czech Republic, 20-26 January 2007

SOFSEM (originally SOFtwareSEMinar) is the annual, internationalconference devoted to the theory andpractice of computer science. Its aim isto foster cooperation among profes-sionals from academia and industryworking in all modern areas of computerscience. The SOFSEM program consistsof series of invited talks, given by promi-nent researchers and professionals,Contributed Talks selected by theProgram Committee from submittedpapers, Working Sessions discussingwork-in-progress, and the StudentResearch Forum.

SOFSEM offers a unique opportunity toquickly obtain a representative and in-depth overview of the areas of computerscience which were selected as the topicsof this year. For an optimal atmosphereof close professional interaction,SOFSEM is organised in tracks withwell-defined topics so as to attract allresearchers and professionals that sharespecific fields of interest. SOFSEM isespecially suited for young computerscientists.

TracksSOFSEM 2007 tracks are:• Foundations of Computer Science• Multi-Agent Systems• Emerging Web Technologies• Dependable Software and Systems.

As it has became a tradition in the lastyears, the proceedings with the invitedand contributed papers will be publishedin the series Lecture Notes in ComputerScience (LNCS, Springer-Verlag) whichwill be distributed at the conference.

SOFSEM 2007 is sponsored by ERCIM.


http://www.cs.cas.cz/sofsem/07/

CCaallll ffoorr PPaappeerrss

CAISE '07 - The 19th InternationalConference on AdvancedInformation Systems Engineering

Trondheim, Norway, 11-15 June 2007

CAiSE'07 invites submissions on thedevelopment, maintenance, procurementand usage of information systems. Thisyear's special theme is 'UbiquitousInformation Systems Engineering', butas usual the conference also welcomespapers on other aspects of IS engi-neering. New information systems spandifferent organizations and geographicallocations, and often need to support theuntethered mobility of the people usingthe systems. They have a large impact onthe every day life of organisations aswell as individuals. In the light of thedevelopment towards ambient, pervasiveand ubiquitous computing this impactwill increase significantly.

TopicsIn addition to the special theme, relevanttopic include:• Methodologies and approaches for IS

engineering:• Innovative platforms and architectures

for IS engineering• Engineering of specific kinds of IS• Quality concerns in IS engineering.

Important Dates• 30 November 2006:

Paper submission deadline • 10 February 2007:

Notification of acceptance • 11-15 June 2007:

Conference and workshops

The proceedings will be published in theseries Lecture Notes in ComputerScience (LNCS, Springer-Verlag).

CAISE 2007 is organised by NTNU andsponsored by ERCIM.


http://caise07.idi.ntnu.no/


7733

EEVVEENNTTSS


ECOOP 2007 - 21st EuropeanConference on Object-OrientedProgramming

Berlin, 30 July- 3 August 2007

The ECOOP 2007 conference inviteshigh quality papers presenting researchresults or experience in all areas relevantto object technology, including work thattakes inspiration from or builds connec-tions to areas not commonly consideredobject-oriented.

TopicsECOOP wishes to embrace a broadrange of topics, so the following list ofsuggested topics is by no means exclu-sive:• Architecture, Design Patterns• Aspects, Components, Modularity,

Separation of Concerns• Collaboration, Workflow• Concurrency, Real-time,

Embeddedness, Mobility, Distribution• Databases, Persistence, Transactions• Domain Specific Languages,

Language Workbenches• Dynamicity, Adaptability, Reflection• Frameworks, Product Lines,

Generative Programming• HCI, User Interfaces• Language Design, Language

Constructs, Static Analysis• Language Implementation, Virtual

Machines, Partial Evaluation• Methodology, Process, Practices,

Metrics• Model Engineering, Design

Languages, Transformations• Requirements Analysis, Business

Modeling• Software Evolution, Versioning• Theoretical Foundations, Formal

methods• Tools, Programming environments.

Research papers should advance the cur-rent state of the art. Many differentresearch methods can be applied, eg,both experimentally based work andmathematical results are welcome.Experience papers should describe novelinsight gained from the practical applica-tion of object technology, in such a waythat it is of interest to a broad group ofresearchers and practitioners.


The Sixth Joint Meeting of theEuropean Software EngineeringConference and the ACMSIGSOFT Symposium on theFoundations of SoftwareEngineering

Dubrovnik, Croatia, 3-7 September 2007

The biennial series of joint ESEC andFSE provides a lively and outstandingforum where researchers and practi-tioners can report and discuss recentresearch results and trends, as well astheir impact on practical application inall areas of software engineering.ESEC/FSE yields the just-right scope,wide enough to portray a comprehensiveand truly-international state of the artoverview, but small enough to keep aninteractive and productive atmosphere.

ESEC/FSE has continuously attractedparticipants from industry and academia,confirming that the venue providestopics that are up-to-date, significant,and interesting. ESEC/FSE 2007 intendsto continue this tradition by offering astrong technical program associated withan exciting list of events, includingkeynote talks by leaders in the field,invited state-of-the-art seminars, work-shops, a doctoral symposium, and the

new Widened Software Engineering(WISE) presentations.

TopicsContributions to the above events on anyfield of Software Engineering areinvited, including, but not limited to:• Requirements Engineering• Specification and Verification• Software Architecture and Design• Analysis and Testing• Development Paradigms and Software

Processes• Evolution and Refactoring• Tools and Environments• Empirical Software Engineering• Software Metrics• Software Quality and Performance • Component-based Software

Engineering• Model Driven Engineering• Web Applications• Distributed Systems and Middleware• Service Oriented Applications• Mobile and Embedded System• Open Standards and Certification• Software Economics and Human

Resources• Dependability (safety, security, relia-

bility)• Case Studies and Experience Reports.

Important Dates• Workshop proposals submission:

5 March 2007• Workshop proposals notification:

15 March 2007• Technical papers submission:

19 March 2007• WISE papers submission:

3 April 2007• Doctoral Symposium abstracts sub-

missions: 14 May 2007• WISE papers notification:

3 May 2007• Technical papers notification:

1 June 2007• Doctoral symposium notification: 10

June 2007• Camera-ready contributions (all): 1

July 2007

The proceedings will be published in theseries Lecture Notes in ComputerScience (LNCS, Springer-Verlag).

ESEC/FSE 2007 is sponsored by ERCIM.


http://www.idt.mdh.se/esec-fse-2007/

WorkshopsECOOP 2007 will host a number ofworkshops addressing different areas ofobject-oriented technology. Workshopsserve as a forum for exchanging latebreaking ideas and theories in an evolu-tionary stage.

Important Dates• Submission of papers:

13 December 2006• Update period: 13-20 December 2006• Deadline for Proposals:

20 December 2006

ECOOP 2007 is sponsored by ERCIM.


ttp://ecoop07.swt.cs.tu-berlin.de/


EURO LEGAL

74

ERCIM News is the magazine of ERCIM. Publishedquarterly, the newsletter reports on joint actions ofthe ERCIM partners, and aims to reflect thecontribution made by ERCIM to the EuropeanCommunity in Information Technology. Through shortarticles and news items, it provides a forum for theexchange of information between the institutes andalso with the wider scientific community. This issuehas a circulation of 10,500 copies. The printed versionof ERCIM News has a production cost of 8 Euro percopy. It is available free of charge for certain groups.

AdvertisingFor current advertising rates and conditions, seehttp://www.ercim.org/publication/ERCIM_News/ or contact [email protected]

Copyright NoticeAll authors, as identified in each article, retaincopyright of their work.

ERCIM News online editionhttp://www.ercim.org/publication/ERCIM_News/ERCIM News is published by ERCIM EEIG, BP 93, F-06902 Sophia-Antipolis CedexTel: +33 4 9238 5010, E-mail: [email protected] 0926-4981

Director: Jérôme Chailloux, ERCIM Manager

Central Editor:Peter Kunz, ERCIM office [email protected]

Local Editors:AARIT: n.a.CCLRC: Martin Prime

[email protected]: Michal Haindl

[email protected]: Annette Kik

[email protected]: Carol Peters

[email protected]: Eleni Orphanoudakis

[email protected] ICT Group:

Michael [email protected]

FNR: Patrik [email protected]

FWO/FNRS: Benoît [email protected]

INRIA: Bernard [email protected]

Irish Universities Association:Ray [email protected]

NTNU: Truls [email protected]

SARIT: Harry [email protected]

SICS: Kersti [email protected]

SpaRCIM: Salvador [email protected]

SZTAKI: Erzsébet Csuhaj-Varjú[email protected]

VTT: Pia-Maria Linden-Linna [email protected]

W3C: Marie-Claire [email protected]

SubscriptionSubscribe to ERCIM News by: • contacting the ERCIM office (see address above) • filling out the form at the ERCIM website at

http://www.ercim.org/

Commission released the current draft of Proposalfor a Council FrameworkDecision on the protection of personal data processed in theframework of police and judicial co-operation in criminal mattersto the Secretary-General of the Council. The current draft saysthat data can only be exchanged with non-EU states "if an ade-quate level of data protection is ensured in the third country or bythe international body to which the data concerned shall be trans-ferred". Five member states support this 'adequacy' test (CzechRepublic Switzerland, Finland, Greece and Portugal) but seven -backing the US complaint - are opposed to its inclusion(Germany, Denmark, Spain, Ireland, Norway, Sweden and UK)24 Aug 2006 the original draft see, http://europa.eu.int/rapid/pressReleasesAction.do?reference=MEMO/05/349&format=HTML&aged=0&language=EN&guiLanguage=en

EU: Visa Information System (VIS) takes shapeThe EU is currently deciding on the purpose, function and scopeof the VIS and law enforcement access to it. The personal data ofeveryone who applies for an EU short-stay visa, including theirphotograph and fingerprints, will be recorded in the VIS (thisincludes persons whose applications for a visa are rejected). Aswith the "US VISIT system", this data will ultimately be used tofacilitate identity checks and verify entry to and exit from the EU.In addition, the "internal security" agencies of the member statesand Europol will have access to the data. See Draft Regulation onVIS and the exchange of data between Member States on short-stay visas, 7 September 2006, See also http://www.libertysecu-rity.org/article108.html

Privacy International released a comprehensive GlobalFreedom of Information Survey. Privacy International provides a comprehensive review ofFreedom of Information Laws and practices in nearly 70 coun-tries around the world. the survey draws attention to the growingmovement around the world to adopt FOI laws. While it alsohighlights that many problems still exist such as poorly draftedlaws, lax implementation and an ongoing culture of secrecy inmany countries. 28 September 2006 the survey is available athttp://www.privacyinternational.org/foi/survey

The European Parliament adopted 'European Parliamentresolution on the interception of bank transfer data from theSWIFT system by the US secret services'. The European Parliament, demanding explanations from EU gov-ernments and institutions regarding their complicity in theSWIFT transfers of financial data to the US on 5 July 2006. See,http://www.europarl.europa.eu/sides/getDoc.do?pubRef=-//EP//TEXT+TA+P6-TA-2006-0317+0+DOC+XML+V0//EN

edited by Yue Liu, NRCCL, University of Oslo, Norway

News about legal information from European directives, andpan-European legal requirements and regulations.


75

IN BRIEF

CWI - Christine Bachoc, Université Bordeaux, and FrankVallentin, CWI, have found new upper bounds for 'kissing'

in higher dimensions.In geometry the kissingnumber is the max-imum number of non-overlapping unit ballsthat can simultaneouslytouch a central unit ball.In two dimensions thekissing number is six.This can be seen wheneuro coins are groupedaround one central eurocoin (see picture). Thekissing number is onlyknown for the dimen-

sions 1, 2, 3, 4, 8 and 24. For the dimensions 5, 6, 7, 9 and 10Bachoc and Vallentin now found sharper upper bounds. Thekissing problem has a rich history. In 1694 Isaac Newton andDavid Gregory had a famous discussion about the kissingnumber in three dimensions. Gregory thought thirteen ballscould fit while Newton believed the limit was twelve. Only in1953, Schütte and Van der Waerden proved Newton right.

An article explaining the results is available athttp://arxiv.org/abs/math.MG/0608426.

ISTI-CNR - Donatella Castelli coordinated the first DigitalLibrary Reference Model workshop in Frascati, near Rome,

in June. At the workshop, aninvited group of internationalexperts discussed the digitallibrary (DL) model proposedby DELOS, the Network ofExcellence on DigitalLibraries. Despite the largenumber of digital library sys-tems in existence, there is so farno clear agreement on whatthese systems are exactly, norwhat basic functionality theyshould provide. This model is afirst attempt at establishing anabstract DL framework, basedon a small number of unifyingconcepts. It seeks to provide acommon semantics that can be

used across different DL implementations. The workshopaimed at achieving global consensus on this model. In herintroductory speech, Patricia Manson, Head of the EC unit for'Technology-Enhanced Learning & Cultural Heritage' said thatshe expected that the results of the workshop would impact onthe definition of FP7 research priorities and help to initiate therelevant political and operational actions needed to deliver theEuropean Digital Library to European citizens.

Kissing coins.

A team of researchers from the Politecnico di Milano andCEFREIL Labs, led by Stefano Ceri, won the 2006 SemanticWeb Services Challenge (SWS), organised by StanfordUniversity. The goal of the SWS Challenge is to develop a

common understanding of var-ious technologies intended tofacilitate the automation ofmediation, choreography anddiscovery for Web Servicesusing semantic annotations.The challenge required thesolution of a Web Servicemediation problem and a WebService discovery problem.The implementation presentedby the Italian team was judgedthe most complete. See:http://deri.stanford.edu/chal-lenge/2006/

Stefano Ceri.

Fraunhofer-Gesellschaft - Ulrich Trottenberg took over thepresidency of the 'Fraunhofer Center of Institutes', SchlossBirlinghoven, on 1st July 2006. The 'center of institutes SchlossBirlinghoven' gathers all research institutes at the Birlinghovencampus. With some 500 scientists, it is the largest center forinformation technology of the Fraunhofer Gesellschaft. Duringhis presidency Trottenbergintends to make the leading roleof the institute center more vis-ible. He is planning to do thisby numerous events/meetingsfor industry, research, politicsand the general public.

Ulrich Trottenberg holds thechair for Applied Mathematicsand Scientific Computing at theDepartment for Mathematics atthe University of Cologne. Healso is the head of theFraunhofer Institute forAlgorithms and ScientificComputing (SCAI) and vicechairman of the FraunhoferResearch Group for Informa-tion and CommunicationTechnology (ICT). In addition he heads the Department ofSimulation and Software Technology of the GermanAerospace Center.

Current research topics of the Fraunhofer Center of Institutesare the extension of the world-wide web to information tech-nology services, the handling of the data flood with automaticextraction of knowledge from data (datamining) and Human-Machine Interaction.

Ulrich Trottenberg.

Donatella Castelli.


ERCIM – The European Research Consortium for Informatics and Mathematics is an organisationdedicated to the advancement of European research and development, in information technologyand applied mathematics. Its national member institutions aim to foster collaborative work withinthe European research community and to increase co-operation with European industry.

ERCIM is the European Host of the World Wide Web Consortium.

FORTH

Fraunhofer ICT GroupFriedrichstr. 6010117 Berlin, GermanyTel: +49 30 726 15 66 0, Fax: ++49 30 726 15 66 19http://www.iuk.fraunhofer.de

Institut National de Recherche en Informatique et en AutomatiqueB.P. 105, F-78153 Le Chesnay, FranceTel: +33 1 3963 5511, Fax: +33 1 3963 5330http://www.inria.fr/

Swedish Institute of Computer ScienceBox 1263, SE-164 29 Kista, SwedenTel: +46 8 633 1500, Fax: +46 8 751 72 30http://www.sics.se/

Technical Research Centre of FinlandPO Box 1000FIN-02044 VTT, FinlandTel:+358 207226041, Fax :+207226027http://www.vtt.fi/

Swiss Association for Research in Information Technologyc/o Prof. Dr Alfred Strohmeier, EPFL-IC-LGL, CH-1015 Lausanne, SwitzerlandTel +41 21 693 4231, Fax +41 21 693 5079http://www.sarit.ch/

Magyar Tudományos AkadémiaSzámítástechnikai és Automatizálási Kutató IntézetP.O. Box 63, H-1518 Budapest, HungaryTel: +36 1 279 6000, Fax: + 36 1 466 7503http://www.sztaki.hu/

Irish Universities Associationc/o School of Computing, Dublin City UniversityGlasnevin, Dublin 9, IrelandTel: +3531 7005636, Fax: +3531 7005442http://ercim.computing.dcu.ie/

Austrian Association for Research in ITc/o Österreichische Computer GesellschaftWollzeile 1-3, A-1010 Wien, AustriaTel: +43 1 512 02 35 0, Fax: +43 1 512 02 35 9http://www.aarit.at/

Norwegian University of Science and Technology Faculty of Information Technology, Mathematics andElectrical Engineering, N 7491 Trondheim, NorwayTel: +47 73 59 80 35, Fax: +47 73 59 36 28http://www.ntnu.no/

Spanish Research Consortium for Informaticsand Mathematics c/o Esperanza Marcos, Rey Juan CarlosUniversity, C/ Tulipan s/n, 28933-Móstoles, Madrid, Spain, Tel: +34 91 664 74 91, Fax: 34 91 664 74 90http://www.sparcim.org

Consiglio Nazionale delle Ricerche, ISTI-CNRArea della Ricerca CNR di Pisa, Via G. Moruzzi 1, 56124 Pisa, ItalyTel: +39 050 315 2878, Fax: +39 050 315 2810http://www.isti.cnr.it/

Centrum voor Wiskunde en InformaticaKruislaan 413, NL-1098 SJ Amsterdam, The NetherlandsTel: +31 20 592 9333, Fax: +31 20 592 4199http://www.cwi.nl/

Council for the Central Laboratory of the ResearchCouncils, Rutherford Appleton LaboratoryChilton, Didcot, Oxfordshire OX11 0QX, United KingdomTel: +44 1235 82 1900, Fax: +44 1235 44 5385http://www.cclrc.ac.uk/

Foundation for Research and Technology – HellasInstitute of Computer ScienceP.O. Box 1385, GR-71110 Heraklion, Crete, GreeceTel: +30 2810 39 16 00, Fax: +30 2810 39 16 01http://www.ics.forth.gr/

Czech Research Consortium for Informatics and MathematicsFI MU, Botanicka 68a, CZ-602 00 Brno, Czech RepublicTel: +420 2 688 4669, Fax: +420 2 688 4903http://www.utia.cas.cz/CRCIM/home.html

Fonds National de la Recherche6, rue Antoine de Saint-Exupéry, B.P. 1777L-1017 Luxembourg-KirchbergTel. +352 26 19 25-1, Fax +352 26 1925 35http:// www.fnr.lu

FWOEgmontstraat 5B-1000 Brussels, BelgiumTel: +32 2 512.9110http://www.fwo.be/

Order FormIf you wish to subscribe to ERCIM News

free of chargeor if you know of a colleague who would like to

receive regular copies of ERCIM News, please fill in this form and we

will add you/them to the mailing list.

Send, fax or email this form to:ERCIM NEWS

2004 route des LuciolesBP 93

F-06902 Sophia Antipolis CedexFax: +33 4 9238 5011


Data from this form will be held on a computer database. By giving your email address, you allow ERCIM to send you email

I wish to subscribe to the

o printed edtion o online edition (email required)

Name:

Organisation/Company:

Address:

Postal Code:

City:

Country

E-mail:

You can also subscribe to ERCIM News and order back copies by filling out the form at the ERCIM website athttp://www.ercim.org/publication/Ercim_News/

FNRSrue d'Egmont 5B-1000 Brussels, BelgiumTel: +32 2 504 92 11http://www.fnrs.be/