TAMS4CPS_Roadmapping-WS-Documentation_05-2015
Transcription
TAMS4CPS_Roadmapping-WS-Documentation_05-2015
Documentation 1st TAMS4CPS Roadmapping Workshop Brussels, 21 May 2015 Compiled by: Sabine Hafner-Zimmermann, Steinbeis-Europa-Zentrum (SEZ) 26/08/2015 The TAMS4CPS project is co-funded by the European Community's Horizon 2020 Programme under grant agreement no 644821. Documentation 1st TAMS4CPS Roadmapping Workshop Table of Contents Participants .................................................................................................................................. 3 1 Introduction................................................................................................................... 4 2 Background.................................................................................................................... 5 3 Workshop agenda and aims ........................................................................................... 6 3.1 Agenda 3.2 Workshop aims ............................................................................................................................................................................. 6 4 Roadmapping process .................................................................................................... 8 4.1 Introduction to M&S for CPS......................................................................................................................................................... 8 4.2 M&S Themes............................................................................................................................................................................... 11 4.3 Participants’ trends and priorities ............................................................................................................................................... 13 4.4 Elaboration of TAMS4CPS roadmap ............................................................................................................................................ 23 5 M&S Roadmap sketches ............................................................................................... 32 6 Next steps .................................................................................................................... 42 ............................................................................................................ 6 © TAMS4CPS Consortium 2 Documentation 1st TAMS4CPS Roadmapping Workshop Participants Abbrev. Name Organisation Prof. Hans Vangheluwe University of Antwerp, Belgium PEM Pär-Erik Martinsson Luleå University of Technology SC Silvia Castellvi Atos Spain EC Eva Coscia Holonix PP Paolo Pedrazzoli University of Applied Sciences and Arts of Southern Switzerland (SUPSI) DM Dimitris Mourtzis University of Patras - Laboratory for Manufacturing Systems & Automation SK Stamatis Karnouskos SAP JH Julius Hermann Óbuda University Les Worth Selex ES SP Sabri Pllana Linnaeus University PJP Paul Palmer Loughborough University MAS Murray Sinclair Loughborough University JSF Prof. John Fitzgerald Newcastle University Zoe Andrews Newcastle University LD Lipika Deka Loughborough University MR Meike Reimann Steinbeis-Europa-Zentrum SEZ Sabine HafnerZimmermann Steinbeis-Europa-Zentrum SEZ Sarah Mortimer Steinbeis-Europa-Zentrum SEZ Claire Ingram Newcastle University CI © TAMS4CPS Consortium 3 Documentation 1st TAMS4CPS Roadmapping Workshop 1 Introduction The roadmapping workshop which took place on 21 May in Brussels was the first one within the framework of the TAMS4CPS project. TAMS4CPS is an EU-funded project with the mission to develop a strategic research agenda for collaboration (SRAC) to foster Trans-Atlantic research in Modelling and Simulation (M&S) for Cyber-Physical Systems (CPS). The project started in February 2015 and will run for two years. The first step in developing this SRAC was to draft a roadmap and identify the most important and pressing issues of M&S for CPS. To this end, nearly 20 researchers and practitioners active and renowned in the field came together to elaborate a first roadmap as well as sketch out concrete solutions for the highly prioritised issues identified. In the following, an outline of the approach and the activities undertaken during this workshop are given. This draft roadmap will be further discussed with the TAMS4CPS expert community and elaborated during several theme-specific workshops conducted during the coming months. At the end of the project, the SRAC will give a comprehensive picture on future research priorities to lay the foundations for future Trans-Atlantic collaboration opportunities. © TAMS4CPS Consortium 4 Documentation 1st TAMS4CPS Roadmapping Workshop 2 Background Technology roadmapping is applied by various organisations to support the process of strategy development and planning with regard to the future development and use of technologies. Roadmaps are frequently developed to identify and visualize strategically relevant aspects and routes of future development/activities. In the following, this information is used to strategically assess existing gaps to achieve or implement these and to detect how to overcome them The main benefit of roadmapping is that it provides information that helps to make better technology investment decisions. It does this by identifying critical technologies or technology gaps that must be filled to meet performance targets, identify future markets etc. Frequently, a matrix-shaped roadmap layout as displayed below is used. It contains two axes, one horizontal, time-based axis and one vertical axis displaying the dimensions relevant for the field in question. This structure allows to depict and integrate a large number of different aspects across a broad time horizon spanning from today until the longer-term future. Roadmapping usually is a workshop-based approach which, for a certain pre-defined theme, aims to in a structured way elaborate ideas on how the future might look like for this theme. Also, the approach aims to improve communication between those with different perspectives but does not aim to provide a definitive picture of the future. To best capture the issues which are most relevant for TAMS4CPS, the layers used for the TAMS4CPS roadmapping were - Trends and drivers: both general, mega-trends and CPS-specific trends, - Needs and requirements for the development of M&S for CPS, - Research Themes along the 5 TAMS4CPS themes on M&S for CPS, - Technologies and technological developments in M&S for CPS, and - Enablers and barriers such as funding, infrastructure, resources etc. In the next chapter, the roadmapping process is described in detail. © TAMS4CPS Consortium 5 Documentation 1st TAMS4CPS Roadmapping Workshop 3 Workshop agenda and aims 3.1 Agenda 1st Roadmapping Workshop 09:15 Registration 09:30 Welcome, Introduction and Overview SEZ, LU, UNEW 10:15 Participants present trends & priorities (5 min per participant) ALL 11:00 Elaborate and Prioritise Trends & Drivers ALL 11:30 Coffee Break 11:45 Elaborate and Prioritise Needs & Requirements of M&S for CPS 12:15 Lunch 13:15 Elaborate, Discuss and Prioritise Research Themes for M&S ALL 14:30 Discuss Enablers and Research Collaboration ALL 15:00 Coffee Break 15:15 Group Work: Develop Mini-Roadmaps for prioritised topics Groups 16:15 Presentation of Mini-Roadmaps Groups 16:30 Wrap-up (including Feedback) SEZ, ALL 17:00 End ALL 3.2 Workshop aims The TAMS4CPS roadmapping process aims to - Lay the basis for the Strategic Research Agenda for Collaboration (SRAC), - Identify technology needs, research and development priorities in M&S for CPS, © TAMS4CPS Consortium 6 Documentation 1st TAMS4CPS Roadmapping Workshop - Ensure clarity of the M&S for CPS definitional framework which was elaborated as TAMS4CPS Deliverable 1.1, - Include the socio-economic perspective, identify industrial and societal needs for the five TAMS4CPS themes. As described above, the roadmapping process was conducted using a dedicated roadmapping methodology. Thus, the process was based on the template below using the dimensions tailored to the TAMS4CPS project needs. The following chapter will give the details of the roadmapping process. Now 2015 Shortterm Medium -term Longterm 2017 2019 2021 Vision beyond 2021 Trends for M&S for CPS Needs & Requirements (M&S for CPS) Trends & Drivers Mega-trends M&S for CPS Themes Architectures principles and models for autonomous safe and secure CPS Systems design, modelling and virtual engineering for CPS Real-time modelling for autonomous adaptive and cooperative CPS MBSE applied to computing platforms and energy management Required Developments in M&S (technologies) Integration of socio/legal/governance models within modelling frameworks Enablers /Barriers Funding and collaboration opportunities Regulatory environment /governance Standards/IPR regime (industry) Business Models Education & Training, Skills, Knowledge resources Infrastructure / Architectures Other © TAMS4CPS Consortium 7 Documentation 1st TAMS4CPS Roadmapping Workshop 4 Roadmapping process 4.1 Introduction to M&S for CPS The workshop was opened by Meike Reimann (SEZ) from the TAMS4CPS project consortium who welcomed the workshop participants. Following, Zoe Andrews (Newcastle University) gave an overview on and a definition of modelling and simulation (M&S) for Cyber-Physical Systems (CPS) as well as a distinction between EU and US understanding, challenges and opportunities. Very briefly, CPS from a European perspective refers to ICT systems (sensing, actuating, computing, communication, etc.) embedded in physical objects, interconnected (including through the Internet) and providing citizens and businesses with a wide range of innovative applications and services.1 For a graphical distinction between CPS and related areas, see also the figure below. From a US perspective, CPS can be described as smart systems that encompass computational (i.e., hardware and software) and physical components, seamlessly integrated and closely interacting to sense the changing state of the real world. These systems involve a high degree of complexity at numerous spatial and temporal scales and highly networked communications integrating computational and physical components.2 Thus, an initial overview might suggest that European usages place more emphasis on the "cyber" aspect of CPS, whereas the US definition pays equal attention to both the "cyber" and "physical" part. Figure: Relationship between CPS and related areas3 Zoe continued giving a definition of modelling and simulation. Following the EU-supported COMPASS-Project, a model can be defined as a partial description of a system, where the description 1 EC, 2013. European Commission. Cyber-Physical Systems: Uplifting Europe’s Innovation Capacity. Report from the th th Workshop on Cyber-Physical Systems: Uplifting Europe’s Innovation Capacity, 29 – 30 October 2013, Brussels. December, 2013. 2 Energetics Inc., 2013. Foundations for Innovation in Cyber-Physical Systems, US Dept. Commerce, Washington DC, US: National Institute of Standards and Technology. 3 Cf D5.2. Technical report: http://www.cyphers.eu/sites/default/files/D5.2.pdf, 2014. © TAMS4CPS Consortium 8 Documentation 1st TAMS4CPS Roadmapping Workshop is limited to those components and properties of the system that are pertinent to the current goal (COMPASS D11.3)4. Modelling in turn can be described as the activity of creating models. Very generally, simulation can be defined as the imitation of the operation of a real-world process or system, e.g. a model, over time.5 The challenges in the field of M&S from an EU perspective according to the CyPhERS project6 include - - - Multi-domain modelling: no established body of knowledge on how to adequately model all the relevant aspects of cyber-physical systems o useful combinations of those aspects and the required level of abstraction – to effectively reason about its physical, technical, or organizational properties o methodical guidelines are missing how to use suitable abstractions of (parts of) a cyber-physical systems at varying level of detail to enable the engineering of those systems with a sufficient level of confidence concerning the quality of the implemented systems Complexity: realization of cyber-physical systems challenges engineering technologies with a new level of complexity o size and heterogeneity of those systems including cross-technology and crossorganization aspects o issues like self-configuration and self-optimization as well as life-update and lifeextension Disruptive business models: established networks may prove inadequate for the new business models and dynamic value-chains implemented by cyber-physical systems and the flexibility of the technical and organizational processes supported by them In contrast, the most relevant challenges from a US perspective7, are - Lack of common terminology, modelling languages, and rigorous semantics for describing interactions – physical and computational – across heterogeneous systems - Current models for human and machine behaviours are not adequate for designing CPS when humans and machines closely interact o - Modelling and measuring situational awareness – human perception of the system and its environment and changes in parameters that are critical to decision-making The difficulty of verifying performance, accuracy, reliability, security, and various other requirements 4 COMPASS D11.3. Convergence Report 3; Document Number: D11.3. Technical report, http://www.compass-research.eu, October 2014. 5 CPS-VO. Cyber-Physical Systems Virtual Organisation. Tagcloud. Available online: http://cps-vo.org/tagadelic, accessed April 2015. 6 Cf CyPhERS D6.1+2. Bernhard Schätz, Martin Törngren, Saddek Bensalem, María Victoria Cengarle, Holger Pfeifer, John McDermid, Roberto Passerone and Alberto Sangiovanni-Vincentelli. Integrated CPS Research Agenda and Recommendations for Action. Document Number: D6.1+2. Technical report: http://www.cyphers.eu/sites/default/files/d6.1+2.pdf, 2015. 7 F Energetics Inc., 2013b. Strategic R&D opportunities for 21st Century Cyber-Physical Systems, Washington DC, US: National Institute of Standards and Technology. © TAMS4CPS Consortium 9 Documentation 1st TAMS4CPS Roadmapping Workshop o e.g., challenges in modelling privacy requirements so that a system can be validated against these requirements - Lack of formalized high fidelity models for large systems, insufficient ways of measuring performance, and inadequate scientific foundations - Modelling the security threat, developing a formal approach to CPS vulnerability assessments, and designing evolutionary and resilient architectures to handle rapidly evolving cyber and physical threats - Call for new governance models – both domestic and global – for providing standards, protocols, and oversight of systems that operate both in physical and cyber space - Fused business models (of the IT industry with those of engineering-based industries) are not yet well-established and can be difficult to convey Opportunities related to the development of M&S for CPS include from the EU perspective8: - Cost-effective development: current approaches to CPS design and verification are already stretching the limits for cost-efficient system development o - Interoperability: making systems inter-operable is not only a technical endeavor, […] it is important that technical solutions are considered in conjunction with for example business drivers and regulations to make sure that "standards" are developed at the right level o - new methodologies (methods and tools) are needed as enablers for cost-efficient development of such systems, paving the way for solid means for verification and certification paving the way for novel CPS applications, and facilitating market expansion Standards: evolve standards that may be preventing or even blocking innovation o support detailed investigations […] to investigate deficiencies and desired evolution of regulations From a US perspective9, the following opportunities are deemed most relevant: - Development of new, formal modelling methods to create robust, physically relevant simulations that accurately recreate scenarios that CPS systems will experience in operation - Creating domain-specific design frameworks that are built on generic but customizable methods and tools would contribute substantially to reducing time to market, development costs, and the complexity of the design process - Models that are adaptive, implementable at varying degrees of sophistication, and optimized for human interventions will help manage risks and safety as systems move toward mixedinitiative modes of operation 8 CyPhERS D5.2. Martin Törngren, Saddek Bensalem, María Victoria Cengarle, John McDermid, Roberto Passerone and Alberto Sangiovanni-Vincentelli. CPS: Significance, Challenges and Opportunities. Document Number: D5.2. Technical report: http://www.cyphers.eu/sites/default/files/D5.2.pdf, 2014. 9 Energetics Inc., 2013b. Strategic R&D opportunities for 21st Century Cyber-Physical Systems, Washington DC, US: National Institute of Standards and Technology. © TAMS4CPS Consortium 10 Documentation 1st TAMS4CPS Roadmapping Workshop - 4.2 Innovative approaches to abstractions and architectures that enable seamless integration of digital and physical systems for control, communication, and computation are needed […] to support and facilitate cost-effective integration M&S Themes TAMS4CPS focuses on the on the four major themes identified in the Artemis JU Strategic Research Agenda (SRA) as well as on a fifth cross-cutting theme. These are briefly described below. Theme 1: Architectures Principles and models for Autonomous Safe and secure Cyber-Physical Systems - Covers all aspects of systems architecting, but particularly focuses on development of modular and composable architectures that take account on non-functional aspects, such as safety and security. - Eventually, such models must include the human element in a disciplined fashion and may be used to support assurance and even certification requirements. - The area of developing and agreeing standards is particularly important for this theme. Theme 2: System Design, modelling and virtual engineering for Cyber-Physical Systems - Concerned with increasingly complex modelling of increasingly complex systems - Autonomous systems interacting with humans will require new developments in M&S that should be extended to reliable verification and validation - A feature of this aspect of modelling will be dynamic models that capture accurately selforganising systems containing embedded software - Virtual engineering as a means to explore more extensive solution spaces will also be a feature of this theme Theme 3: Real time modelling for Autonomous adaptive and cooperative Cyber-Physical Systems - Models that can be used to control dynamic systems (more efficient use of resources and adaptability over the life-cycle to ensure sustainability) - Includes aspects of machine learning and distributed decision-making by CPS - Focus om Human machine interfaces Theme 4: Model-Based Systems Engineering (MBSE) applied to Computing Platforms and energy management - Concerned with energy efficient computing and includes the better management of large distributed networks of devices © TAMS4CPS Consortium 11 Documentation 1st TAMS4CPS Roadmapping Workshop - Emphasis of this theme is on the use of MBSE to describe, and hence manage large networks that dynamically reconfigure - Environmental modelling will also be important in this theme Theme 5: Integration of socio/legal/governance models within modelling framework - Focus on the integration of heterogeneous models that afford an integration of embedded software with models of the real world - Models of technical systems must necessarily make assumptions about the operational environment and the rules of operation. However, to better understand the complexities of massive CPS in the everyday world, models must be developed that include social, legal, and governance aspects of the overall system - This is an area of growing importance both for safe operation and for understanding better how the full power of CPS can be exploited © TAMS4CPS Consortium 12 4.3 Participants’ trends and priorities Following the introduction to the theme and the process of the workshop, the participants briefly presented themselves and their M&S priorities and vision to the audience. The issues brought up during this introductory round are displayed below. This step aimed to introduce the participants with each other as well as their background and fields of interest, and to establish a common basis and understanding by the participants which facilitates the following thematic discussions. Dimitris Mourtzis, Lab for Manufacturing Systems and Automation (LMS) Timeline Now Short Medium Long Vision Developments in M&S for CPS Computer-controlled machine tools and equipment; robots performing repetitive tasks, fenced off from people Hardware in the loop simulation modelling for systems-of-systems Advanced data mining for analysis and trend identification Adaptable and Selflearning models for CPS Multi-disciplinary and multidomain real-time simulation for CPS Consider / Exploit Internet of Things in manufacturing modelling and simulation Cloud-based data storage with increased security Developments towards standardization Development of web-based engines and software Integration with existing simulation tools Rationalization of existing software and conversion to parallel computing / webbased paradigms AutomationML – OPC-UA integration Collaboration & Relevant Funding Schemes in EU and US EU: H2020 2014-2015 Calls CPSoS – CSA H2020 CAPP-4-SMEs – FP7 CyPheRS project – FP7 etc. Smarter, connected processes for agile and efficient production Real-time Situational Awareness Smart Grid / Factory / Worker Symbiotic (robots work safely with people in shared spaces) Knowledge management and reasoning for manufacturing Holistic Virtual Factory Models and Simulation USA: Cyber-Physical Systems (CPS), Cyberinfrastructure Program EuroCPS – H2020 USA: Engineering and Systems Design (ESD) program - NSF USA: Cyberinfrastructure Program N/A Interoperable, smart, and autonomous hardware and software components Novel AI technologies, allowing for the use of extensive computing power Natural language processing together with technical information (CAx) N/A USA: Engineering and Systems Design (ESD) program - NSF EU: H2020 FoF-ICT The TAMS4CPS project is co-funded by the European Community's Horizon 2020 Programme under grant agreement no 644821. Documentation 1st TAMS4CPS Roadmapping Workshop Paolo Pedrazolli, SUPSI Now Timeline CPS M&S Trends & Drivers Short Medium Long Open and Cloudbased Simulation Multi-disciplinary and Multidomain simulation simulation methodologies and multidisciplinary tools for the design, engineering and management of CPS-based Factories, in order to strategically support production-related activities during all the phases of the factory life-cycle, from the integrated design of the product process - production system, through the optimization of the running factory, till the dismissal/reconfiguration phase. Simulation for Lifecycle Management Needs & Requirements (M&S for CPS) Digital Consistency & Security Handling of big data Digital continuity Simulation scalability Synchronization of Digital and Real World Multidisciplinary integrated simulation Vision John Fitzgerald, Newcastle University Timeline Now Short Medium Long Vision CPS M&S Architectural Frameworks & Modelling (SysML, UML) Domain-specific: automotive, transport, smart grid. Transition to low-carbon economy Urban population growth CPS integration (SoCPS?), support for conformance and verified emergence; Trends & Drivers Co-modelling & co-simulation Model integration (FMI etc.) Semantics & formal techniques Marshalling big data/cloud. Threats to infrastructure © TAMS4CPS Consortium Growing verification capability & efficiency (domain-specific) 14 Climate change CPS for the social good (energy, transport, sustainability): Documentation 1st TAMS4CPS Roadmapping Workshop “Smart” resilience and need to address them. SoS & IoT Needs & Requirements (M&S for CPS) Guidelines for model production Architectural Patterns Socialise CPS concept: recognition across domains and countries Demonstrators of wellfounded M&S in CPS. First integrated toolchains demonstrated. Clear business cases for M&S in CPS. Curricula & communities of interest in CPSEng integrating digital and other Eng. Disciplines. Developments in M&S for CPS First semantic integrations under development. Foundations for semantic interoperability to support co-simulation Not CPSs, but CPS thinking; Effective tools to engage stakeholders and domain experts in CPS design Documented cases of robust multi-disciplinary M&S. CPS Eng. graduates in leading roles. Effective multidisciplinary design teams for CPS. Quality CPS engineering education. Efficient large-scale cosimulation based on multiparadigm models. European ecosystem in CPS design technology CPS Eng. graduates in lead technical roles Semantic interoperability supporting verification and model integration. Evidence management tools. Efficient Verification & simulation methods (for design space exploration). Enabling smooth trade-off across C/P boundaries. Semantically well-founded verification of CPS-level properties. CPS you can trust. © TAMS4CPS Consortium 15 Documentation 1st TAMS4CPS Roadmapping Workshop Murray Sinclair, Loughborough University Timeline Now Short Medium CPS M&S Creating hybrid models of CPS Creating hybrid models of CPS Creating hybrid models of CPS Trends & Drivers Developing M&S approaches for highly-variable, probabilistic, interdependent components Developing M&S approaches for highly-variable, probabilistic, interdependent components Developing M&S approaches for highlyvariable, probabilistic, interdependent components Using models for V&V for very large CPS Using models for V&V for very large CPS Using models for V&V for very large CPS Modelling networks of business contracts Modelling networks of business contracts Use of M&S to control CPS, including the updating problem Use of M&S to control CPS, including the updating problem Using M&S to model sustainability and its amelioration Using M&S to model sustainability and its amelioration Understanding what M&S of complex systems is telling us Understanding what M&S of complex systems is telling us Maintaining ‘Quadruple trust’: protection (from loss or corruption), security (from malicious attacks), privacy (private from all but the intended users), safety (from unintentional loss, corruption, or disclosure Maintaining ‘Quadruple trust’: protection (from loss or corruption), security (from malicious attacks), privacy (private from all but the intended users), safety (from unintentional loss, corruption, or disclosure Needs & Requirements (M&S for CPS) © TAMS4CPS Consortium Long Use of M&S to control CPS, including the updating problem M&S for combined CPS Using M&S to model sustainability and its amelioration Understanding what M&S of complex systems is telling us Using M&S to control CPS we no longer understand. Maintaining ‘Quadruple trust’: protection (from loss or corruption), security (from malicious attacks), privacy (private from all but the intended users), safety (from unintentional loss, corruption, or disclosure 16 M&S for systems within CPS that learn Using M&S to control CPS we no longer understand. Vision Documentation 1st TAMS4CPS Roadmapping Workshop Modelling Intrinsic complexity (problem) and induced complexity (organising to deal with it) Modelling Intrinsic complexity (problem) and induced complexity (organising to deal with it) Modelling Intrinsic complexity (problem) and induced complexity (organising to deal with it) Exploring options for transferring complexity to more amenable parts of the modelled system Exploring options for transferring complexity to more amenable parts of the modelled system Exploring options for transferring complexity to more amenable parts of the modelled system Using M&S to analyse business model & architecture to move complexity to where it can best be addressed Using M&S to analyse business model & architecture to move complexity to where it can best be addressed Using M&S to analyse business model & architecture to move complexity to where it can best be addressed Standards for M&S Standards for M&S Modelling ethical behaviour in systems (to meet legal/governance issues) Modelling ethical behaviour in systems (to meet legal/governance issues) Modelling individuals within communities (to meet needs, e.g. health care) Modelling individuals within communities (to meet needs, e.g. health care) M&S for systems within CPS that learn M&S for systems within CPS that learn © TAMS4CPS Consortium 17 Documentation 1st TAMS4CPS Roadmapping Workshop Stamatis Karnouskos, SAP Source: Leitão, Colombo, & Karnouskos (2015), Technical Report, Unpublished/under review. Area Key Challenges Difficulty Priority Maturity in Real-time control of CPS systems high high 4-7 years Real-time CPS SoS high medium 3-5 years Optimization in CPS and their application high medium 4-7 years On-CPS advanced analytics medium high 3-5 years Modularization and servification of CPS low high 3-5 years Energy Efficient CPS medium medium 3-5 years Lifecycle Management of CPS medium medium 5-8 years Management of (very) large scale CPS and CPS-SoS high high 5-8 years Security and Trust Management for heterogeneous CPS high high 5-8 years Safe programming and Validation of CPS SoS high high 5-10+ years Resilient risk-mitigating CPS high high 5-10+ years Methods and tools for CPS lifecycle support high high 3-7 years CPS Capabilities CPS Management CPS Engineering © TAMS4CPS Consortium 18 Documentation 1st TAMS4CPS Roadmapping Workshop CPS Infrastructures New operating systems and programming languages for CPS and CPS SoS medium low 3-6 years Simulation of CPS and of CPS-SoS medium high 3-6 years Interoperable CPS services medium high 2-5 years Migration solutions to emerging CPS infrastructures medium high 3-6 years Integration of heterogeneous/mobile hardware and software technologies in CPS low medium 2-4 years Provision of ubiquitous CPS services medium medium 3-5 years Economic impact of CPS Infrastructure high high 3-6 years Autonomic and self-* CPS high medium 7-10+ years Emergent behaviour of CPS high medium 7-10+ years CPS with Humans in the Loop high high 2-5 years Collaborative CPS medium medium 5-8 years Artificial Intelligence in CPS high high 7-10+ years Cross-Domain large-scale information integration to CPS Infrastructures medium low 6-9 years Transformation of CPS data and information analytics to actionable knowledge high high 4-8 years CPS Ecosystems CPS Information Systems © TAMS4CPS Consortium 19 Documentation 1st TAMS4CPS Roadmapping Workshop Knowledge-driven Decision Making/Management high medium 6-10+ years Paul Palmer, Loughborough University / Webnebulus Ltd. Timeline Now Short Medium Long Vision CPS M&S Increasing complexity and functionality Deployment of IVP6 based systems. Standards for sub system interfaces. Plug and play for complex system components. Reducing cost fabrication, cost 3-D Printing Platform independence Standards for physical components Self configuring interfaces between subsystem components Trends & Drivers Custom complex parts and subsystems. Smart building blocks IPV6 Needs & Requirements ethernet and text de facto standards for linking subsystems: (M&S for CPS) Are they secure? Modeling of of system interaction and validation of security and privacy CPS Domain certified modelling prior to manufacture Are they private? Interaction between subsystems? © TAMS4CPS Consortium 20 M&S of changes to deployed systems of mixed architectures Rapid simulation and modelling before commit to manufacture of systems comprising a mix of new and predeployed components. cf Internet Documentation 1st TAMS4CPS Roadmapping Workshop Gyula Hermann, Óbuda University, Budapest, HUNGARY Timeline Now Short Medium Long Vision CPS M&S Product modelling as a system Requirement engineering Legal impact Trends & Drivers Self adaptive product model Smart modelling Change and increase of information content Impact of technological change Single life cycle model Virtual executable model Intelligent sensor networks Application of mathematical results related to large networks (graph based) New method for information exchange Needs & Requirements Increase of flexibility Inclusion of new aspects (M&S for CPS) Multilevel integration of current levels Windows into other systems Environmental impact Standardised modules and systems Include uncertainty Include business modelling Include exception handling Increase adaptability Learning capabilities Failure mode and effect modelling Developments in M&S for CPS Modelling for preventive maintanence Meta language for system description and simulation Extensive graphic tools for simulation Modelling and simulation of small scale manufaturing systems Simulation of traffic flow and related costs © TAMS4CPS Consortium 21 Simulation with uncertainty included Behaviour model Documentation 1st TAMS4CPS Roadmapping Workshop Hans Vangheluwe, University of Antwerp Timeline Now CPS M&S Growing complexity Trends & Drivers - heterogeneity Short Medium Long Vision - dynamic structure - emergence - adaptability - human(and groups of)-in-the-loop Needs & Requirements Deal with above complexity (M&S for CPS) Supporting M&S tools Integrated vision of CPS systems development (on longer silos). Standardization Integrate analysis (of non-functional properties) with simulation, testing, maintenance, ... Scale-ability Modular composition Developments in M&S for CPS FMI FMI++ (dynamic, multiformalism) interoperability Modelica Integration of domain-specific and general-purpose modelling Semantic adaptation Modelverse Collaboration & Funding Schemes in EU and US COST Action MPM4CPS © TAMS4CPS Consortium 22 Multi-Paradigm Modelling (multi-formalism, multi-abstraction, process) 4.4 Elaboration of TAMS4CPS roadmap The roadmapping process was conducted in five steps. An interactive workshop approach was applied using a large template (see also section 3.2) which was completed layer-wise starting from the top dimension and moving down to the bottom: 1. Elaboration and priorisation of trends and drivers 2. Elaboration and priorisation of needs and requirements of M&S for CPS 3. Elaboration, discussion and priorisation of research themes for M&S (along the five TAMS4CPS themes) 4. Discussion of enablers for M&S as well as research collaboration in the field of M&S 5. Group work to refine prioritized topics in form of a mini-roadmap Elaboration and priorisation were conducted in a similar manner for all dimensions. First, the participants received post-it notes which they populated with the issues they thought were most important for the respective dimension. Then, these were explained and pinned onto the template. Following, the issues were clustered and given an overall labelling. Finally, the participants were given a number of adhesive dots to identify the most important issues from their point of view. In the following, you find the list of the issues clusters which were ranked highest in each dimension. To further process the issues which came up during the workshop, the participants were asked to also write down their acronym (see participants list on page 3), the theme addressed (T1 – T5) and the time horizon envisioned (Now, S, M, L, Vision) onto the post its. These are, where available, included below, too. The time horizons will be incorporated into the analyses to come. The overall results matrix is shown below. The TAMS4CPS project is co-funded by the European Community's Horizon 2020 Programme under grant agreement no 644821. Documentation 1st TAMS4CPS Roadmapping Workshop Trends & Drivers of M&S for CPS No of points Issues cluster and sub-issues 11 Complexity and automation - 8 Modelling of complex manufacturing systems (M, JH) Process optimisation/effectiveness increase (S/M, SK) Automation pyramid evolution, right hierarchical levels, also at manufacturing level (M, PP) - Net based simulation engines and sensors connection for supply chain (DM) - Embedded intelligence in HW equipment for self-configuration in Production Environments (L, DM) Integration and interoperability - 8 Interoperability/ Integration of models from different sectors/ domains e.g. smart homes with emergency systems with traffic management (LD) - Integration of macro-micro simulation mod. (e.g. factory-machine level) - Virtualisation and Cloud Computing (DM) - Automation systems are integrated with corporate IT (cloud solutions) - Convergence of product and system modelling (JH, M/L) - (Co-)Model/ simulation integration (e.g FMI) (now->S, JSF) - (cross layer) enterprise-integration - Multi-domain integration platforms (M, SC) - Standards for interfacing models (for hybrid systems) (S, MAS) - Tool support for integrated models of hardware, software and human systems Resilience and sustainability 5 - Modelling fault tolerance in distributed systems across organisations (CI) - CPS for sustainability & cost reduction in lifecycle (EC) - Simulation for Life-Cycle Management (PP) - Maintenance of systems of models over lifecycles (V&V included) (MAS) Language and platform 4 - Virtual executable model (M,SH) - Modelling long term evolution + placing reliance on interim system versions (CI) - Collaborative cross-disciplinary modelling+simulation teams (CI) - Architectural modelling languages (e.g. Sysml) (Now->S, JSF) - Increasing capability of formal techniques (e.g. model checking) (JSF) - Platform independent modelling language (LD) Data 2 - M&S for high data availability analysis (S, SC) - Big data everywhere and need to handle - Real-time data analytics (SK) - Large quantity of information available, need to make sense out of it Social and legal aspects - M&S societal challenges (L, SC) CPS address aging, lack of workforce (SP) © TAMS4CPS Consortium 24 Documentation 1st TAMS4CPS Roadmapping Workshop - 1 Modelling ethical behaviour by systems interacting with humans/users/consumers (MAS) - Modelling legal/contracts/IPR issues (MAS) Internet of things - 0 Driver IPVG deployed to support IoT, whole new set of issues around configuration and security (S, SP) - sensors, IoT, Actuators are pervasive (S) - Scada is a new target for bachers (PEM) - Smart systems everywhere (S, SP) - Production Systems are becoming more distributed (PEM) Other issues - platforms for integrating different simulator platforms (av ICT, physical, Al) Models to manage system complexity (S, MAS) CPS in global scale, distributed, heterogeneous systems (EC) Driven need to model legacy systems (PJP) M&S new business models (M, SC) Needs & Requirements No of points Issues cluster 8 Learning and self-adaptability - 8 High system adaptability and modularity so CPS in manufacturing are Plug and produce (M, DM) - Self adjustment of digital models based on real-time sensor/ device data (DM) - Self-adaptability, configuration repairing..(EC) - Learning capabilitites (M, JH) - Models for CPS systems that learn (M, MAS) Integration and interoperability 7 - Interoperability and Modularity of Models & simulation solutions (EC) - Self Orchestration-integratabiliy, interoperability (PEM) - Modelling and simulation of CPS with mixed new and legacy sub-systems (PJP) - Adressing heterogeinity (S, SP) - Solutions to support multidisciplinary modelling of CPS (EC) - multidisciplinary integrated simulation Safety and security - 5 Simulation cyber attacks and threats (M) Inclusion of uncertainty, failure mode and effort analysis (M, JH) Testing for brittleness in CPS (M,MAS) Managing resilience and agility - when needed, what to do, mobilising resources (S,MAS) - M&S of reliability (S,SP) - Scalable M&S tools (S, SP) Complexity and automation © TAMS4CPS Consortium 25 Documentation 1st TAMS4CPS Roadmapping Workshop 4 - Automatic generation of M&S models for cloud manufacturing (S, DM) - Manufacturing in the cloud - Modelling systems for real time monitoring control (S, SC) - Simplification (manage complexity) (S/M, SK) - Cost effective solutions (S/M,SK) Standards 3 - Guidelines for model generation (Now, JSF) - CPS modelling to certified standards prior to system deployment (PJP) Digital continuity 2 - Synchronisation of digital and real world (S, PP) - Digital continuity across factory life-cycle (S, PP) CPS engineering curricula (M, JSF) 1 Societal and legal issues 1 - Architectural patterns for CPS (S, JSF) - Enable large scale stakeholder collaboration (M, SK) - Modelling human factors behaviour in CPS (M) Business models 0 - Inclusion of business modelling + enabler (L, SH) - Simple service discovery - "google 4 CPS" (PEM) Others: - Demonstrators for CPS M&S + enabler (S, JSF) © TAMS4CPS Consortium 26 Documentation 1st TAMS4CPS Roadmapping Workshop M&S for CPS research themes Most working time was dedicated to the population of the research themes dimension, as this dimension was divided into the five TAMS4CPS themes described in section 4.2 above. Consequently, a broad variety of issues was taken up and discussed by the participants. It could be noticed though, that most issues were assigned to theme 2 (which had much overlap with theme 1) and theme 3. This was also mirrored in the voting and clustering of issues which is displayed in the table below. Topic A (CPS verification & simulation) is relevant for themes 1 and 2, Topic B (Automatic updates of virtual models when CPS components are moved) is relevant for theme 2, Topics C (Predictive modelling & real-time decision support) and D (Brokerage Platform) are relevant for theme 3. No of points Issues cluster Relevant TAMS4CPS theme 15 CPS verification & simulation (Topic A) 1&2 - 8 Efficient verification & simulation methods for CPS design space exploration (T2M,M/L,JSF) (A) 6 Points - Verification+tools for emergent CPS properties + SoCPS + CPSoS (T2L,M/L,JSF) (A) 1 Point - Ability to verify safety/security/resilience in a CPS (T2L,L,CI) (A) 3 Points - Verification/compliance of CPS properties in large ecosystems (T1,L,SK) (A) 5 Points Automatic updates of virtual models when CPS components are moved 2 (Topic B) - 8 Automatic updates of virtual models when CPS components are removed, changed or replaced (PEM) (L->) 4 Points - adaptable, self learning, simulation models for CPS based on knowledge management and CBR (DM) (Theme 2/S) 1 Point - formal semantic framework for sound co-modelling (S/M,T2S,JSF) 2 Points - architectural frameworks to support self adaptation (T2/3M) (M, Cl) 1 Point Brokerage Platform (Topic D) 3 - 7 Facilitate models and tools for service development on top CPS (T2S,M,SC) (D) - 5 Points - Technology to model systems behaviour when classes of data are missing or are deliberately excluded (T3, MAS) - 3 Points - Technology to model product and system behavior in the context of changing environment (T2S,JH) (D) Predictive modelling & real-time decision support (Topic C – not further 3 elaborated during workshop) - Real time decision support (M2M connection, HMI, distributed and © TAMS4CPS Consortium 27 Documentation 1st TAMS4CPS Roadmapping Workshop - heterogeneous data) (T3N,S,SC) 4 Points Predictive modelling of different prod. cases and how they effect quality, energy, environment, etc. (PEM) 3 Points In addition to the selected topics, the other issues which came up and which were not grouped under specific topics/headings are listed below according to the relevant TAMS4CPS theme. Theme Theme 1 Architecture principles and models for autonomous safe and secure CPS Theme 2 Systems design, modelling and virtual engineering for CPS Theme 3 Real-time modelling for autonomous adaptive and cooperative CPS Issues - Automatic threat model assessment for CPS system of systems (T1L,L,SK) - Standardization of basic/ core CPS services & data models (T1,S/M, SK) 1 Point - Standards for describing functionality of components or subsystems so models may be more easily implemented (T1S, S/M, PJP) 1 Point - New proprietary formats for transfer of exchange of models and data between virtual engineering tools (T1S,S/M,PJP) 3 Points - New proprietary formats for transfer of exchange of models and data between virtual engineering tools (T1S,S/M,PJP) 3 Points - Integration of stochastic models in CPS descriptions (T1-2S,M,JSF) 2 Points - Advanced data analytics for analysis and identification for smart processing agile production (T2S, DM) - Semantic meta-data model for CPS-based factory representation (T2N, PP) 1 Point - Technology to model learning (T2N, MAS) - Application of new mathematical tools (results in the field of large networks) into current/future modelling/simulation (T2N, JH) - Hybrid M&S environments: contin.+discrete, deterministic+random (T2S, S,SP) 1Point - Tools+notations that can capture key safety/security resilience concepts (T2M,M,CL) - Tools that present complex data in ways that humans can understand+make decisions on (T2M,M,CL) 2 Points - CPS behavior models to empower multi-disciplinary engineering (T2M) 3 Points - Ethical models that can generate large simulations on real-time data in usefully short amount of time (T2M, M/L,CL) - Languages for describing CPS reconfiguration (T3M, M/L,JSF) 1 Point - M&S for systems which live - M&S on the fly (T2L,M/L,CI) 1 Point - Integration of stochastic models in CPS descriptions (T1-2S,M,JSF) 2 Points - Engineering of collaboration (automation) systems. M&S of SoS, Cloud-Cloud, complex Systems, Physical Prop. (T3S, PEM) 1 Point - Integrated tools that are flexible and do not impose lifecycle models (T3N, S, Cl) - Technology to assume trustworthy behavior by autonomous systems © TAMS4CPS Consortium 28 Documentation 1st TAMS4CPS Roadmapping Workshop Theme 4 Model-Based Systems Engineering (MBSE) applied to Computing Platforms and energy management Theme 5 Integration of socio/legal/governance models within modelling frameworks - - (T3S, MAS) Theme 3/2016-2017: computer controlled machine tools, equipment and robots fenced-off from people (T3S, DM) Self-X: self-learning and self-adaptive CPS (M,T3S,SC) 1 Point Models for Fog computing vs. cloud computing to support local decision-making (T3M, EC) Languages for describing CPS reconfiguration (T3M, M/L,JSF) 1 Point Modelling industrial human behavior and grocery behavior including cultural effects (T3M, MAS) 1 Point Knowledge integration with other systems continuously, solutions to support DSS along the CPS lifecycle (T3M,L,EC) 1 Point methodologies to map complex high level processes into behavioral models of CPS (T3L,EC) No issues addressed Models to evaluate effects of new regulations (T5S, MAS) Trade-space models for incorporating metrics for sustainability (T5S,MAS), 1 Point Modelling industrial human behavior and grocery behavior including cultural effects (T5M, MAS) Developments in M&S themes were not discussed during the overall roadmapping activity but were rather supposed to be discussed during the smaller working groups and during the coming themespecific project workshops. Enablers & Barriers Also for enabler and barriers, mainly cooperation issues came up which will be further refined during the following theme-specific workshops and will be a major issues in the SRAC later-on. Thus, at this point, these aspects were collected but not further addressed during this first roadmapping workshop. All issues named are compiled in the table below. Funding and collaboration opportunities - collaboration and inclusion of stochastic behavior in system and product modelling - collaboration of realtime modelling of autonomous system (especially applications in transport) - Collaboration on the inclusions of learning capacity - Collaboration on architectures for trust-worthy behavior of CPS, how should CPS behave to customers, users, disabled etc and how do we assure this? (MAS) - Barrier-familiarity with CPS terminology - practitioners don't self-identify as CPS (CI) EU funding: © TAMS4CPS Consortium 29 Documentation 1st TAMS4CPS Roadmapping Workshop - CPS-H2020 (DM) - ECSEL, Industry-region approach - ICT - FoF/Spire/etc. US funding-collaboration: - USA cyber infrastructure progr. - USA-NSF-(ESD) - USA-cyber physical systems EC/NSF Joint programme (S/M,JSF) - Regulatory environment /governance/ Standards/IPR regime (industry) Business Models Education & Training, Skills, Knowledge resources Infrastructure / Architectures Other - cross-program topic definition to solve horizontal needs: Interop, SEC,... collaboration on inclusions of human models in the range of CPS models. Reality check project assessment: o from promise to results 5yrs later o vision check o roadmap adjust (SK) Common methodology for CPS evaluation/ comparison (SK) Popular "Challenge problems" for model-based CPS Engineering ("the driverless car" of CPS Eng.) (S/M,JSF) - development of new business models to support the new manufacturing EPA based on simulation/CPS (DM) Collaboration - cross-sector - multidisciplinary - business & sales - engineering - soft skills such as "design" - create common understanding of similarities and differences among CPS,IoT,M2M => dispersion of effort? - Engineering degree + CSA courses incl. cross-disciplinary CPS modelling training (S,CL) - Workforce need different skills, including the elderly -usability - Cross-disciplinary education (CS+EE+Math+Econ) (SK) - engineering education based on the new paradigm to develop skill/competence -learning-teaching factory (DM) - promotion of best practices in different sectors to promote them. create awareness, attract investors (EC) open platforms to design and simulate CPS at low cost (EC) enabling technology for sharing model and simulation components (S/M, PJP) Popular on widely accepted language for simulation and modelling (S/M,PJP) © TAMS4CPS Consortium 30 Documentation 1st TAMS4CPS Roadmapping Workshop TAMS4CPS main roadmapping template © TAMS4CPS Consortium 31 Documentation 1st TAMS4CPS Roadmapping Workshop 5 M&S Roadmap sketches Following the roadmapping exercise, the three topics A, B, D which were prioritized highest and which were deemed most relevant/interesting by the participants were further elaborated in three working groups in the afternoon by using a similar template as for the general roadmapping activity. Though, this template was to be elaborated in a more focused way taking into account the selected topic only. The last step of the elaboration of the mini roadmap was the presentation of the roadmap to the whole workshop audience. This was done by the so-called solutions template which further detailed and described the topic and the solution resulting from it. Down below, both the topic-specific templates and the solution sheets can be found. Further elaboration of these issues will follow during the course of the project. © TAMS4CPS Consortium 32 Topic A: CPS-verification & simulation by JSF, SP, SK, LD The TAMS4CPS project is co-funded by the European Community's Horizon 2020 Programme under grant agreement no 644821. Documentation 1st TAMS4CPS Roadmapping Workshop © TAMS4CPS Consortium 34 Documentation 1st TAMS4CPS Roadmapping Workshop Solution A: CPS-verification & simulation by JSF, SP, SK, LD Description of Solution Impact of Solution Links to Key Drivers Vision: CPS’s that merit the trust • • • Managing development risk Better market deployment Safer, secure & resilient CPS Data, integration & interoperability, language & platform Resilience & sustainability Key skills, facilities and technology gaps • • • Generating evidence for trust/ confidence Theoretical analysis <static> <static models> Simulation <HPC, co-simulation & modelling of cyber & physical> What research would be needed? • • • Foundations -> analyse heterogeneous aspects through theoretical analysis <composability> Methods -> Model constructions <stating & defining key properties, model design space exploration Tools -> Tools to support the method <hybrid>, tools to manage complexity of models What needs to be done next? • • Robustness of existing tools Extending tools by two ways -> incorporating wider range -> incorporating capability © TAMS4CPS Consortium 35 Documentation 1st TAMS4CPS Roadmapping Workshop Topic B: Automatic updates of virtual models when CPS components are removed by DM, JH, CI PEM, MAS © TAMS4CPS Consortium 36 Documentation 1st TAMS4CPS Roadmapping Workshop © TAMS4CPS Consortium 37 Documentation 1st TAMS4CPS Roadmapping Workshop Solution B: Automatic updates of virtual models when CPS components are removed by DM, JH, CI PEM, MAS Description of Solution Holistic virtual factory – modelling & simulations Impact of Solution • • Efficient customer driven production Seamless integration of CPS-components with systems Links to Key Drivers • • • • Customization & produce to order Demand profiles external & internal Customer integration Traceability of goods (LCA) Key skills, facilities and technology gaps • • • • Natural language processing with technical capabilities Adaptive simulation model Real time situation awareness Automation ML/ arrowhead automation platform/ cloud solution What research would be needed? • • • • • • • Self-configuration of control loops Architectural framework for self-adaptation CBR-Models (Advanced) Contractual Standards data-formats self-awareness What needs to be done next? Interoperable, smart, autonomous hardware © TAMS4CPS Consortium 38 Documentation 1st TAMS4CPS Roadmapping Workshop Topic D: Brokerage Platform by EC, PP, SC © TAMS4CPS Consortium 39 Documentation 1st TAMS4CPS Roadmapping Workshop © TAMS4CPS Consortium 40 Solution D: Brokerage Platform by EC, PP, SC Description of Solution Brokerage platform offering services & solutions to CPS – using industries and IT providers Impact of Solution reduction of costs and increased deployment of M&S solutions for CPS Links to Key Drivers CPS pervasiveness, complexity, integration & interoperability Key skills, facilities and technology gaps • • • technologies for safety & privacy of data, standards & interoperability solutions, Plug & play technologies at system level for SW & HW What research would be needed? • • • IPR management, new business models standards & interoperability solutions education & mindset creation What needs to be done next? • • catapult centers develop infrastructure The TAMS4CPS project is co-funded by the European Community's Horizon 2020 Programme under grant agreement no 644821. Documentation 1st TAMS4CPS Roadmapping Workshop 6 Next steps The TAMS4CPS roadmapping workshop which took place on 21 May 2015 in Brussels was the first workshop in a series of general and theme-specific workshops throughout the project. These activities are supported by several webinars to inform a wider audience on project progress and encourage active involvement in project activities. This is detailed in the project outline below. During the following months 10 theme-specific workshops (2 per theme) will take place. In parallel, the Strategic Research Agenda for Collaboration (SRAC) will be drafted. The results of the initial roadmapping will feed both into the SRAC and into these 10 workshops. Furthermore, the outcomes of the theme-specific workshops will inform the further development of the SRAC during the course of the project. Thus, the workshop documentation at hand is the first building block of the main project outcome, the Strategic Agenda for Transnational Research Collaboration. The TAMS4CPS project partners are grateful for any comments on and additions to the information contained in this documentation by experts from both sides of the Atlantic Ocean, and we look forward to keeping up discussions and interaction on the issue of M&S for CPS. © TAMS4CPS Consortium 42