The objective of the meeting is to identify opportunities and challenges in developing methods and tools for model-based engineering systems that are enabled by, and dependent on, networked computing technology. This category of “smart” systems grows ever wider, and ranges from small autonomous devices to large-scale infrastructure1. The availability of data, and the growing power to process it flexibly and at scale, brings great potential, but also challenges the state of the art in design methods and tools. Some of the most fundamental questions relate to the ways in which engineers can be enabled to work across traditional boundaries between disciplines. Our proposed meeting will bring scientists and engineers from a wide range of backgrounds together to appraise the state of the art in these diverse areas, develop shared vision, and prioritise challenges for future work at a range of technology readiness levels.

The rapidity of technological development has made time-to-market key to commercially successful innovation. At the same time, growth in communication capabilities has led to new interdependencies between engineered systems and the other (independently evolving) systems around them. This means that better integration is required between design disciplines and between life cycle phases. New technological possibilities – for example in electronics, virtual reality and in 3D printing – are emerging and may disrupt existing solutions2. There is therefore an urgent need to ensure that researchers from diverse disciplinary backgrounds are enabled to combine different kinds of models into well-founded but heterogeneous collections that describe the key characteristics of these new emerging smart systems of systems3.

There are different dimensions to the increasing level of smartness in systems. One dimension is the relationship of the engineered system in relation to the overall (eco-)system of which it forms a part.

Here the technology roadmap of the ecosystem elements towards their targeted markets are of main importance. A second dimension relates to how the individual engineered system evolves to raise its level of smartness from monitoring, control, optimisation to autonomous behaviour.

Nowadays users expect that systems from different suppliers will interoperate seamlessly in ways that may not have been considered when the individual constituent systems were conceived. In particular, in a business-to-business value chain it is of the utmost importance to have a vision for the evolution of interoperability between all relevant systems in the same ecosystem and this can be seen at different levels from smart components, through smart products, smart connected products, smart product systems and ultimately smart (eco-)systems of systems.

The ecosystem level in particular utilises artefacts that have been produced in one product lifecycle phase in another lifecycle phase. For example, CAD drawings might subsequently be used for 3D printing or to support augmented reality views post-deployment. There could also be monitoring of deployed products that can feed information back to either development or production based on big data analysis.

For each individual product different levels of smartness can be achieved, each enabled by particular technologies:

1. Monitoring: Sensors and external data sources enable monitoring of aspects of the system and its environment. Typically the main enabling technology is the Internet of Things (IoT)4.

2. Control: Given a monitoring capability, embedded software, possibly alongside a cloud solution, confers the ability to personalise and control some system functionality. Typically the main enabling technology is that of Cyber-Physical Systems (CPSs)5.

3. Optimisation: Given monitoring and control capabilities, it becomes possible to optimize performance by predicting forthcoming issues. The enabling technology is typically based on big data analysis6 or machine learning7.

4. Autonomy: In certain domains, a degree of autonomy is sought, in which decisions are made by a system in accordance with its own principles. In practice, a balance is often sought between autonomy and the capacity to collaborate and interact with external systems, for example to create systems that can optimise performance and carry out self-diagnostics in a safe manner. Expertise in fields such as safety, security, dependability and must be involved if the system has an ability to do damage8.

Model-based technologies for systems engineering have been the subject of research and innovation activities worldwide. Such technologies aim to deliver languages, methods and tools for analyzing engineered systems and systems-of-systems from a very early stage of design. Experience is showing that such model-based techniques help to manage development risk downstream, for example reducing the number of prototypes required prior to release. They also enable analytic approaches to the assurance of key system-level properties related to dependability and performance. Recent advances allow diverse engineering disciplines to integrate their models of cyber and physical system elements in simulation environments, and are beginning to deliver tool-supported analysis of such systems that deliver monitoring and control. However, the technologies of model-based engineering networked, smart systems with learning for optimisation and autonomy are far from advanced, in part because of the separation of research and engineering practice in the range of disciplines involved. It is in this context that we propose our Shonan meeting.

We aim to gather scientists and engineers from a wide range of backgrounds to identify the opportunities and challenges in developing methods and tools for model-based engineering of smart systems. Our goal is to work towards common understanding of these opportunities and challenges in spite of the diversity of participants’ backgrounds, and so we will ask participants to participate as representatives of their disciplines as well as individuals.

We aim to balance the composition of the participant group so that we have potential developers of methods and tools – including commercial tools – as well as designers of smart products who may use such methods and tools. We expect to include representation from several possible application areas, such as the following:

• Vehicles: systems that – with varying degrees of autonomy – move and may convey a payload either individually or in swarms. There is considerable interest in this area because of systemlevel performance, safety, security, and regulatory aspects affecting design.
• Infrastructure: systems are often interventions within existing systems that have a long-lasting substantial physical presence, such as the increasing smart-isation of buildings or the electricity grid.
• Mobile devices and human experience: systems that deliver information or experience between humans. This is a volatile market in which innovation is rapid but the expectations of a consistent user experience drive the need for engineering of emergent properties from independent networked devices.
• We expect to hold sessions around three kinds of subject:
• Application areas: These sessions will seek to spread understanding of the characteristics of the smart systems in application areas such as those mentioned above.
• Methods and tools: These sessions will focus on the cross-cutting methods and tools that underpin model-based engineering for smart systems.
• Models and model-based engineering: Our experience in multi-disciplinary projects tells us that fundamental concepts of model-based engineering are often shared between different application domains and engineering disciplines, but are frequently expressed using different terminologies. We therefore consider it vital that we hold “lingua franca” sessions in which we develop an understanding of the concepts that we share.

The meeting will involve two kinds of session: seminars and workshops. In a seminar, speakers will give presentations that form the subject of discussion. We will provide clear guidance that seminar presenters act as the representatives of their area and provide an overview of the state of the art, not concentrating solely on their own work. Workshops take the form of group-based discussions that use specific methods for idea divergence and convergence such as brainstorming and KJ Method in order to develop a type of output9 defined in plenary.

1 When we refer to “engineered systems” here, we include interventions in existing systems, such as the integration of new sensors into a building, for example.

2 Disruptive IoT Innovation, Article No :1274 | July 14, 2014 | by Avi Itzkovitch, UX Magazine,
https://uxmag.com/articles/disruptive-iot-innovation

3 How Smart, Connected Products Are Transforming Competition”, by Michael E. Porter and James E. Heppelmann, Harvard Business Review, November 2014, https://hbr.org/2014/11/how-smart-connectedproducts-are-transforming-competition

4 See, for example https://smartanythingeverywhere.eu/

5 See, for example http://into-cps.au.dk/

6 See, for example https://en.wikipedia.org/wiki/Big_data

7 See, for example https://en.wikipedia.org/wiki/Machine_learning

8 See for example the predictions for automonous cars in ”Autonomous Vehicle Implementation Predictions Implications for Transport Planning”, Todd Litman Victoria Transport Policy Institute, November 2016,
http://www.vtpi.org/avip.pdf

9 See http://www.theworldcafe.com/key-concepts-resources/world-cafe-method/