Platforms4cps Conclusions and Recommendations

Platorum\Recommendations

Platforms4CPS has identified recommendations in 4 key areas:

• Research priorities
• Supporting innovation needs
• Societal and legal issues that need addressing
• Strategic business support

In order to grasp the business opportunities highlighted in the sectors covered by Platforms4CPS there is a need to address these 4 key areas to ensure that:

• The right technology areas are supported
• There is successful transfer of new ideas to European companies via innovation mechanisms
• Societal concerns which are barriers to uptake of new technologies such as trust, privacy, regulation, liability, and security of employment are addressed
• European citizens can rely on trustable systems

In the shorter term, these can begin to be addressed under Horizon 2020 and existing Digitising European Industry activities via engagement with an expansion of the Digital Innovation Hubs, linking PPPs to work in synergy and supporting the development of platforms and large-scale pilots in key domains such as Automotive, Agriculture, Medicine, etc. Further in the future, the recommendations address Horizon Europe linking with developing ideas within the Commission such as the Edge 2030 vision.

Key needs for the future

• Increase digital capacity and capability through Digital Innovation Hubs
• Enhance multi-disciplinarity, cross-fertilisation (application domain & engineering domain)
• Foster collaboration, European coordination and de-fragmentation across Europe
• Support large-scale demonstrators in key areas, e.g. autonomous driving, etc.
• Tackle the issue of the confused landscape of business support for SMEs
• Explore CPS enabled business models and business services, facilitate access of SMEs
• Provide help to SMEs in allaying fears that are significant barriers to adoption, such as risks around cybersecurity
• Encourage the development of common standards to connect different technologies.
• Establish a “Science of Design for CPS”
• Address the skills shortage, particularly in digital engineering capabilities and encourage systematic engagement between education and industry to encourage life-long learning and re-skilling to avoid a future digital divide
• Revitalise EU Engineering education, raise the status of engineering embracing multi-disciplinarity and incorporate CDIO (Conceive Design Implement Operate) ideas to pro-vide T-shape (broad and deep) education considering that around two-thirds of children in primary school today will work in jobs which do not even exist yet
• Ensure that European citizens can rely on European supplied trusted systems

At a societal level, there is a need to work on a legislative framework that addresses privacy, liability and ethics. Although GDPR provides a solution to privacy there is a need for clear guidelines on data ownership, management and exploitation to provide a level playing field across Europe. Finally, social acceptance of increased automation and AI should be promoted in co-operation with trade unions concerning issues such as employment quality and quantity, welfare, health and privacy.

Short-Term - Horizon 2020

In the short term, the European Commission should build upon existing activities and target actions that lead to a synchronisation of National Initiatives, PPPs and large-scale pilots. At a national level member states have already started initiatives like “Industry 4.0” in Germany, “Smart Industry” in the Netherlands, “L’industrie du Futur” in France or the “High Value Manufacturing Catapult” in the UK. Here there are already major investments in the PPPs and large-scale pilots and there is a desire to promote cross-coordination between these to address vision areas such as autonomous cars, health, etc. This is already beginning via the lighthouse projects initiated by the PPPs.

The EC has put in place actions shown under Horizon 2020 to support partnerships and platforms, the regulatory framework, Digital Innovation Hubs to engage with SMEs, and digital skills. The strategy being adopted is fully supported by the findings of Platforms4CPS. It is clear that there is a need for greater cooperation between Member States and a need to synchronise existing and emerging regional, national and EC digitisation efforts. This has a major multiplier effect for Europe.

Overall the Digitising European Industrystrategy mobilise round €50B targeted at digitisation activities by 2020 as shown in Figure 13 with a main pillar addressing strengthening key parts of the digital value chain including PPPs, industrial platforms, large-scale pilots and testbeds. The partnership approach with industry is particularly relevant for the CPS domain which tackles key business sectors, e.g. Automotive, Aerospace, Rail, etc. as well as key societal challenges such as health and well-being of an ageing population. The targeted budget of €3.2 billion being invested in Public Private Partnerships under Horizon 2020 in PPPs (ECSEL, Big Data, Photonics, 5G networks, Cyber-security) and “Platforms and Pilots” (such as the connected smart factory and IoT large-scale pilots) are helping to strategically align activities across Europe.

Considering Cyber-Physical Systems, four elements of the “Digitising European Industry” initiative are of particular importance: i) creation of a pan-European network of Digital Innovation Hubs and ii) synchronisation of national initiatives, iii) support of digital platforms and standards, and finally iv) putting in place a supporting regulatory framework.

Digital Innovation Hubs. There is a strong need to engage with SMEs and support innovation and transfer of technology to SMEs. The most appropriate means for achieving this is via Digital Innovation Hubs, clusters and regional initiatives. The “Digital Innovation Hubs” are seen as an important step forward for mastering digital transformation, where companies, particularly SMEs, can gain experience and insight on novel development of cutting-edge technologies and have the chance to turn these developments into opportunities for their business. The investment of 100M€ a year under Horizon 2020 in this area is welcomed and is essential in providing innovation ecosystems that help fostering the implementation of CPS across Member States. Successful actions and initiatives like the I4MS (Innovation for Manufacturing SMEs - FoF) and SAE (Smart Anything Every-where - ICT) should be built upon with I4MS Phase 4, Smart Anything Everywhere Phase 3 and support of Robotics as a new activity. Already the effects of this are being supplemented via national actions being pursued in 10 member states that also have DIH strategies. Notably between 2016-2017, the EC funded 600 projects in 150 DIHs engaging over 1000 SMEs and Midcaps, however, there is still a need to further broaden this and also improve the coverage of DIH’s with the implementation of 30 new DIH specifically in Eastern Europe (EU13). Overall the European Commission should foster coordination of national and regional initiatives to bring together all relevant constituencies from EU Member States with the aim of creating an EU-wide network of Digital Innovation Hubs.

Supporting Regulatory Framework. The future will be strongly driven by the exploitation of data. Here there are a number of barriers and industry is looking for guidance in a number of areas. The GDPR regulation which was introduced in March 2018 is seen as a very positive step by industry providing much needed guidance. It is also important in allaying societal fears about loss of privacy and how personal data is being used. Further actions by the European Commission in development of the Data Package which covers the free flow of non-personal data, access and reuse of data and access of private data for public purposes is also likely to open up new business opportunities. This will be further supported by the PSI directive also known as European legislation on the re-use of public sector information which is trying to make as much business data, government data and scientific data as possible available to European companies.

Within the CPS domain the areas of AI and autonomous systems are rapidly developing areas and it is expected that these will be key areas in the future. To support these areas there is a need for regulation to address liability issues. Currently, existing regulation in this area is being scrutinised such as that for the liability of defective products and services, as well as safety according to Machinery Directive 2006/42/EC, however, it is clear that as these fields develop there will be a need for more regulation to cover such areas as accidents in autonomous driving and to ensure the transparency of AI. Cybersecurity has also become a major concern for industry as a result of recent high-profile attacks. Certification for cybersecurity is thus a key need and this is being addressed via introduction of the European Cybersecurity Certificates Scheme. Within this generalised overall framework specific cybersecurity mechanisms will be defined for specific products. The area of security is currently a key concern for many companies engaged in development of CPS and is an area that is seen key for the future adoption.

In the longer term, from 2020 onwards, research and innovation activities will be supported by the new Framework Programme Horizon Europe. Ideas for this programme are currently under development and there are many consultations taking place. There is an emphasis on missions which target “moonshot” activities and it is expected that funded projects will contribute to these missions. An example of a mission related to CPS is “an integrated transport system reducing car congestion by 50% in 10 European cities by 2030”. Several other of the proposed missions also address CPS topics.

Emerging Themes - There are a number of emerging themes that have been identified within Plat-forms4CPS workshops. The areas of “autonomous systems”, “Artificial Intelligence” and “trust” were clearly highlighted as key themes for the future. There is a need to support this with funding for low power processing at the edge and a concerted action to master AI at a European level. In order to build trustable systems there is a need to maintain sovereignty of key value chains.

Processing at the Edge - There is a move towards localised intelligence at the “edge” in order to react promptly in time-critical applications. It is not possible to guarantee safety, latency and predictability for autonomous cars if there is a reliance on remote connection to the cloud, so processing needs to be performed locally. Privacy and security concerns also drive the need for processing data at the edge rather than transmitting or storing data in the cloud. Notably edge computing is more amenable for privacy/security and is also GDPR compliant. In order to provide high performance computing and new computing techniques, such as neuromorphic computing, at the edge there is a need for energy efficient computation for battery powered and energy harvesting powered devices as well as for electric vehicles. This will require a 2-3 orders of magnitude improvement in energy consumption.

AI and Autonomous Systems - There is a need for understandable, accountable autonomous systems, which act ethically. Already there is an initiative to support an “AI-on-demand platform” with a desire to connect and strengthen AI activities across Europe. However, with the importance of AI for the future there is a need to provide even greater support for AI developments in key sectors. A positive step is that a declaration of cooperation on AI has been signed by 25 European Countries which will produce a coordinated plan for AI by the end of 2018. AI will have major implications on future systems and European businesses require the necessary tools and skills to adopt and exploit AI-based solutions. In particular, a programme is required to encourage the business adoption of AI technologies to solve problems and deliver practical business value. In addition to adopting/developing AI solutions Europe must develop expertise and provide help in building investment and business cases. Governments and employers need to encourage and provide continual education and training for existing employees throughout their careers to encourage the development of skills in AI. The skills to develop and deploy AI solutions depend upon Science, Technology, Engineering, and Mathematics (STEM) skills and so there is a need provide incentives to pursue these areas particularly at the further education level with support from industry in the development of curricula.

Trust and acceptance - Trust is difficult to build, but in order to generate trust there is a need for a factor of 10 reduction in software bugs, a requirement for better usability, a need for resistance to cyber-attacks, and an approach to explainable AI technologies. Already complexity is a challenge, but tools will be required to improve the productivity of companies to produce dependable soft-ware automation systems, robots and AI. There also needs to be a public debate on ethics and trust with involvement from government, academia and industry as well as the general public.

Sovereignty - There is growing concern over sovereignty across key value chains, e.g. Aerospace and Automotive, as the political climate in the world is changing fast. There are two key threats. The first is that if Europe becomes reliant on foreign hardware and software in key value chains then restrictions on exporting products to other countries may prevent European trade. More seriously if systems rely on foreign made components it will not be possible to guarantee safety and security of future systems. There are increasing concerns about security issues from backdoors in the hardware which may lead to systems and infrastructure being compromised by foreign governments and terrorists. There is thus a pressing need for full European sovereignty in key applications such as defence and security, but also in critical applications such as autonomous cars and infrastructure.

Research Recommendations

A key challenge identified by CPS experts was how to manage complexity at both the design stage and in the operation of systems to provide trustworthy systems for the future. Engineers need to deal with many heterogeneous components and also complex interactions with humans. This requires new approaches to Systems Engineering that can deal with decomposition into components to manage complexity and new or improved CPS engineering approaches.

Gaining a lead in “Edge Computing” was also identified as a major opportunity for Europe. Many systems are relying on cloud computing and Big Data processing at the moment, however, the ability to move processing to local assets allows systems to react promptly in time-critical applications, e.g. autonomous driving. Here the latency of remote connections to central cloud processing would not be acceptable for safety and predictability of response. This coupled with increasing concerns over privacy and security makes edge processing highly advantageous. Successful exploitation of edge computing, however, requires development and demonstration of high performance computing, energy efficient computation for battery and energy harvesting powered operation and new computing techniques such as neuromorphic computing.

Humans are an integral part of the system and may interact with the system in a number of ways raising fundamental questions on the degree of automation required and how CPS and humans collaborate. There is also a move from hand held devices to wearables or even implantable devices. The interaction and collaboration between CPS and humans will thus intensify with intuitive, assisting systems and humanoid robots. In future systems will have to predict and adapt to human needs, preferences and capabilities. Research and development will thus have to cross silos with respect to disciplines such as biology, computing, ethics and engineering in a variety of application domains.

Within CPS diverse functions are integrated to meet systems-level attributes such as safety, security, performance and usability. Above this there is a need for co-engineering to link the system atributes to manage traceability in the product lifecycle and deal with trade-offs between conflicting attributes. Here automation can be used to deal with complexity management, in particular, to track the impact of system changes during the product lifecycle. For instance, to alert design experts that performance or safety may be compromised by a security patch or via addition of new technology, e.g. from an SME. Here there are strong linkages with incremental certification and agile system engineering.

Grand Challenge	Recommendation	Potential Implementation
Trustworthy CPS for Autonomous and Smart AI – Societal Scale CPS	Develop a science of design for CPS with multiple links to application domains	Create a platform for trustworthy CPS €20M with focus on lower TRL, more fundamental multi-domain research. The aim would be to define the research roadmap and implementation strategy for the Science of Design for CPS, which would then be coordinated by a CERN-like organisation
CPS Edge Computing	Support research actions on edge computing algorithms and architectures	Develop a platform for edge computing and promote this via demonstrators
Humans-in-the-Loop	Address the complex interactions between humans and systems with increasing autonomous functionality	Fund multi-displinary research that brings together human factors and CPS engineering
Co-engineering of CPS system attributes	Advance techniques to manage and automate traceability and trade-off optimisation between safety, security, performance and usability	Establish a research field for co-engineering. Benefits include faster certification, system integration and modification

Innovation Recommendations

EU innovation initiatives such as ICT Innovation for Manufacturing SMEs (I4MS) and Smart Anything Everywhere (SAE) provide a good starting point for addressing some of the barriers to innovation highlighted by Platforms4CPS. Digital Innovation Hubs, clusters and regional initiatives need to work together and engage with SMEs to support innovation and transfer of technology. These should be further developed and expanded to connect the many fragmented national and regional initiatives that exist. One such approach would be to establish a CERN-like CPS vehicle to create strong links between competence, demonstration, and innovation centres on an EU scale as well as showcase experiments and large-scale pilots which also have a role to play in alleviating public concerns and regulatory, legal and ethical issues. This vehicle would coordinate the implementation of an over-arching research agenda on design and engineering of CPS.

As well as Digital Innovation Hubs being a means to communicate important information to the right people, they also play a role in studying the technical challenges faced with adopting technology. Recent studies have indicated that productivity could be increased by up to a factor of three through the proper use of existing technology. This is a concern because if existing systems and industrial engineering processes are not suited to the uptake of new technologies, then these technologies will either not be used, or where they are used, will not produce the desired benefit. Thus, it is proposed to study existing industrial engineering processes with migration examples to create national repositories that can be used to provide guidance for organisations to self-analyse and evolve procedures as well as tooling to help in adopting new technologies.

The industry is faced with exploding complexity and new tools will be necessary to deal with complexity and to reduce the expense of developing dependable high-quality software. Agile (open source) platforms as well as the federation of platforms will be needed that can integrate new and legacy systems. This requires the development of toolchains that can support all aspects of the development cycle from design to testing and roll-out of new systems. As systems will evolve over time there will also be a need to continually support systems as new functionality is incorporated.

A key aspect that has been highlighted is the lack of engineers and skills to support future digitalisation in Europe. To counter this there is a need to revitalise EU engineering education. This should not only provide continual education and training for existing employees, but also support the new skills that need to be developed. The role and status of engineering needs to be raised and promoted to society. Incentives are required to encourage students to pursue STEM skills with a goal of providing a multidisciplinary engineering background particularly at the further education level incorporating CDIO (Conceive Design Implement Operate) ideas with support from industry in the development of curricula.

Grand Challenge	Recommendation	Potential Implementation
Defragmentation/Collaboration	Link existing activities to boost communication, avoid fragmentation and silos	Support Digital Innovation Hubs, training, and coordinate via a CERN-like vehicle
Improve the uptake of Technology in CPS Industrial Processes	Build supportive approaches to migrate existing industrial engineering processes allowing swifter time to market for technologies	Joint-venture funding and incentives that support and document the evolution of new technologies
CPS Engineering, Interoperability, Complexity	Foster development of Euro-pean toolchains for CPS	Coordinate projects to develop CPS toolchains via the CERN-like organisation
Skills/Competence Provision EU Competitiveness	Revitalise EU Engineering education and raise the status of engineering, embracing multi-disciplinarity as well as incorporating CDIO (Conceive Design Implement Operate) approaches	Provide incentives for engineering education based on best-established practices such as (CDIO)

Societal and Legal Recommendations

In terms of societal and economic trends IT-systems are becoming ubiquitous, with an increasing societal dependence and an increasing vulnerability. This drives the demand for safety and security of such systems. There is a need for funded projects to encourage technology adoption, but at the same time there is a need to raise public awareness. Key issues are:

• Trust and public acceptance
• Liability
• Cybersecurity
• Safety
• Definition of the ethical basis for AI considering key rules that need to be adopted

Societal acceptance of increased automation and AI needs to be promoted in co-operation with trade unions to allay fears about employment quality and quantity, welfare, health and privacy. There is a need to address ethical issues and support this with clear regulation to ensure that AI systems are acting in the best interests of people. One of the key problems is transparency of what is encoded in the AI. In many systems there is a need to be able to ensure the completeness of the training set to ensure inclusivity (e.g. handicapped people). Already, there is ongoing work within the EC on an ethical and legal framework to support AI with guidelines expected to be published in 2019. There will also be a need to support skills in response to the socio-economic changes that are likely. The lack of actions addressing ethical training for industry was noted and it is recommended that specific efforts are targeted at industry in Horizon Europe programmes.

The future relies heavily on the use and processing of data. Although GDPR provides a solution to privacy there is a need for clear guidelines on data ownership, management and exploitation to provide a level playing field across Europe. Likewise, the need for security has become better recognised and here there is a need for guidelines and supporting legislation. Finally, industry needs clear guidelines on liability for products and services which are exploiting these new technologies.

Grand Challenge	Recommendation	Potential Implementation
Raise Public Awareness	Raise awareness of privacy and security in industry and to the general public	Create a support action for societal dialogue
Ethics of AI	Create an ethical framework for AI, supporting transparency, and regulating for miss-use	Introduce legislation to enforce transparency and ethical adoption of AI
Privacy/ Security, Liability	Enforce GDPR, mandate in-built security mechanisms for key applications and clarify liability law for new products and services	Put in place enforcement measures for GDPR, enforce built-in security for European products and put in place appropriate legislation for products and services

Business Recommendations

There are many opportunities for European SME’s, Mid-caps and Large Industrial Enterprises in the Automotive, Rail, Aerospace, Maritime, Manufacturing, Health and Energy sectors. These arise from increased automation, connectivity, optimisation of systems and processes (e.g. traffic management), health monitoring and also from services (e.g. mobility as a service and infotainment). There are intense competition and concerns due to the significant investments and dominance of companies outside of Europe, e.g. China, South Korea, USA, in the electronics domains. However, if society is to rely on increased automation and highly connected systems there is a need to ensure that European products can be trusted. In order to support this there is a need to maintain sovereignty in key CPS technologies. At a European level there is a need for secure components, cyber-security and data ownership. Devices and software will need to provide a resilient and secure digital space to allow trustable European systems can be built. This is needed for essential technologies not only in sectors such as Aerospace and Defence, but also in autonomous applications such as cars, trains, ships as well as infrastructure applications, e.g. Energy Grid, Banking, etc.

Grand Challenge	Recommendation	Potential Implementation
Maintaining European Sovereignty in key CPS technologies	Develop European value chain for trustable CPS	Significant funding to secure key components of strategic European value chains

Your feedback is welcome!

We invite you to register to the Platforum and give your appreciated feedback on the presented results.
You are also welcome to join the discussion on this topic in the forum here.
For registration or login issue, please contact us for help.