3 ECS Key Application Areas

This Major challenge focuses on how Industry 4.0 should address the future regulation or market requirements emerging from the European Green Deal and zero-carbon (or below carbon-neutral) operations.

Nearly 200 countries have committed to the Paris Agreement on climate change to limit global warming to below 2ºC. Any update about new challenges and activities are regularly published by the EU Parliament¹⁰. The rapid transformation of all sectors is therefore required. In fact, many European countries have set even more ambitious targets, and Chips JU could have a great bearing in reducing environmental impact through sustainable manufacturing, including energy- and resource-efficiency and by applying circular economy strategies (eco-design, repair, re-use, refurbishment, remanufacture, recycle, waste prevention, waste recycling, etc).

There are some so-called rare earth metals to save, and any kind of careless materials usage can be proven uneconomical and risk to the environment. The vision of Sustainable Process Industry through Resource and Energy Efficiency (SPIRE) categorises the high-level goals discussed above into more practical action, as follows.

• Use energy and resources more efficiently within the existing installed base of industrial processes. Reduce or prevent waste.

• Re-use waste streams and energy within and between different sectors, including recovery, recycling and the re-use of post-consumer waste.

• Replace current feedstock (raw material to supply or fuel a machine or industrial process) by integrating novel and renewable feedstock (such as bio-based) to reduce fossil, feedstock and mineral raw material dependency while reducing the CO2 footprint of processes or increasing the efficiency of primary feedstock. Replace current inefficient processes for more energy consumption reduction.

• Resource-efficient processes when sustainability analysis confirms the benefits.

• Reinvent materials and products to achieve a significantly increased impact on resource and energy efficiency over the value chain.

ManuFUTURE Vision 2030 combines these objectives as shown in Figure 3.3.1

A new upcoming EU co-programmed partnership called Innovative Advanced Materials for Europe (IAM4EU) will be supported by the IAM-I¹² (the Innovative Advanced Material Initiative) association, which already released a template for collecting inputs for its SRIA¹³. This materials value chain initiative is of paramount importance for the EU manufacturing Industry and will have a great impact also in Digitalization and Semiconductor Design & Production.

• Monitoring flows of energy, materials, waste and Lifecycle assessment: It is already commonplace in many industry sectors (food, medicine, etc) that material and energy streams need to be fully traced back to their starting point. As more and more products, raw materials, etc, become critical, this implementation strategy must be expanded. Flows need to be monitored. Sustainable manufacturing needs comprehensive environmental data and other measurements that may have been in place when the relevant manufacturing or production was initiated. On the other hand, this is a very typical application for many types of IoT sensor and systems that can be informed by careful “Life Cycle Assessment” (LCA). LCA is a prerequisite for holistic environmental evaluation, and it is a simple but systematic method, that requires a mixed combination of extensive and comprehensive models and data.

• Virtual AI assistants: Discharges or losses mostly happen when production does not occur as planned, due to mistakes, the bad condition of machinery, unskilled operation, and so on. Human factors cause most of the variation in the running of continuous processes. There should therefore be a focus on how to implement an AI-supported assistant to help operators by providing advice and preventing less than optimal changes and addressing the dynamicity of the production environment.

• Human–machine interfaces and machine-to-machine communications: Augmented reality (or virtual reality) will be used to support a number of tasks. Enhanced visualisation of data and analytic results will be required to support decision-making.

• Human operators in more autonomous plants and in remote operations: the relationship between machines and the human factor needs to be rethought. In terms of the logic of human-machine interface, from touch displays, to wearable devices and augmented reality, but also to maintain the centrality of the human factor within the new contexts. The “Skills 4.0” are necessary for the management of new technologies for data administration, for privacy, for cybersecurity and much more.

• Human safety: With the localisation of personnel, machines and vehicles, situation-aware safety (sensing of safety issues, proximity detection, online human risk evaluation, map generation, etc) will become increasingly vital.

• Competence and quality of work in a human-centred manufacturing: At a strategic level, the European automation and industrial IT industry depends on its ability to attract skilled personnel to maintain their competence over time. A higher level of formal training may be required for workers in production and maintenance. Greater specialisation is constantly introducing products and processes that require greater company-specific training.

• Green Deal: Policy initiatives aimed at putting Europe on track to reach net-zero global warming emissions by 2050 are key to the European Commission’s European Green Deal. Following the highly challenging objectives of the Green Deal, all industries must focus on high efficiency, low energy usage, carbon-neutrality or zero-carbon usage, zero waste from water, soil and air – all measured, calculated or estimated on product, factory, global and lifecycle levels. European industry must research and discover new materials while paying a great deal of attention to recycling, re-use, and de-manufacturing and re-manufacturing. As a “niche” and not exhaustive example new RICS-V based computing hardware will be needed to reduce energy used at data centres. Extra 3 x performance will be gained compared to ARM Cortex-A75.

Many of these advances will require extensive development in the other engineering, business or social domains, even at the individual level, that are outside of the Chips JU focus. However, it is also obviously the case that a growing part of these approaches will be implemented through the significant help of electronics and software technologies. The need for Chips JU technologies is diverse, and it is not useful to indicate one single technology here. High performance, high precision, careful and professional engineering and decision-making are needed – often at a much higher level than today.

Major challenge 3 focuses on connected and smarter cyber-physical systems (CPS), industrial internet, big data, machine learning, optimization, and other AI-based methods. Local edge-based intelligence is seen as an opportunity for Europe. AI optimized and Open hardware becoming more important to support European AI Framework. This Major challenge extends toward AI-enabled, adaptable, resilient factories, including the human as a part of a “socio-technical” system. AI methods, extending various systems used in modern industries, like(predictive) condition monitoring and maintenance, scheduling, planning, or quality control, will be applied to not only support reconfigurable first-time-right/zero-defect manufacturing, but also to support human decision-making (considering uncertainties). Consequently, AI-enhanced systems will enable resilient manufacturing ecosystems based on new business models, increase safety & security in working environments, and improve productiveness and quicker return from investments. In this context, Explainable AI (XAI) is another an emerging field for better understanding of decision making to increase human trust and robustness of emerging technologies. An important challenge here is to lead not only the digital transformation of Industry 4.0/5.0, but also the next generation of Chips JU platforms supporting AI-driven human-centric autonomous Industry 4.0/5.0 operations. Condition monitoring techniques can be applied to many types of industrial components and systems, although often at additional cost. Commonly, the business value required from condition monitoring depends on the higher availability of equipment and, for production processes, information provision to be able to plan and act on maintenance proactively instead of reactively, as well as to offer decreased cost and improved on-time delivery. Other business values that may be of interest are safety and the optimal dimensioning/distribution of spare parts and maintenance staff. Thus, serious breakdowns and unplanned interruptions to production processes can largely be avoided using condition monitoring.

AI will impact several main areas, all of which are relevant to Digital Industry improving:

• Productivity, by exploiting AI in the design, manufacturing, production and deployment processes.

• Flexibility, by using AI throughout the value stream, from supply to delivery, increasing the autonomy and resilience of each process in the value chain.

• Customer experience, by using AI to make products faster and with better quality, and to provide more efficient services.

• Assistance to human operators in circumstances of rising complexity by using AI to support the decision-making process with ever-increasing levels of complexity and dynamics.

• AI radical innovations would enable more productivity and new products that just supporting, maintaining and improving existing ones

These fundamental impacts will be experienced in all areas of every market sector. In manufacturing and production, AI will deliver productivity gains through a more efficient resources, energy and material use, better design and manufacturing processes, and inside products and services, enhancing their operation with more refined contextual knowledge.

The agenda here is cross-sectorial, focusing on AI applied in any domain. However, the impact of AI in Digital Industry is of particular significance. AI in manufacturing is for example in one of the main focus of AI, Data and Robotics¹⁴ (ADR)and area of expected investments, already roadmapped by ADR since 2021¹⁵

1. European AI framework

Figure 3.3.2 sets out the context for the operation of AI public/private partnerships (PPPs), as well as other PPPs or joint undertakings (JUs). It clusters the primary areas of importance for AI research, innovation and deployment into three overarching areas of interest. The European AI framework represents the legal and societal fabric that underpins the impact of AI on stakeholders and users of the products and services that businesses will provide. The AI innovation ecosystem enablers represent essential ingredients for effective innovation and deployment to take place. Finally, the cross-sectorial AI technology enablers represent the core technical competencies that are essential for the development of AI systems.

In June 2024 a big step forward has been made by the EU Parliament publishing the AI Act¹⁷, which is the first European regulation on Artificial Intelligence. Nevertheless ADRA (the AI, data and Robotics association) published in 2023 the "Strategic Orientation towards an AI, Data and Robotics Roadmap 2025-2027"¹⁸. Hence the European AI framework is still moving forward.

2. AI in manufacturing

• AI for dynamic production planning and management: This involves taking real-time decisions to optimise the factory operation by quickly modifying the productions schedule, based on the current state of the shop floor, predicted sales orders, unexpected events such as machine breakdowns or changes in job priorities, etc.

• Virtual models spanning all levels of the factory life and its lifecycle: A holistic and coherent virtual model of the factory and its production machinery will result from the contribution and integration of modelling, simulation and forecasting methods and tools that can strategically support manufacturing-related activities.

• AI for green/sustainable manufacturing: The development of software-based decision-support systems, as well as energy and resource management, monitoring and planning systems, will lead to overall reduced energy consumption, more efficient utilisation and optimised energy sourcing.

• AI applied in supply chain management: Planning and managing logistics for real-time operations, collaborative demand and supply planning, traceability, and execution, global state detection, time-to-event transformation, and discrete/continuous query processing would therefore be a challenge in view of the distributed nature of these elements.

• AI for advanced manufacturing processes: The ability to design functionality through surface modifications, functional texturing and coatings, enabling improved performance, embedded sensing, adaptive control, self-healing, antibacterial, self-cleaning, ultra-low friction or self- assemblies, for example, using physical (additive manufacturing, laser or other jet technologies, 3D printing, micromachining or photon- based technologies, physical vapour deposition, PVD) or chemical approaches (chemical vapour deposition CVD, sol–gel processes) will deliver high functionality and hence high-value products.

• AI for adaptive and smart manufacturing devices, components and machines: Embedded cognitive functions for supporting the use of machinery and robot systems in changing shop floor environments. Open Hardware based AI solutions will also support European AI Framework

3. AI for decision-making

Decision-making is at the heart of AI. Furthermore, Explainable AI will increase “human-centric” approaches in autonomous industry.

• AI can support complex decision-making processes or help develop hybrid decision-making technologies for semi-autonomous systems.

• Human decision-making, machine decision-making, mixed decision-making and decision support.

• Sliding or variable decision-making, dealing with uncertainty.

• AI for human interaction with machines.

4. AI for monitoring and control

• AI for control technologies.

• AI for monitoring services.

• AI for maintenance systems for increased reliability of production systems.

• AI services for continuous evaluation and mitigation of manufacturing risks.

• AI for quality inspection.

The volume and value of industrial services are increasing by between 5% and 10% every year. The share of services has exceeded the share of machinery for many machines, system and service vendors – not just for a final assembly factory, but also for companies in supply chains. Companies are willing to take larger shares of their customers’ businesses, initially as spare part suppliers, but increasingly for remote condition monitoring, as well as extending this to a number of those tasks previously considered as customer core businesses. From a customer point of view, such a shift in business models lies in the area of outsourcing.

Industry as services is changing the production through use of externalisation moving local tasks to external and ever more specialised providers, benefiting of greater flexibility, resilience and adaptability on production; a new proposition of supply chain is embedding all production phases, from the procurement of raw materials and semi-finished products up to the customer services and/or the design of parts or whole final product.

While many businesses have become global, some services are still provided locally, at close to customer’s locations, while other services are provided centrally by the original vendor or companies specialised in such services. Similarly, as there may be extensive supply chains underpinning the vendor companies, the respective services may also extend to supply chain companies. The industrial era is becoming a service era, enabled by high-end Chips JU technologies. This distributed setting conveniently fits into modern edge-to-cloud continuum innovative architectures as computing power engines and infrastructures enabling emulation, training, machine learning and communication platforms enabling real-time interconnections.

The importance of service businesses to the future is evident as they enable a revenue flow beyond traditional product sales, and more importantly they are typically much more profitable than the product sales itself.

The service business markets are becoming more and more challenging, while high income countries are focusing on the high-skilled pre-production and lifecycle stages. Fortunately, in the global service business market, Europe can differentiate by using its strengths: a highly skilled workforce, deep technology knowledge and proven information and communications technology (ICT) capabilities. However, to ensure success it needs new innovations and industry-level changes.

• Remote operations, teleoperation

  • Remote engineering and operations, telepresence: Operating or assisting in operations of industrial systems from remote sites.

  • Edge/cloud solutions: Implementing distributed service applications on effective edge cloud systems.

  • 5G with very low latency will be used for remote operations.

• AI Services for monitoring and collaboration

  • Collaborative product-service engineering, lifecycle engineering: Extending R&D to take into account how products and systems will be integrated into the industrial service programme of the company. This should possibly be enhanced by obtaining further knowledge to provide services for other similar products (competitors!) as well their own installed base.

  • Training and simulation: Complex products such as aircrafts, drones, moving machines and any tele-operated machineries need a simulation environment for proper training of the human driver/operator.

  • Condition monitoring, condition-based maintenance, anomaly detection, performance monitoring, prediction, management: The traditional service business sector is still encountering major challenges in practice. It will therefore require an extension to the above, as targets of services are expanded to other topics in customer businesses in addition to spare parts or condition monitoring.

  • Strategic partnership and data capitalization, "data as currency", will be key factors that will enable new business model inside AI service business

• Fleet management, Edge and local/global decision making

  • Decision and operations support: In most cases, decision-making is not automatic, whereas in the future it could be based on remote expert assistance or extensive diagnosing (AI-based, etc), engineering, and knowledge management systems.

  • Fleet management: This could benefit on the basis of sold items, by obtaining knowledge and experience from range of similar components and machines in similar or different conditions.

• Business services integration

  • Local and global services: Organising services locally close to customers and centrally at vendors’ sites.

  • Full lifecycle tutoring: Monitoring activities, level of stress and performance-oriented behaviour during the product’s life, from anticipating its end of life to properly handling its waste and recycling, including improved re-design for the next generation of products.

A “digital twin” is a dynamic digital representation of an industrial asset that enables companies to better understand and predict the performance and behaviour of their machines, and also to find new revenue streams. Currently, connectivity to the cloud allows an unprecedented potential for the large-scale implementation of digital twin technology for companies in various industries. A physical asset can have a virtual copy running in the digital twin environment, increasing revenue through continuous operational data. Similarly, several plants can operate in the same virtual space benefiting from globally shared data and remote automated decision-making. Factory/mill level digital twins are ever more autonomous and intelligent and can be adaptative by on-line learning. If needed they can interact to cloud level to evaluate model and optimize parameters using transfer learning at cloud.

Simulation capability is currently a key element in the European machine tool industry’s attempt to increase its competitiveness. In the Industry 4.0 paradigm, modelling plays a vital role in managing the increasing complexity of technological systems. A holistic engineering approach is therefore required to span the different technical disciplines and provide end-to-end engineering across the entire value chain.

In addition to virtual commissioning, modelling and simulation can more widely respond to many digitalisation challenges:

• visualising physical or real-world phenomena of products, production, businesses, markets, etc.

• helping designers to perform their core tasks – i.e., studying alternative designs, optimising solutions, ensuring safety, and providing testing for automation and Internet of Things (IoT) solutions.

• The effects of changes can be safely and more comprehensively assessed in advance in a virtual domain rather than using real plants, equipment or even mock-ups.

• Simulators offer versatile environments for users or operator training.

• It is evident that former computer-aided design (CAD)-driven digitalisation is shifting the focus towards simulation-based design.

Simulators may be used online and in parallel with its real counterpart to predict future behaviour and performance, provide early warnings, outline alternative scenarios for decision-making, etc., although they have years of research behind them, such tracking simulators are to be co-designed and improved exploiting also recent investments in computing infrastructures (e.g. HPC EU families, EPI initiative), with a special focus on the industrial context.

Telepresence technologies can also be considered as the predecessor for an extended reality (XR) presence. The combination of new and advanced technology like e.g. XR, Artificial Intelligence of Internet of Things (AIoT), Edge, High Performance Computing (HPC), 5G and open integration platforms offers significant potential for innovation, which would benefit the evolution of European digital industry.

As an example, XR is a combination of virtual and augmented reality, and an XR presence is a continuum between a physical reality presence and a virtual reality presence. The main driver here is improving competitiveness through better productivity, more effective worker safety and better quality. The industrial applications have followed the prospects offered by the gaming industry and consumer applications. One of the reasons for its increased take-up is the declining cost of electronic components and sensors.

As some major smart glass producers will provide technology and a platform for consumers, other EU industrial groups could do the same, and the EU industrial ecosystem could take benefit of such devices and provide industrial use cases. These can be extended then to state-of-art applications. “Back-end” services are Digital Twins and Condition Monitoring systems that will provide everything that optimizes the system work, providing critical and useful information for the Field Worker. Longer vision for the integration human actions and back-end data servers can be used to build knowledge graphs to help other users to work like experts. As a summary: smart glasses can extend human understanding and knowledge if services and information flow can be utilized and formed to usable knowledge.

The EC "Data Act" initiative¹⁹ issued in 2024 promotes reliable and secure access to data within the European data economy and helps to advance the digital transformation by managing various aspects of data sharing, exchange and use. In terms of industrial applications and global interoperability between different legal entities, this means that data ownership and data use must be properly managed. For example, when digital twin processes data supplied by a tool manufacturer (data holder), then the user of this tool must be informed about this activity.

• Digital Twin: Design process digitalisation, telepresence

  • Heterogeneity of systems: Information sharing and standards and means to ensure interoperability of digital twins and their information sources are important to facilitate information synchronisation. Having all relevant engineering disciplines (processes, assembly, electronics and electrical, information systems, etc.) evolving together and properly connected over the lifecycle phases is therefore crucial. This also involves multi-domain simulation, joint simulation of multi-simulation systems coupling.

  • Immersive telepresence for industrial robotics from design toward production lines and any other operational scope.

  • Digital twins applied to sustainability and circular economy: Simulate the usage of energy, use of raw material, waste production, etc., with the goal of improving energy efficiency and circular economy performance.

• Virtual commissioning, interoperability:

  • Virtual commissioning: Digital twins applied to virtual commissioning to bring collaboration between different disciplines and models from domains of engineering (mechanic, electronic, automation) in the same environment.

  • Interoperability: Applications cannot yet be used across platforms without interoperability.

• Simulators: Tracking & Simulator based design

  • Tracking mode simulation: Model adaption based on measurements. Generating simulators automatically from other design documentation, measurements, etc. Generation of simulators from 3D, data-driven models, etc.

  • Simulator-based design: Digital twin for testing the designed model by replacing the required physical components with their virtual models. This offers continuous design improvement (the digital twin provides feedback and knowledge gained from operational data), design optimisation, etc.

• Digital twins combined with data-driven models (knowledge and data fusion):.

  • Combination of data-driven and knowledge-based models along the complete lifecycle (product and production). The real challenge is to combine physics and knowledge-based models (digital twins) with data- driven models (models created using AI from massive acquired experimental data), capitalising on the strength of information present in each of them.

  • New ways to generate large 3D scenes have been based, for example, on NeRF (Near Radial Field) and/or the Gaussian Splats.

• Humans & Knowledge integration:

  • Human-in-the-loop simulations: Methods and simulations for human-in-the-loop simulations and integration of digital twins in learning systems for workers.

  • The 3D Internet platform: to integrate all of the aforementioned aspects into a single powerful networked simulation for humans to get the Situational Awareness for an industrial process as a whole.

  • Live 3D Digital Twins: to provide the state awareness (animation & color etc.) in AR/VR/XR.

Machines are usually more precise and efficient than humans when carrying out repeatable tasks. Thus, replacing or aiding work processes susceptible to human errors, quality defects and safety issues with machines will have an impact on quality and redundant waste. The application of AI in robotics and the same robotics without AI extends the opportunity for automation of manual tasks, increasing the sustainability, safety & security of the production and its green transition. This will reduce the environmental impact -efficient reduction of the waste and optimise product quality toward zero-defect, process, and manpower. This will also achieve a more safe and secure working area to ease the human-machine, and machine-machine co-working.

There are many kinds of autonomous systems, (co)-robots and working machines. Just to give an insight into their widespread adoption, can be categorised by purpose, as follows:

• Industrial machines and robots:

  • Manufacturing (e.g. welding, assembling, spray gun robots).

  • Material handling (e.g. conveyors, warehouse robots, trucks).

• Consumer robots:

  • Domestic (e.g. robotic lawn mowers or vacuum cleaners).

  • Care (e.g. lifting or carrying robots).

• Healthcare and medical robots:

  • Robotic surgery, hospital ward automation.

  • Medical tests and hospital care, remote healthcare.

  • Medical imaging, exoskeletons.

• Moving machines:

  • Mining machines (e.g. drilling machines, dumpers, conveyors).

  • Forestry (e.g. forest harvester), agriculture (e.g. tractors, appliances).

  • Construction (e.g. excavators, road graders, building robots).

  • Logistics and sorting centres (e.g. cranes, straddle carriers, reachers, conveyor belts, sorting machines, trucks).

  • Military robots and machines.

• Transport:

  • Vehicles, trucks and cars, trains, trams, buses, subways.

  • Aviation (e.g. aeroplanes, helicopters, unmanned aerial vehicles, UAVs).

  • Marine (e.g. vessels, ships, auto-piloted ships), submarine (e.g. auto-piloted submarines).

• Utilities and critical infrastructures:

  • Extraction (e.g. drills for gas, oil).

  • Surveillance (e.g. quadcopters, drones).

  • Safety, security (e.g. infrared sensors, fire alarms, border guards).

  • Energy power plants sensors and actuators (e.g. production and distribution).

  • Transportation (e.g. moving bridges, rail exchanges).

The main aims and evolution trends of robots and autonomous systems in digital industry are oriented toward:

• Production efficiency, speed and reduced costs,

• Higher precision and quality, safety in working conditions,

• To scale up “smart and high-end manufacturing”.

As is evident from the above, robots and machines and in particular autonomous systems are involved in all application chapters of this SRIA in addition to Digital Industry – i.e. Digital Society, Health and Wellbeing, Mobility, Energy, and Agrifood and Natural Resources- and are positively impacted by any improvement on both technological layers of the SRIA; foundational and cross-sectional technology.

There is undoubtedly a move to increase the level of automation and degree of digitalisation in industry, which will ultimately lead to fully autonomous systems.

However, between low and high technology manufacturing (two extremes: entirely manual and fully autonomous), there will always lie a large area of semi-autonomous equipment, units, machines, vehicles, lines, factories and sites that are worth keeping somewhat below 100% autonomous or digitised. The reasons for this include:

• A fully autonomous solution may simply be (technically) near to impossible to design, implement and test.

• If achievable, they may be too expensive to be realised.

• A fully autonomous solution may be too complex, brittle, unstable, unsafe, etc.

• A less-demanding semi-automatic solution may be easier to realise to a fully satisfactory level.

When the extent of automation and digitalisation are gradually, reasonably and professionally increased, often step by step, they may bring proportionally significant competitive advantages and savings that strengthen the position of digital industries overall. However, since the extent of automation and digitalisation remains well below 100%, any potential negative effects to employment are still either negligible or non-existent. On the contrary, the competitive advantages due to the adoption of robotics and autonomous system solutions increases the market positions of companies and, generally, enhances the need for more people in the respective businesses. An evolution of work organization and leadership may be required to enable the full potential of the digital industry.

1. Autonomous functions of systems:

Advances in Artificial Intelligence simplify the development of fully autonomous systems, which can fully automate work tasks in various application domains by increasingly reducing the need for human intervention and, at the same time, allowing to highlight additional innovative functions. A prominent example can be found in the automotive domain, where autonomous vehicles are expected to play a key role in the future of urban transportation systems. Such a challenge in a so complex scenario will promote significant advances in any autonomous functions of the many systems which are integral part of the Digital Industry ecosystem.

Autonomous robotics is the key enabling technology for the implementation of autonomously functioning shopfloors. Immediate benefits will be additional safety, increased productivity, greater accessibility, better efficiency, and positive impact on the environment. The following figure provides a generalised view of technologies and functionalities used in autonomous systems developed in the automotive domain. Nevertheless, this schematic representation of required building blocks can be generalized to many other types of autonomous systems.

The following figure shows matching of ISA95 standard that define control and other enterprise functions with building blocks of autonomous systems. The autonomy is expected to increase in the level 2 and Level 3 of ANSI/ISA95²⁰.

2. Safety and security in autonomous systems:

Current standards of safety requirements for autonomous machines categorise safety into four approaches.

• On-board sensors and safety systems for machines that work among humans and other machines but is restricted to indoor applications.

• An isolated autonomous machine that works in a separated working area, mostly an intensive outdoor environment where other machines or humans are monitored.

• Machine perception and forecast of expected and/or unexpected human activities aimed at: (i) assisting human activities and movements with a proactive behaviour; (ii) preserving human health and safety; and (iii) preserving the integrity of machinery.

• An operator is responsible for reacting to a hazardous situation encountered by the autonomous machine when being provided with enough time between alert and transferring responsibility.

3. Requirements management and conceptual modelling of autonomous systems:

With the increasing complexity of autonomous functionality in both AV and ADAS systems, traditional methodologies of developing safety critical software are becoming inadequate. This observation can be made not only for autonomous driving, but in any industrial field of application of autonomous systems. Since autonomous systems are designed to operate in complicated real-world domains, they will be expected to handle and react appropriately a near endless variety of possible scenarios, meeting expectations from various stakeholders such as the internal engineering teams, people involved in the autonomous systems managements (e.g. passengers, drivers, workers, e-Health patients), regulatory authorities, and commercial autonomous vehicles/robots fleet operators.

3. Human–machine interaction in autonomous systems:

Improvements on sensing capabilities, on actuation control, IIoT (Industrial IoT) and SoS distributed capability are, in robotics and autonomous systems, the key enabler for:

• Human-robot interaction or human-machine cooperation.

• Transparency of operations between human and Advanced Machine Systems (AMS) in uncertain conditions.

• Remote operation and advanced perception, AS oversight and tactical awareness.

• Autonomy intended to enhance human capabilities.

• Natural human interaction with autonomous systems.

• Assisted, safety-oriented and proactive robot interaction with humans.

4. Digital design practices including digital verification and validation (V&V):

• Automatic or semi-automatic V&V.

• A digital design environment, digital twins, physical mock-ups.

• Sub-task automation development, generation of training data and testing solutions and field data augmentation, according to a handful of global machine manufacturers.

• Machine state estimation (assigning a value to an unknown system state variable based on measurements from that system).

5. Simulators and autonomous systems:

• Process model based and product 3D-models approaches, environment and object models, and simulation tools.

• Early design phase simulators.

• Robotic test environments.

• Empirical or semi-empirical simulators, making use of both real and simulated data collected from previous experiments.

• Off-road environments.

6. Autonomous capabilities development in a digital environment

• Autonomous decision taking.

• Self-evolving capabilities.

• Exploitation of knowledge in cognitive flexibility and in adaptability of the reaction.