Digital Twins – success factor for Virtual Engineering

Use Digital Twins to network your systems flexibly and test them virtually

The digitalization of systems is a megatrend that encompasses all areas of application. System boundaries are disappearing and networking across system boundaries is coming into focus.

Digital representatives, for example Digital Twins, are one means of achieving this. They map the current state of the system, provide uniform interfaces to data and services, and can predict a system’s reaction to changes. A Digital Twin enables virtual tests, shortens development times and thus the time-to-market of new products. Virtual Engineering comprises the activities involved in creating and using digital representatives. Thus, Virtual Engineering supports developers and engineers in optimally integrating Digital Twins into processes.

Digitaler Zwilling im Automotive / Virtual Engineering - Fraunhofer IESE
© Fraunhofer IESE
Digitale Zwillinge bilden alle relevanten Eigenschaften ihrer realen Gegenstücke ab.

What is a Digital Twin? And what is Virtual Engineering?

Definition Digital Twin

A Digital Twin is a virtual representation of a physical asset. It represents the current state of a system and can thus predict the behavior of a system. This enables virtual tests at development time, or the evaluation of decisions at runtime. The implementation of self-optimizing systems is made significantly easier with Digital Twins. In addition to being a mere representation of a system, Digital Twins can also realize bidirectional interfaces that allow influencing a system. This makes the Digital Twin a tool for the cross-sector coupling of systems.

A distinction is made between the Digital Shadow and the Digital Twin. The Digital Shadow represents the state of a real system, all relevant components, as well as the environment digitally. This also includes, for example, users, certificates, and processes. Uniform interfaces ensure coupling across system boundaries and enable cross-system interaction. By integrating simulation models, the behavior of a system can be predicted and virtual test environments can be realized, for example. A Digital Twin additionally has the ability to exert a controlling influence on a system. Integrated safety concepts then ensure that the Digital Twin and the real system exhibit predictable behavior at all times.

Digital Twins – Examples

Nowadays, many systems are interconnected and communicate with each other. Nevertheless, communication is often limited: The use of different protocols and formats frequently prevents direct data exchange. In a production environment, for example, an Enterprise Resource Planning (ERP) system can therefore not easily access data from the shop floor. However, this data is necessary for digital production, as processes increasingly need to be documented automatically and flexible production processes need to know the status as well as the capacity utilization of a plant at any time. The Digital Twin realizes a complete image of the production that comprises the current state of the plant and allows making controlled changes to it.

In automotive engineering, Digital Twins can realize high-precision Software-in-the-Loop (SiL) test environments in a virtual testbed by coupling all relevant system components – buses, virtual ECUs, failure models, and driving functions. This is already reducing test costs today. Highly automated and autonomous driving functions have to complete a large number of test kilometers to get approval, so they can no longer be tested at all by any other means.

Digital Twins are therefore particularly suited for systems in which systems that previously acted in isolation are digitally integrated with each other. Likewise, HW/SW codesign and the simulation of system environments are also supported by Digital Twins. This challenge can be found in traditional industries such as automation, automotive and commercial vehicle technology, or medical technology, but also in cross-sector coupling, for example in the integration of vehicles with mobility services.

Emergence and Market Relevance

Originally, Digital Twins were introduced to reduce the high cost of testing aircraft by means of simulations. For this purpose, models of the real systems were to be used that represent all relevant properties, and not just a subset of these. Today, Digital Twins are therefore used in many areas. For a long time now, not only aircraft have been tested with them; in automotive engineering, too, virtual ECUs are tested in a Digital Twin of the vehicle, which reflects its hardware, buses, failure patterns, environment, and the driving behavior. In a production environment, Digital Twins are used to identify optimization potential in complex processes, but also to test changes in production first on a virtual representation of the production line before transferring these to the real world.

In the future, systems will increasingly have to make decisions autonomously. If functions of a vehicle are to be dynamically upgraded at a later point in time, or if production is to be dynamically optimized, the systems involved have to be able to assess the effects of these changes. Due to these requirements, Digital Twins will gain even more importance in the future.

Relationship to Virtual Engineering

Virtual Engineering supports development processes with the help of digital models, for example by simulating system components and SiL tests, in order to integrate components virtually. Virtual Deployment is used to test software components on the target platform and to check whether, for example, response times are adhered to or whether the response to failure patterns corresponds to the expected behavior. Digital Twins of components such as ECUs, software components, and processes are the building blocks for successful Virtual Engineering.

Virtual Engineering uses Digital Twins to accelerate development cycles, develop alternative product and solution concepts, and get early feedback on the impact of decisions as well as the expected result. Virtual Engineering therefore encompasses the complete process, the required infrastructure, and the data management strategies for Virtual Engineering. By shortening development and change processes, the use of Digital Twins saves hard cash.

Back to Overview

Digitaler Zwillinge in der Fabrik / Virtual Engineering - Fraunhofer IESE
© Fraunhofer IESE
Schnellere Time-to-Market: Test- und Freigabeprozesse mit Digitalen Zwillingen virtualisieren.

Digital Twins and Virtual Engineering:
What are the opportunities?

Faster Time-to-Market

Time is money! Often the winner is not who was able to place the best product on the market, but the one who placed it first. In addition to the actual development and implementation, however, it is often tests and troubleshooting that delay completion. Digital Twins realize digital representations of systems as virtual test environments. Changes, the integration of systems, and new developments can be tested in this way. With a Digital Twin, you can also reliably reproduce rare situations, such as hardware defects, operating mistakes, or transmission errors.

 

Cost savings

It is often difficult to understand how complex systems work. There is a reason why the first steps on the path to Industrie 4.0 are “seeing” and “understanding” one’s own systems. Digital Twins realize a holistic representation of a system. They can merge data from various sources and thus identify optimization potential. These interfaces are also suitable for long-lived systems and therefore enable lasting access to data from different sources.

 

Higher quality

Changes in systems often entail unintended misbehavior, so-called side effects. Digital Twins provide a virtual representation of the target system and its environment and thereby enable the establishment of a Continuous Integration / Continuous Engineering environment even for embedded systems. Your system is thus automatically tested with every change, and you can change your software quickly and efficiently and also detect unwanted side effects efficiently.

 

Enabler for autonomous systems

While current systems have only a limited scope for making decisions, autonomous systems, such as autonomous vehicles, make significantly more decisions independently. To make the right decision in the current situation, perception of the environment is necessary. Autonomous systems are therefore often interconnected and share sensor information across physical system boundaries. In these cases, simulations are a prerequisite for quality assurance and safety assurance. We carry out research for you on how to realize self-optimizing systems and autonomous decisions with Digital Twins. Do you want to profit from this research?

 

Enabler for Industrie 4.0

Predictive maintenance, which prevents unplanned downtime, is almost an Industrie 4.0 classic. The data required for this is provided by Digital Twins. Technically, this is realized with the help of an asset administration shell and submodels. More complex Industrie 4.0 applications also become possible with Digital Twins. Virtual commissioning shortens setup times, and flexible production can also manufacture small lot sizes in a cost-effective manner. Our open-source Industrie 4.0 middleware Eclipse BaSyx realizes the platform for your digital manufacturing. Contact us!

 

Back to Overview

Digital Twins: What are the challenges?

Communication
and interfaces

The Digital Twin represents the current state of a system. For this purpose, all relevant data, such as telemetry data, sensor values, data sheets, and test results from different sources must be integrated into a digital model. In doing so, it must also be taken into account that interfaces, as well as sensors and their parameters, for example, may change during the lifetime of a system. Interfaces must therefore define reusable data models that are independent of the components used.

Handling
of data flows

A Digital Twin is a representation of a system. Often, this representation does not contain all raw data, but rather a simplified view of it. In certain situations, for example when errors need to be traced, it is then necessary to access more extensive raw data. A distributed architecture must ensure that the data can be preprocessed where it arises and that, on the one hand, only necessary data is distributed through the network, but on the other hand, all information is made available when required.

Uniform understanding of data

If Digital Twins are used across company boundaries, different data formats and units of measurement are often used. In addition, the meaning of data is not fully defined. If Digital Twins are used across companies, for example to provide mobility services or to implement supply chains, then a common understanding of data must be ensured.

System behavior

In order to accurately predict a system’s behavior, the Digital Twin must represent all aspects of the real system. However, since a simulation model can usually only simulate a few of these aspects, numerous simulation models must be integrated into one simulation to create a Digital Twin. To do so, execution and calculation models must be integrated with each other, and it must be ensured that the coupling error remains within an acceptable range.

System and
system environment

A particular challenge in the simulation of autonomous systems is the creation of suitable digital representations of both the technical systems and their operational environment. In automated driving, for example, the operational environment or “Operational Design Domain” must include all relevant aspects in order to be able to draw valid conclusions from the simulation results with regard to safety.

Dependability

A real Digital Twin is characterized by the fact that it can exert a controlling influence on a real system. Independent of whether the Digital Twin adjusts parameters on the basis of data or actively controls the system, functional safety, reliability, as well as availability requirements must be taken into account when developing a Digital Twin.

How does a Digital Twin work?

In complex systems, different interfaces are often used. There are sound reasons for this: Bus protocols are optimized for real-time-capable or high-frequency transmission of small amounts of usage data, whereas Ethernet, for example, shows the greatest performance in the transmission of large amounts of data. Very simple end devices often do not implement any protocols at all, but communicate a certain voltage level, for example. To convert that into a numerical value, the characteristic curve of the device must be known. This leads to very real problems: For example, characteristic curves are often stored in the control software. There is no meaningful separation between the representation of information and information processing. If parts of systems must be adapted or replaced, extensive changes are the result. A Digital Twin therefore provides a digital system representation that maps the state of the system to a logical model. This makes it irrelevant to the user how a particular value is obtained. What is important is that the user can use the Digital Twin to access a representation of the system that contains the required information.

 

Acatech Maturity Model for Industrie 4.0

The Acatech maturity model for Industrie 4.0 describes the possible maturity levels. Although these are actually defined for the maturity of a production, they can be used across the board to describe Digital Twins:

  • Maturity level 3 is the first one that is relevant for Digital Twins. It calls for connectivity, but not for uniform interpretation of data and service. It is sufficient to provide data. The users must ensure that the data is interpreted correctly.
  • Level 4 of the maturity model requires that the Digital Twin holds information in a uniform format. Depending on the usage, formats must be converted or metadata must be used. This metadata describes, for example, whether a value was measured or how reliable a value is.
  • Maturity level 5 expects a predictive model describing system behavior in a defined environment. This can be used, among other things, to predict changes in behavior.
  • Maturity level 6 offers the possibility to optimize a system and thus exert an influence over it. This is only possible with a Digital Twin; a Digital Shadow is no longer sufficient for this.

 

The users of Digital Twins are diverse

Users can be operators who use a dashboard to monitor the state of a plant and identify optimization potential. However, the user of a Digital Twin can also be another Digital Twin – for example, the plant twin can aggregate its state from the Digital Twins of the devices and the workers. Digital Twins are also used during development and during testing activities.

 

In order for Digital Twins to represent all properties of a system, several simulation models often have to be coupled with each other.
© Fraunhofer IESE
In order for Digital Twins to represent all properties of a system, several simulation models often have to be coupled with each other.

Digital Twins make test processes easier

During development, there are often only a few prototypes or systems available for tests. This leads to long waiting times and high costs. In automotive engineering, for example, new driving functions are tested in a dedicated test environment, the Hardware-in-the-Loop (HiL) testbeds. These replicate relevant system components, but they are expensive and therefore a scarce resource. Digital Twins of vehicles are a significantly cheaper alternative, which, once they have been developed, can be replicated an unlimited number of times and can accelerate test processes. Digital Twins also make it possible to clock up the huge number of kilometers required for the qualification of autonomous driving functions. For this purpose, Digital Twins are integrated into a co-simulation. (The figure above shows an example of this.)

In order to be able to make meaningful statements about the system behavior, simulation models of all system components have to be integrated with each other in a Digital Twin of the overall system. The Digital Twin depicted in the figure comprises the networking infrastructure with failure models, driving dynamics, sensors, and actuators, as well as the system environment. We integrate simulation models as components, which can be interconnected via the Functional Mockup Interface (FMI), for example. The Digital Twin is created by coupling these models, so that all relevant system properties are captured.

The models of the communication buses realize an exact representation of the timing behavior of the communication, including arbitration and, if necessary, error signaling and correction. Platform models describe the target platform. Our simulation platform combines these models into an integrated simulation, so that a Digital Twin can be assembled from existing components, if required. The particular challenge lies in integrating live data from a real system. Numerical effects must also be taken into account here in order to reduce coupling errors.

 

Back to Overview

How can Fraunhofer IESE support your company in Virtual Engineering?

We will be happy to support you with our tools and our knowledge in the digitalization of your systems. Your concrete issue is not listed here? Contact us anyway. Just like you, we enjoy working on new, revolutionary solutions and projects.

Architectures for and with Digital Twins

A Digital Twin often requires new architecture concepts. Data models must be developed, suitable technologies selected, and quality properties considered. Will the networks used be able to cope with the increasing requirements for communication? What happens in case of a failure? How to realize a service-based architecture? We will be happy to develop an architecture concept for your systems and advise you on all matters related to digitalization.

Virtual testbeds

Test your functions in virtual electronic control units (ECUs), create virtual vehicles with buses, ECU platforms, and failure models, integrate existing simulators into your virtual testbeds. For example, our connectors combine Simulink models and Functional Mockup Units (FMU) of the FMI standard into a joint co-simulation. With our C- or Java-based interfaces, we will develop a simulation environment that is customized to your needs.

Continuous Integration / Continuous Deployment

With Continuous Integration (CI) and Continuous Deployment (CD), you can implement new functions very quickly in a continuous process and test them. New functions thus reach the necessary market maturity faster. Our co-simulation solution FERAL supports you in implementing CI/CD solutions even for embedded systems, such as those very frequently found in automotive engineering and production.

Digital manufacturing processes

Use our open-source solution Eclipse BaSyx to create a virtual data space with asset administration shells. The Digital Twins for your processes, resources, and products will digitalize your manufacturing processes. Realize a virtual control room that presents the state of your products, devices, and processes at one glance and enables predictive maintenance, product tracking with automated documentation, or production for small lot sizes. With Eclipse BaSyx, you can digitalize your manufacturing processes and use an open platform. In this way, you solve the challenges of today while at the same time laying the foundation for the future. Do you need support or additional features? Feel free to contact us.

Safe coupling

Future systems will be specialists and team players! In order to solve a task optimally, systems must get together, like in a convoy on the highway. With Digital Dependability Identities, you can already determine dynamically today whether a coupling is safe or not.

Safety concepts

Proven safety concepts can only be used to a limited extent in the context of highly automated and networked systems. Dynamically coupled systems therefore require dynamic safety concepts. The information needed for this purpose does not only need to be available at runtime, but must also be available across systems. A fundamental approach is therefore to shift safety assurance to runtime with safe Digital Twins. Systems become aware of their safety requirements and are enabled to dynamically adapt and guarantee them. Are you still missing a suitable concept? We will be glad to help you!  

Data sovereignty

Data is the new gold. Do you want to give away your data for free and let unauthorized people see your secrets? Make sure that only authorized people can access your data. With MY DATA Control Technologies, we offer you a data usage control solution that you can integrate into your systems.

Platforms and tools for digitalizing your systems

FERAL

 

You want to build a Digital Twin, for example as a virtual testbed for driving functions, but you don’t know how to integrate tasks, bus communication, networks, failure behavior, and functions with each other? You are thinking about a Continuous Integration solution for your development projects? You want to know whether your network is suitable for the communication requirements of Industrie 4.0?

Fraunhofer FERAL simulates systems of systems. With that, we also realize your simulation solution.

Eclipse BaSyx

 

We are instrumental in driving the development of Eclipse BaSyx, an open-source middleware for Industrie 4.0. We do this because we know that Industrie 4.0 is primarily a software revolution that requires the incorporation of state-of-the-art software concepts into production systems.

In the meantime, Eclipse BaSyx is not only used in production, but also, for example, to monitor buildings and infrastructure and make data available in real time.

MY DATA Control
Technologies

Because your data is important, we offer you our data usage control solution. With this, you determine who may use your data, when, and for what purpose

DRAMSys

You develop embedded systems, but the DRAM memory is too slow? DRAMSys is a framework for the optimization of DRAM memory systems through simulation analysis. Optimize memory usage and even memory controllers for your application domains.

Why should your company collaborate with Fraunhofer IESE in Virtual Engineering?

 

 

Digitale Zwillinge bei der Fahrzeugentwicklung / Virtual Engineering  - Fraunhofer IESE
© Fraunhofer IESE
Wir erstellen Digitale Zwillinge Ihrer Systeme.

We support you in the conception and implementation of Virtual Engineering with Digital Twins.

In this way, you avoid expensive wrong decisions and rely on state-of-the-art technologies for the integration of your systems.

 

We develop modern simulation solutions for you.

This ensures that you can test your systems virtually, save valuable time and resources, and benefit from short development cycles.

 

We develop strategies for open systems with you.

In this way, we ensure that your systems can also be integrated into a digitalized environment.

 

We realize completely new safety concepts with you.

With this, we enable systems and system functions that you would have considered inconceivable until now.

 

We develop virtual testbeds with you.

With these, you can integrate system components, buses, failure models, and user behavior and qualify/certify your systems virtually in the future.

 

We digitalize your production.

This allows you to identify optimization potential, commission devices and manufacturing lines virtually, and enable data analysis.

 

We realize Continuous Integration / Continuous Deployment with you.

With this, we shorten your development cycles significantly and find unexpected errors as well as side effects already during development.

 

The Fraunhofer-Gesellschaft is the largest applied research organization with a wide range of expertise in more than 70 institutes.

This ensures you access to technology and engineering expertise from all disciplines relevant for your ecosystem in all domains within the Fraunhofer partner network. 

We will be happy to support you in mastering your challenges in the area of Digital Twins and Virtual Engineering.

We develop customized solutions, work on the current state of the art, are independent and neutral. Benefit from our experience!

 

Contact us!

 

We will be happy to support you and make time for you! Schedule an appointment with us, by email or by phone.

Which references does Fraunhofer IESE have regarding Virtual Engineering and Digital Twins?

We work with pioneers in the field of Digital Twins. Together with Robert Bosch GmbH, for example, we are researching simulation solutions that integrate the relevant vehicle components into a Digital Twin for the virtual integration of ECUs. John Deere uses our simulations solutions to evaluate architecture decisions according to various quality characteristics of highly automated platforms. And meanwhile, even the inventor of the DRAM standard (Rambus) is licensing our DRAM simulators in order to further develop this technology and make it even better.

Collaboration Partner


“For Fraunhofer IESE and Rambus, this cooperation is an important partnership and a huge step towards a broader application and transfer of the DRAMSys tool in companies."

 


James Tringali
Technical Director, Rambus

 

Memory Technology

DRAMSys with Rambus

DRAMSys is a Digital Twin of the DRAM subsystem for the evaluation of bandwidth, latency, and power consumption. In cooperation with Rambus, further development of the DRAM simulators is underway.

Automotive Engineering

Dynamic Validation

With Digital Twins, we solve central challenges regarding the safety assurance of autonomous vehicles. We published more about this topic at “The Autonomous Event – Safety & Architecture 2021”.

Collaboration Partner


“BaSyx makes the commissioning of a new production line about 30 percent faster.”

 


Gerhard Schaller
Director Digitalization Operations für die ZF Division Electrified Powertrain

 

Industrie 4.0

Projekt BASYS 4

In the BaSys 4 project, we are developing the open-source Industrie 4.0 middleware Eclipse BaSyx for the implementation of Digital Twins.

 

Industrie 4.0

Safety Engineering with Digital Twins

With our Digital Twins, we are realizing modular, reusable safety concepts for lot size 1 and Plug&Safe solutions. (link to follow)

Automotive Engineering

Robert Bosch

Virtual validation: Together with Robert Bosch GmbH, we are developing virtual Hardware-in-the-Loop tests (HiL) for the integration of driving functions.

Customer Statement


“The simulation framework FERAL of Fraunhofer IESE is characterized by its high flexibility. This enables us to cope with numerous tasks.”

 


Dr. Roland Samlaus
Robert Bosch GmbH

 

Customer Statement


“Eclipse BaSyx is an important step for process automation in enterprises and in public administration. In the partnership with Fraunhofer IESE, Cisco supports the research on BaSyx.”

 


Jonas Rahe
Öffentliche Hand, Cisco Systems GmbH

 

Industrie 4.0

Industry 4.0 Made Easy

Together with Intel and Cisco, we describe the easy implementation of Industrie 4.0 with Digital Twins and FlexPod. (link to follow)

Industrie 4.0

DigiPro4BaSys

We are performing a retrofit with Digital Twins with the goal of enabling a holistic view of the factory. (link to follow)

Industrie 4.0

SAP AAS

With SAP AAS, we are implementing asset administration shells and their serialization based on Eclipse BaSyx. (link to follow)

Customer Statement


“With open-source projects such as Eclipse BaSyx, initial prototypes can already be realized today, which accelerates the step towards product-ready implementation of Industrie 4.0 concepts.”

 


SAP AG

 

 

Podcast

Listen instead of reading

In the MORGEN DENKER podcast, Dr. Thomas Kuhn explains how Fraunhofer IESE solves problems with the help of Virtual Engineering and Digital Twins. Listen in now!

Publications from the area of Digital Twins and Virtual Engineering

Publications of Fraunhofer IESE in the subject area of “Digital Twins and Virtual Engineering”

    • F. Schnicke, T. Kuhn, P.O. Antonino: Enabling Industry 4.0 Service-oriented Architecture through Digital Twins. IEEE European Conference on Software Architecture (ECSA), 2020.
    • Z. Müller-Zhang, P.O. Antonino, T. Kuhn: Dynamic Process Planning based on Digital Twins and Reinforcement Learning. International Conference on Emerging Technologies and Factory Automation, 2020.
    • P.O. Antonino, F. Schnicke, Z. Zhang, T. Kuhn: Blueprints for architecture drivers and architecture solutions for Industry 4.0 shopfloor applications. Proceedings of the 13th European Conference on Software Architecture-Volume 2, 2019
    • T. Kuhn, P.O. Antonino. F. Schnicke: Industrie 4.0 Virtual Automation Bus Architecture. IEEE European Conference on Software Architecture (ECSA), 2020.
    • T. Kuhn, F. Schnicke, P.O. Antonino: Service-Based Architectures in Production Systems: Challenges, Solutions & Experiences. IEEE 2020 ITU Kaleidoscope: Industry-Driven Digital Transformation (ITU K), 2020.
    • T. Kuhn, P.O. Antonino, A. Bachorek: A Simulator Coupling Architecture for the Creation of Digital Twins. IEEE European Conference on Software Architecture (ECSA), 2020
    • P.O. Antonino, J. Jahic, B. Kallweit, A. Morgenstern, T. Kuhn: Bridging the Gap between Architecture Specifications and Simulation Models. IEEE International Conference on Software Architecture Companion (ICSA-C), Northeastern University, Washington · United States, 2018
    • A. Bachorek, F. Schulte-Langforth, A. Witton, T. Kuhn, P.O. Antonino: Towards a Virtual Continuous Integration Platform for Advanced Driving Assistance Systems. IEEE International Conference on Software Architecture Companion (ICSA-C), 2019
    • M. Jung, L. Steiner, N. Wehn: The Open Source DRAM Simulator DRAMSys4.0. IEEE/VDE 24. Workshop „Methoden und Beschreibungssprachen zur Modellierung und Verifikation von Schaltungen und Systemen” (MBMV 2021), March. 18-19, 2021, Munich
    • M. Jung, C. Weis, N. Wehn: DRAMSys: A flexible DRAM Subsystem Design Space Exploration Framework. IPSJ Transactions on System LSI Design Methodology (T-SLDM), October, 2015.
    • E. Cioroaica, S. Chren, B. Buhnova, T. Kuhn, D. Dimitrov: Towards creation of a reference architecture for trust-based digital ecosystems. Proceedings of the 13th European Conference on Software Architecture-Volume 2, 2019
    • E. Cioroaica, B. Buhnova, T. Kuhn, T. Schneider: Building Trust in the Untrustable. IEEE/ACM International Conference on Software Engineering: Software Engineering in Society Track (ICSE-SEIS), 2020
    • E. Cioroaica, S. Cjrem, O-E-K. Akouf, A. Larsson, R. Chillarege, T. Kuhn, D. Schneider, C. Wolschke: Towards Creation of Automated Prediction Systems for Trust and Dependability Evaluation. IEEE International Conference on Software, Telecommunications and Computer Networks (SoftCOM), 2020
    • J. Reich, J. Frey, E. Cioroaica, M. Zeller, M. Rothfelder: Argument-Driven Safety Engineering of a Generic Infusion Pump with Digital Dependability Identities. Springer International Symposium on Model-Based Safety and Assessment (IMBSA), 2020
    • J. Reich, D. Schneider, I. Sorokos, Y. Papadopoulos, T. Kelly, R. Wei, E. Armengaud, C. Kaypmaz: Engineering of Runtime Safety Monitors for Cyber-Physical Systems with Digital Dependability Identities. International Conference on Computer Safety, Reliability, and Security (SAFECOMP), 2020