A Cross-Domain Systematic Mapping Study on Software Engineering for Digital Twins

Full text

Journal of Systems and Software 00 (2022) 1–59
Journal
Logo
A Cross-Domain Systematic Mapping Study on Software
Engineering for Digital Twins
Manuela Dalibor1,∗
, Nico Jansen1, Bernhard Rumpe1, David Schmalzing1, Louis Wachtmeister1, Manuel Wimmer2,
Andreas Wortmann3
Abstract
Digital Twins are currently investigated as the technological backbone for providing an enhanced understanding and management
of existing systems as well as for designing new systems in various domains, e.g., ranging from single manufacturing components
such as sensors to large-scale systems such as smart cities. Given the diverse application domains of Digital Twins, it is not
surprising that the characterization of the term Digital Twin, as well as the needs for developing and operating Digital Twins are
multi-faceted. Providing a better understanding what the commonalities and differences of Digital Twins in different contexts
are, may allow to build reusable support for developing, running, and managing Digital Twins by providing dedicated concepts,
techniques, and tool support. In this paper, we aim to uncover the nature of Digital Twins based on a systematic mapping study
which is not limited to a particular application domain or technological space. We systematically retrieved a set of 1471 unique
publications of which 356 were selected for further investigation. In particular, we analyzed the types of research and contributions
made for Digital Twins, the expected properties Digital Twins have to fulfill, how Digital Twins are realized and operated, as well
as how Digital Twins are finally evaluated. Based on this analysis, we also contribute a novel feature model for Digital Twins from
a software engineering perspective as well as several observations to further guide future software engineering research in this area.
©2011 Published by Elsevier Ltd.
Keywords: Software Engineering, Digital Twins, Manufacturing, Industry 4.0
1. Introduction
Research and industry leverage Digital Twins to monitor and control (cyber-physical) systems in various domains,
including autonomous driving [104], biology [266], medicine [26], smart manufacturing [256], and many more. They
promise tremendous potential to reduce cost and time and improve our understanding of the represented systems. The
various Digital Twins serve different purposes, including analysis [214], control [292], and behavior prediction [66],
and they are used at different times relative to the represented system, e.g., before it exists to explore its design
space [274] or during its runtime to optimize its behavior [1]. Despite plethora of definitions [70, 115, 166, 179, 253]
there is little consensus about what a Digital Twin actually is. This also is reflected in many of the available definitions
∗All authors contributed equally to this research.
Email addresses: dalibor@se-rwth.de (Manuela Dalibor ), jansen@se-rwth.de (Nico Jansen ), rumpe@se-rwth.de
(Bernhard Rumpe ), schmalzing@se-rwth.de (David Schmalzing ), wachtmeister@se-rwth.de (Louis Wachtmeister ),
manuel.wimmer@jku.at (Manuel Wimmer ), andreas.wortmann@isw.uni-stuttgart.de (Andreas Wortmann )
1Software Engineering, RWTH Aachen University, Germany
2Department of Business Informatics – Software Engineering, Johannes Kepler University Linz, Austria
3Institute for Control Engineering of Machine Tools and Manufacturing Unit (ISW), University of Stuttgart, Germany

being (1) ambiguous, by deferring to another undefined term, such as a “virtual representation” [311], a “computable
virtual abstraction” [251] , or a “a virtual projection of the industrial facility into the cloud” [53]; (2) narrow, by fo-
cusing on specific use cases, domains, or technologies, such as a “digital model of the real network environment” [88]
or a ”virtual representation based on AR-technology” [214]; or (3) utopian, due to all-encompassing aspirations, such
as an “integrated virtual model of a real-world system containing all of its physical information” [92], a “complete
digital representation” [65]. Instead of producing more of such definitions, we aim to uncover the nature of Digital
Twins as documented in literature bottom-up. To this end, we conducted a systematic mapping study [385, 386, 387]
on Digital Twins to investigate the following research questions (RQs):
•Who uses Digital Twins for which purposes (RQ-1)?
•What are the conceptual properties of Digital Twins (RQ-2)?
•How are Digital Twins engineered (RQ-3), deployed (RQ-4), and operated (RQ-5)?
•To which extend are Digital Twins evaluated (RQ-6)?
Following a detailed search strategy involving five digital libraries, we initially found 1472 unique publications.
Out of these, 624 publications were identified as potentially relevant of which 356 publications were finally selected
and categorized using a comprehensive classification scheme focusing on the contribution types, research types, prop-
erties, implementation methods, deployment details, operation decisions, and evaluation means of Digital Twins. The
resulting research landscape developed by this study can help to understand, guide, and compare future research in
this field across different domains. In particular, this paper identifies common and varying Digital Twin features and
identifies challenges that seem to be less investigated. The contributions of this paper, hence, are
1. A first systematic and comprehensive study on Digital Twins across different domains, applied implementation
technologies, and purposes.
2. A novel feature model of common Digital Twin features to guide researchers and practitioners in making deci-
sions about the design, development, deployment, and operation of Digital Twins.
3. A synthesis of observations on the landscape of Digital Twin research that may guide further research.
In the following, Section 2 presents related studies and discusses how our survey differs in width, depth, and
research method. Afterward, Section 3 details our research method, questions, queries, and data collection. Then,
Section 4 answers these research questions individually and Section 5 investigates correlations between the answers to
these questions. Based on our findings, Section 6 presents a novel model of common Digital Twin features. Section 7
discusses threats to the validity of our survey, and finally, Section 8 concludes the paper with on outlook on future
work.
2. Related Studies
Research has produced a variety of studies on Digital Twins, their features, and their application. This section
relates these studies to our investigation.
Studies on Features and Characteristics of Digital Twins
The survey presented in [376] reduces an initial corpus of 579 publications to 233 included publications to identify
a Digital Twin classification schema of eight features. The systematic survey on themes in Digital Twin research
presented in [398] investigates a corpus of 92 publications obtained by using Google Scholar as data source solely and
the artificial cut-offcriterion to consider the first 500 search results only. In this study, the authors identify 19 themes of
Digital Twin research, such as managing a physical entity, fidelity of the Digital Twin, and the twinning process itself.
Ultimately, the authors suggest a more detailed comparison of Digital Twins and publications from related fields.
Through the study, the authors identify 13 key characteristics of Digital Twins, including the nature of the twinned
system and its environment (both of which can be physical or virtual) and how its connections are realized (physical-
to-virtual and vice versa), and similar. Hence, the study predisposes interesting assumptions about Digital Twins,

such as that there is a connection between the twinned system and the Digital Twin. Another systematic literature
study considers 82 publications out of an initial corpus of 1300 publications to identify features and characteristics of
Digital Twins in the oil & gas industry [401]. For the oil & gas industry, the authors of [401] identify asset performance
management, asset risks, and support for virtual training to navigate and operate assets as the most important drivers
for Digital Twin research. These priorities reflect the complex and often remote nature of assets in the oil & gas
industry, and we expect these not to apply to general Digital Twin research in the same order. Moreover, that survey
identifies challenges to Digital Twin engineering and identify the lack of scope and focus, the lack of standardization,
and security issues as most important in their domain.
In the context of Industry 4.0, various studies touch on the topic of Digital Twins. For instance, a strategic roadmap
towards Industry 4.0 [365] identifies Digital Twins as the goal of the smart manufacturing strategy for the transition to
Industry 4.0. Here, Digital Twins are considered as a vision combining data analytics provided by intelligent enterprise
resource planning (ERP) systems and data collected from the manufacturing machines. Achieving this vision is left
as subject for future work.
Studies on Engineering Digital Twins
Various studies address the question of how Digital Twins are engineered by investigating the requirements [366],
architectures [374, 364], and the technologies [377] used with Digital Twins. Some of these focus on a specific
domain, such as the study of 43 publications on engineering Digital Twins in manufacturing presented in [374]. In
contrast, others have a wider scope, such as the survey of over 137 publications reported in [377] or the study of 52
publications on requirements for Digital Twins [366]. In [374], the authors categorize a corpus of 43 publications
on Digital Twins according to the type of contribution (case study, concept, definition, review), level of integration
(which describes the nature of the contribution as being a Digital Twin, a Digital Shadow, or a model), application
domain, and employed technologies (including AutomationML, simulation, SAP, RFID, etc.). Based on their data,
the authors conclude that research on the Digital Twin is still “in its infancy”. Our study presented in this paper may
allow to confirm or reject whether this still is the case.
In [377], the authors investigate many interesting research questions from the nature of Digital Twins to their
essential features to potentials for the evolution of the Digital Twin idea. Their process for selecting the included 137
publications is not further discussed. In another survey on requirements for Digital Twins in the context of Industry
4.0 [366], the authors analyze 52 publications obtained via Web of Science and combine the insights from their corpus
with interviews of six industry representatives. Overall, they identify real-time data handling, integration, and fidelity
as the main requirements for Digital Twins in their context.
Similar studies focus on particular aspects of Digital Twin engineering, such as the relation of Digital Twins
to product lifecycle management systems [367], the synchronization of Digital Twins with their counterparts [372],
or architectures for installed base management systems [364]. The authors of [367] investigate an initial corpus of
256 publications and reduce it to include 123 publications ultimately. Analyzing the resulting corpus, the authors
conclude that, among communication, representation, and computation, microservices are a quintessential technical
basis for Digital Twins. Moreover, they find that Digital Twins are primarily used in manufacturing and that Digital
Twins contribute not only to asset control but also to strategic business aspects. The survey on the synchronization of
Digital Twins with their counterparts [372], identifies granularity of the synchronization, the management of real-time
and historical data, proper data distribution, and operability with production resources as the main synchronization
challenges for employing Digital Twins. The authors of [364] categorize 18 selected publications on installed base
management systems architecture for manufacturing and investigate which aspects (such as communication, data
quality, security, or Digital Twin) these address. They find that only three architectures for installed base management
systems provide capabilities to serve as Digital Twins and suggest future work in this direction.
Another study investigates technologies and tools for Digital Twins [399]. The authors identify Digital Twins as
5-tuples consisting of physical entities, virtual models, data, services, and connections. Based on this assumption,
the authors discuss how the different parts of Digital Twins relate, which tasks the different parts of Digital Twins
have and which kinds of tools, such as “tools for data storage”, “tools for behavior modeling”, etc., are necessary to
realize Digital Twins. The study lists specific instances to guide practitioners in selecting suitable tools for engineering
Digital Twins for these categories. In another study of 26 publications on Digital Twins and Digital Shadows [397],
the authors investigate the areas that Digital Twins are applied to and which kinds of paradigms are employed to

achieve this [379]. According to this study, manufacturing is the most prominent domain for Digital Twins, whereas
Industry 4.0, artificial intelligence, and simulation are the most important paradigms.
A review on sustainable, intelligent manufacturing with Digital Twins discusses which equipment, systems, and
services are required to achieve this vision [368]. The paper presents a framework of sustainable intelligent manufac-
turing through Digital Twins featuring from a very abstract vantage point, which suggests that, among others, artificial
intelligence, 5G, the Internet of Things, are part of this vision. However, the review does not suggest processes, meth-
ods, or tools for engineering or operating Digital Twins.
Studies on the Application of Digital Twins
Other studies investigate the application of Digital Twins. Some of these also are focused on specific domains,
such as the study of 13 publications applying Digital Twins in construction [375], the survey of 26 publications on
the use of Digital Twins in manufacturing [369], or the survey about 23 publications on Digital Twins in smart,
interconnected factories [370]. In the latter, the authors identify a “Digital Twin lifecycle” being one of the key
enablers for smart manufacturing networks. Some studies with narrow focus include larger corpora, such as the
survey about 110 publications on Digital Twins in smart manufacturing [371]. Some of these studies are less narrow,
such as the investigation of 50 studies on Digital Twins in multiple industrial domains [373]. The latter’s findings
are that (asset) prognostics and health management is the most popular application area for industrial Digital Twins,
that modeling is essential for engineering Digital Twins, and that main challenges in Digital Twin application is
bridging the gap between cyber parts and physical parts. The survey presented in [378] investigates how Digital Twin
implementations could be evaluated and presents a grading schema for Digital Twins based on an initial corpus of 16
publications.
Studies on Literature about Digital Twins
Moreover, a few meta-studies investigate, for instance, how Digital Twins are described in the literature [380],
what the most frequently used terms for describing Digital Twin challenges in high-value manufacturing are [363], or
which definitions are used to describe Digital Twins [400]. For instance, in [363], the authors identify the 75 most
often used terms to describe Digital Twins (the top 3 being “system”, “data”, and “physical”) and analyze which topics
these belong to. Based on that analysis, the authors produce 11 clusters of Digital Twin topics (including engineering,
standards, scalability, cost & time, cyber-pyhsical system (CPS), data, user interaction, and more) and summarize
the challenges in high-value manufacturing relating to these clusters. The study presented Digital Twins [400], does
not consider terminology used to describe Digital Twins but considers explicit definitions only. Based on an ad-
hoc literature analysis, the authors collected 19 definitions and identify eight Digital Twin dimensions, including
“connectivity mode”, “CPS intelligence”, “simulation capabilities”, and “human interaction”.
Synopsis
Most of the mentioned related studies have a particular scope and depth as well as a certain level of systematic
rigor. Hence, they only address a subset of the research questions investigated within this study, consider smaller
corpora, or cannot be fully reproduced with the information presented in the corresponding publications. For the
latter, there is often a lack of information about the selected data sources, search query, and inclusion/exclusion
criteria. Thus, a larger and detailed study on Digital Twin concerns across different domains in a fully reproducible
manner is still missing, especially when it comes to the software engineering of Digital Twins.
3. Research Method
A systematic mapping study identifies publications within a research field and classifies these according to pre-
defined, structured criteria [386]. Thus, it provides an overview of the topics and contribution types for a research
domain to analyze the current status, challenges, and general progress. We have based our study on generally ap-
proved guidelines [386, 392] and practices of other mapping studies [361, 393, 394, 395]. To conduct this study, we
have used the following five-step process (based on [386]): (1) Definition of the research questions, (2) Search for
primary publications, (3) Identification of inclusion and exclusion criteria and filtering of primary sources based on
these, (4) Classification of primary studies by keywording, and (5) Extraction and aggregation of data.

Define Research
Questions Conduct Search Screening of
Papers
Keywording via
Abstracts
Data Extraction
and Mapping
Research
Scope All Papers Relevant
Papers
Classification
Scheme Systematic Map
Outcome Phases
Figure 1: Phases and outcomes of a systematic mapping study [386]
Figure 1 visualizes this process with its phases and their outcomes. In the first phase, the scope of the mapping
study is defined. This includes the research questions as well as the overall topics of the publications to be considered.
In the second phase, we collected the corpus of potentially relevant publications for our study. Afterward, in the third
phase, we analyzed the corpus according to defined criteria and reduced it to conduct our study only on thematically
relevant contributions. Based on abstract and keywords, we then derived an initial classification scheme (cf. Phase
4). Finally, the relevant publications were examined (based on full reads) and classified in Phase 5, mapping the
identified classes to the number of findings and findings cross-related as well as mapped to software engineering
phases to provide answers to the research questions described in the following.
3.1. Research Questions
We aim to identify publications on Digital Twins to investigate how they are defined, how they are engineered
and used, and to document the current state-of-the-art. Furthermore, we analyze the different goals that application
domains pursue with Digital Twins concerning their real-world counterparts and overall lifetime. This general research
interest results in the following research questions. However, not every paper is expected to provide an answer to every
question. Therefore, some information on certain facets may not be available (N/A).
1. Who uses Digital Twins for which purposes?
These questions aim to identify which application domains research in Digital Twins targets and which concerns and
challenges the publications on Digital Twins address.
RQ-1.1 Which domains employ Digital Twins? With this question we aim to understand where Digital Twins are
meant to be employed. This might shed light onto domains that are either particularly interested in Digital
Twins or particularly suited for their application.
RQ-1.2 What is the purpose of these Digital Twins? Digital Twins might be investigated for a variety of reasons.
This question aims to identify these.
2. What are the conceptual properties of Digital Twins?
A central focus of our study lies on investigating the conceptual properties of Digital Twins and their fundamen-
tal concepts. We explore, which properties and parts are associated with the twin and determine the subjects that
can be twinned. Moreover, we identify whether a Digital Twin is unique to its counterpart and how these entities
communicate. The questions concerning this topic are:
RQ-2.1 What is the real-world counterpart (i.e., the observed entity)? This question aims to understand what is
represented by Digital Twins.
RQ-2.2 How are multiple Digital Twins of the same observed entity supported? Research contributes different per-
spectives on supporting multiple Digital Twins for (aspects) of the same system. We aim to identify how the
different perspectives are supported.

RQ-2.3 When is the Digital Twin used in the lifecycle of the observed entity? Digital Twins can be used before the
twinned system exists, during its deployment, for its operations, or even after. We aim to find out how the
different lifecycle phases are supported by Digital Twins.
RQ-2.4 What stage of the observed entities lifecycle use does it represent? Orthogonal to RQ-2.3, Digital Twins
can represent different lifecycle stages of the twinned system, e.g., during system runtime a Digital Twin of the
system as-designed might be employed as no other, more up-to-date twin, is available.
RQ-2.5 How does the Digital Twin interact with its real-world counterpart? Some schools of Digital Twin thought
propose that a software system can be a Digital Twin only if there is a direct interaction from it to the twinned
system. With this question, we aim to find out whether this is a common perspective.
RQ-2.6 What (if anything) does the Digital Twin optimize? A Digital Twin might optimize the behavior of the
twinned system, itself, or nothing at all, e.g., if it is only monitoring the twinned system. We aim to find out
which optimizations are supported by Digital Twins.
RQ-2.7 What does a Digital Twin consist of? There is a scientific debate whether a Digital Twin is a model, a
software system, or whether it even includes physical parts (such as hardware for augmented reality). With this
question, we aim to find out what are common parts of Digital Twins.
RQ-2.8 Are Digital Twins decomposable? (De)composition is a quintessential software engineering for supporting
reuse. We aim to find out whether research on Digital Twins supports it as well.
3. How are Digital Twins engineered?
These questions aim to identify the means to construct Digital Twins. To this end, it explores how the different parts
and properties are realized for implementing these twins. We focus on technical details such as concrete realization,
communication, or associations with product lifecycle management. Furthermore, we investigate different tools and
technologies that have proven to be promising or essential for constructing Digital Twins. Corresponding related
research questions are:
RQ-3.1 How are Digital Twins implemented? There might be different possibilities for realizing Digital Twins. This
question aims at exploring these.
RQ-3.2 Which tools are used to engineer Digital Twins? In addition to the method of implementation, we are also
interested in which tools are used during development. With this question, we aim to investigate whether there
are certain trends concerning the tools used for the Digital Twin implementation.
RQ-3.3 Are Digital Twins developed with their own development process or are they developed together with the
observed entity? Since the term twin already suggests a strong similarity to the observed system, we leverage
this question to further investigate how this similarity affects the development process of Digital Twins.
RQ-3.4 How is quality assurance for the Digital Twin supported? With this question, we want to explore whether
Digital Twins use the same or different methods for quality assurance than the observed system.
RQ-3.5 Has the Digital Twin own requirements? Like most systems developed using engineering methods, a Digital
Twin might have requirements to fulfill. With this question, we intend to investigate whether and to which
extent such requirements are already considered during the development of Digital Twins.
4. How are Digital Twins deployed?
After their construction, Digital Twins must be deployed appropriately. This research question investigates the initial
configuration and system environment of twins. Furthermore, we consider concrete technologies for interconnectivity,
resulting in the following research questions:

RQ-4.1 Where is the Digital Twin deployed? This question aims at uncovering whether Digital Twins operate in the
cloud, on the edge, directly on the twinned system, or somewhere else.
RQ-4.2 How are Digital Twins connected to the observed entity? Digital Twins can be connected to their counterpart
to exchange information. With this question, we want to find out which technologies are used to connect them.
5. How do Digital Twins operate?
Finally, we analyze the operation of Digital Twins, including in- and output, as well as underlying data structures.
Furthermore, we investigate the possibilities of current Digital Twins to autonomously perform decision-making.
Hence we investigate the following research questions:
RQ-5.1 Does the Digital Twin feature decision-making functions? A probable use case for Digital Twins can focus
on its application to make decisions for a system. With this question, we aim to identify the different approaches
to realize these artificial decision-making processes.
RQ-5.2 To which events, inputs, or data does a Digital Twin react to? Digital Twins usually rely on information
about the physical entity’s state and user inputs. With this question, we identify how the Digital Twin gains
required information and which events trigger its actions.
RQ-5.3 Which output does it produce? A frequent use case of Digital Twins is representing the physical entity’s
state. With this question, we aim to find out how Digital Twins interact with their environment and how they
communicate to and influence their operating context.
6. How are Digital Twins evaluated?
We analyze how the included publications evaluated their contributions. For quantification, we identify and assign
classes of the different technology readiness levels. Furthermore, we explore whether the publications provide any
metrics related to the proposed Digital Twins that could be reused in future research.
RQ-6.1 Which technology readiness levels do Digital Twin evaluations employ? When constructing Digital Twins,
there can be a vast range between how the results are evaluated and to which extend they are ready for applica-
tion in an industrial context. Thus, this question aims at classifying how mature the proposed twins are.
RQ-6.2 Does the Digital Twin yield any measurable advantages? As their deployment comes with a specific goal of
what Digital Twins can achieve or improve, this question investigates potential benefits.
3.2. Search Queries and Data Sources
The search strategy is of major importance for the identification of relevant publications in order to answer the
research questions. In order to do so, formulating an appropriate search query and selecting the relevant libraries
is required. As we aim to find out who uses Digital Twins Digital Twins independently of a concrete domain or
application context, we do not restrict our search term any further. Therefore, we ultimately searched in the selected
databases for “Digital Twin”, keeping the search query simple and pragmatic. The selected databases are ACM Digital
Library, IEEE Xplore, Scopus, SpringerLink, Web of Science. We opted to omit Google Scholar due to its problems
with structured literature retrieval [382] and to ensure quality of included sources.
As we conducted a full-text search for “Digital Twin”, we omitted using other related terms, such as “digital
thread” or “digital shadow” as we expect publications contributing to Digital Twin research should at least use this
term in either related work or referenced literature. However, we cannot guarantee to not miss a small amount of
relevant publications, but argue that searching this way seems more appropriate than just searching in titles and
abstracts for keywords. Moreover, we also did not put any lower bound as year limit and included papers published
until October 2019. We extracted the results as comma-separated lists and manually merged these into a single list of
unique publications.

3.3. Screening Publications
The inclusion of a study into the classification phase of a systematic mapping study usually is decided based on
its quality and accessibility as well as on its title, abstract, and keywords [386]. To reduce the corpus and enable
reproduction of the study, we used the following explicit inclusion and exclusion criteria.
Inclusion criteria. From the initial corpus we identified the potentially relevant publications based on the following
four criteria:
1. Studies published in peer-reviewed sources namely journals, conferences, and workshops.
2. Studies are electronically accessible.
3. Studies are available in English.
4. Studies where we could deduce from title, abstract, or keywords that their main topic of study is the conception
or application of Digital Twins.
Exclusion criteria. Publications fulfilling the inclusion criteria were still excluded based on the following four criteria:
1. Studies from sources without systematic peer-review processes, such as books, magazines, and websites.
2. Short papers of less than 5 pages excluding references, such as editorials, reviews, or tool demonstration teasers.
3. Studies where Digital Twins are related work, further applications, or a broader context only.
4. Studies presenting literature reviews on Digital Twins (which are already discussed in Section 2).
While we did not limit the search to any time frame, the final corpus considered relevant does not include any
papers from before the year 2011. This is due to the term only then gaining popularity and contributions not meeting
our inclusion criteria.
"Digital Twin"
Springer
704
IEEE
137
WoS
379
ACM
33
Scopus
749
Figure 2: Data collection initially produced 1472 unique documents, out of which 356 were identified as relevant for our study.
We each analyzed and classified the first 30 (about 2%) publications of the 1472 unique publications of the corpus
to build a shared understanding of Digital Twins, the research questions, and the classification scheme. We then dis-
cussed the analysis results to align our understanding of the publications, our analyses, and the research methodology.

To filter publications based on unambiguous exclusion criteria, we evenly distributed the remaining 1442 publications.
Afterward, we determined inclusion based on whether a publication’s main contribution is towards Digital Twin re-
search by screening titles, abstracts, and keywords only. We delayed the inclusion decision to the classification phase
for publications where abstract and title screening did not suffice to determine inclusion. In this classification phase,
we then decided the inclusion based on the publications’ full text to not exclude relevant publications with sub-optimal
phrasing of abstract or title.
Eliminating 530 duplicates and 848 publications outside the scope of our study left 624 publications for review.
These publications were again distributed between the authors of this paper for a detailed review and classification.
Furthermore, we discussed the classification, exclusion, and inclusion of publications to align and refine our under-
standing whenever needed. During these discussions, we excluded additional publications and refined our shared
understanding of the classification scheme. During the reviews and discussion, further unrelated publications were
excluded. However, we did not exclude publications based on their venue or comprehensibility alone, and we also did
not perform any additional quality evaluations.
3.4. Classification Schema
To investigate Digital Twins appropriately, we have developed a corresponding classification scheme. This scheme
is inspired by [386] and adapted for the landscape of Digital Twin research. The specific facets are based on our
research on digital twins (e.g., [357, 360, 358, 359] and have been revised and/or refined iteratively while discussing
the papers among the authors as well as with digital twin experts of the “Internet of Production”4excellence cluster
and the “Christian Doppler Laboratory for Model-Integrated Smart Production”5.
After the initial screening, we analyzed the remaining 624 potentially relevant papers in the classification phase.
We have read the remaining papers completely to extract all relevant information and excluded publications that turned
out to be irrelevant for Digital Twins. We categorized the 356 remaining papers as follows.
Contribution Type Facet. Distinct papers may include different facets of contribution. Thus, inspired by [386],
we classified the publications for the type of contribution. By this means, we used five contribution types [386] to
examine the overall kind of benefit the analyzed papers provide:
•Analyses: Papers presenting investigations without constructive contributions, e.g., [76, 178, 186].
•Concepts: Papers presenting ways of reasoning about things, such as new metamodels or taxonomies, e.g., [1,
99, 207].
•Methods: Papers presenting ways of doing things, e.g., [4, 19].
•Metrics: Papers presenting ways of measuring things, e.g., [32, 264, 333].
•Tools: Papers presenting novel software tools related to implementing Digital Twins, e.g., [12, 24, 88].
We classified each publication uniquely to a contribution type.
Research Type Facet. A further important question relates to the research type of elaboration. It describes how the
findings are conducted and presented. Again inspired by [386], we further distinguished the publications by their
research type. We adjusted the originating classes to better match our corpus, e.g., by excluding philosophical facets,
as these did not occur in our study. The five resulting research types based on [386] are:
•Evaluation: Papers evaluating existing techniques, cf. e.g., [51, 81, 130].
•Experience: Papers reporting personal experiences, e.g., [24, 71, 73].
•Solution: A novel solution is presented and argued for with case studies, e.g., [1, 11].
4Internet of Production website: https://www.iop.rwth- aachen.de
5CDL-MINT Christian Doppler Laboratory website: https://cdl- mint.se.jku.at

•Validation: Papers presenting novel techniques and experimenting with them, e.g., [50, 52, 103].
•Vision: Research agendas, e.g., a vision of model-based logistics engineering presented in [5, 18].
These five facets provide an overview of the research focus of the analyzed papers. The classification was disjoint,
and we discussed contributions when in doubt.
RQ-1.1 - Digital Twin Application Domain Facet. When considering the application domains of Digital Twins,
smart manufacturing often comes to mind first. To better understand why this is the case and which domains employ
Digital Twins, we investigate RQ-1.1. To classify the different domains, we employed the Level 1 classes of the
Statistical Classification of Economic Activities in the European Community [388], which comprise all economic
areas currently considered by the European Parliament. Consequently, the application domain facet of our survey
comprises all 20 level 1 classes specified below:
•A- Agriculture, Forestry and Fishing
•B- Mining and Quarrying
•C- Manufacturing
•D- Electricity, Gas, Steam and Air Conditioning Supply
•E- Water Supply, Sewerage, Waste Management and Remediation Activities
•F- Construction
•G- Wholesale and Retail Trade; Repair of Motor Vehicles and Motorcycles
•H- Transportation and Storage
•I- Accommodation and Food Service Activities
•J- Information and Communication
•K- Financial and Insurance Activities
•L- Real Estate Activities
•M- Professional, Scientific and Technical Activities
•N- Administrative and Support Service Activities
•O- Public Administration and Defense; Compulsory Social Security
•P- Education
•Q- Human Health and Social Work Activities
•R- Arts, Entertainment and Recreation
•S- Other Service Activities
•T- Activities of Households as Employers; Undifferentiated Goods and Services Producing Activities of House-
holds for Own Use
•U- Activities of Extraterritorial Organizations and Bodies

Where the research is generic to an application domain, it is classified as “J -Information and Communication”. We
have chosen this category because Digital Twins are fundamental software systems. Contributions that do not address
a specific application domain therefore describe general information and communication systems for the application
of Digital Twins. Leveraging this classification scheme, we applied a single application domain to each publication.
RQ-1.2 - Purpose Facet. Digital Twins usually exist not only for their own sake, but to fulfill a specific purpose
concerning their physical counterpart. Concerning RQ-1.2, we aimed to understand these purposes and therefore
differentiated between the following dimensions:
•CPS Data Processing, Integration, Persistance summarizes purposes related to data processing, integration,
and persistence checking, such as knowledge collection [6], privacy enhancement [18], or data integration into
a shop floor environment [272].
•CPS Maintenance subsumes purposes related with maintaining a CPS, such as predictive maintenance [183],
fatigue testing [103], or damage evaluation [60].
•CPS Monitoring describes purposes related to collecting, analyzing, and visualizing data about the state of
a Cyber-Physical System (CPS), such as real time monitoring of building operation efficiency [100], health
monitoring [318], or process parameter monitoring [277].
•CPS Behavior Prediction summarizes purposes to predict future CPS behavior, such as fuel consumption pre-
diction [95], driver behavior prediction for crash analysis [104], or predict pulsation and velocity inside the
vessel of a human heart [245].
•CPS Behavior Optimization subsumes purposes related to optimizing a CPS’s behavior, such as path planing
for robots [24], running mode optimization of CNC tools [122], or reduce fatigue damage [296].
•CPS Validation describes purposes related with verification and validation activities, such as structural integrity
analysis [282], damage modelling for automotive low-carbon structural steel validation [98], or robot algorithm
validation [188].
•CPS Reuse describes purposes related to CPS reuse, such as design reuse [336], or reconditioning [148].
•Design Space Exploration subsumes purposes in the context of design space exploration, e.g., variation analysis
[330] or virtual prototyping [260].
•Enterprise decision making summarizes purposes that evolve around complex enterprise processes and decision
making, such as macro perspective analysis [107], or smart process planing [138].
•Teaching describes the purposes related to initial and continuing education, such as manufacturing machine
exploration [228], teaching of manufacturing cell handling [241], or robot manipulation training [343].
•Visualization &Representation summarizes purposes directly related to visualizing a physical counterpart, such
as represent a production site in a virtual environment [7], or visualizing object properties in augmented /mixed
reality [169].
Since Digital Twins might have more than one purpose, e.g., the health monitoring the approach presented in [318]
relates to monitoring as well as maintenance, the selection mentioned above is not disjoint.
RQ-2.1 - Counterpart Facet. A Digital Twin is little without its counterpart. To better understand what it is that is
“twinned”, we classify our corpus’s publications according to the various counterparts described. Our classification
schema of Digital Twin counterparts comprises:
•Biological Beings, such as factory employees [215], fishes [266], or sports players [288].
•Individual Systems, such as automated cars [219], gas turbines [108], or manufacturing machines [67].

•Processes, e.g., business processes [77], medical processes [273], or recycling processes [239].
•Products, such as reinforced plastics [330], sunroof ring frames [210], or wearable masks [40].
•Systems of Systems, such as complete factories [1], oil wells [131], or railway systems [207].
•Other counterparts, e.g., arbitrary physical bodies [162] or unspecified manufacturing resources [291].
To distinguish whether a publication reports on an individual system, product, or system of systems, we discussed
these publications and together decided about their specific focus regarding the Digital Twins’ counterpart(s). We
also encountered some publications that report on Digital Twins for more than one counterpart or a combination of
counterparts, such as processes and related systems of systems [85] or products and related systems [348]. Such
contributions add to multiple counterpart facets accordingly.
RQ-2.3 - Digital Twin Lifetime Facet. Digital Twins can reflect, monitor, and support all phases of the physical
entity. In some application scenarios building the physical entity may be very time- or cost-intensive. In these cases,
a Digital Twin can be applied during the design phase of the physical entity to communicate design decisions or to
simulate multiple designs. At runtime, a Digital Twin may monitor the physical entity’s actions and suggest further
steps, e.g., for minimizing raw material waste or energy consumption. Consequently, we introduce a Digital Twin
lifetime facet that distinguishes:
•Design-time, to characterize Digital Twins that are employed during the the design phase of the physical entity,
e.g., to evaluate different product variants [219].
•Runtime to characterize Digital Twins that are employed while the physical entity is already operating. These
Digital Twins may predict future behavior or control and optimize the physical entities’ next steps [19].
Whether a publication describes a Digital Twin that is used at design- or runtime was usually explicit. In cases
where a reviewer could not classify the described Digital Twin, we also discussed the respective paper among the au-
thors. We also encountered publications that report on Digital Twins for more than one lifetime or even Digital Twins
that were transferred from design- to runtime [20]. Such contributions add to multiple lifetime facets accordingly. In
contrast to RQ-2.4 this question focuses on the lifetime where the Digital Twin is applied and not on the lifecycle step
for which the Digital Twin reflects the physical system.
RQ-2.4 - Digital Twin Lifecycle Facet. Digital Twins application scenarios can be distinguished along the system’s
lifecycle specified in the ISO/IEC 15288 system lifecycle [389]. This norm distinguishes roughly three product
lifecycle phases: the design phase including the conceptualization and modeling of the system, the manufacturing
phase where the system is brought into existence, and the operation phase where the system operates and fulfills its
intended purpose [405].
Depending on the twinned system’s lifecycle phase, Digital Twins serve different purposes [403]. For instance,
Digital Twins might represent a system as it is designed for design-space exploration of that system before it is
constructed or Digital Twins focus on the running system as it is in use to represent its current state and serve for the
maintenance prediction. Consequently, we introduce a Digital Twin lifecycle facet that distinguishes:
•As-Designed, to describe Digital Twins that represent the physical counterpart during its design phase. These
Digital Twins are e.g., useful for optimizing the production process [16].
•As-Manufactured also integrates data that characterizes the manufacturing process of the physical counter-
part. Thus, it may include runtime data that provides insights for maintenance [314] of predicting material
fatigue [41].
•As-Operated describes Digital Twins that represent the usage and operation of the physical counterparts, e.g.,
for supervising and optimizing [56, 53] or for predicting future behavior [29, 31].

To distinguish whether a publication describes a Digital Twin as designed, manufactured, or used, we discussed
the categories a-priori in detail. If case a reader could not classify the described Digital Twin we also discussed the
respective paper among the authors. We also encountered publications that report on Digital Twins for more than one
lifecycle [353, 402]. Such publications contribute to multiple lifecycle facets accordingly. In contrast to RQ-2.3 this
question focuses on the lifecycle stage the Digital Twin represents and on the time when the Digital Twin is used.
RQ-2.5 - Interaction Facet. Literature exhibits various understandings of Digital Twins from precise models used at
system-design time [5, 8, 16] that are used to prescribe a system to be to software systems twinning another system
and directly manipulating its behavior [176, 213, 253]. With this facet, we, thus, investigate whether Digital Twins
tend to support direct interaction with the observed system. To this end, we distinguish two cases:
•Direct Interaction comprises Digital Twins that are directly connected to their counterpart through various
communication measures, such as message busses, networks, or Internet technology.
•No Direct Interaction describes Digital Twins in which interaction either is indirect, e.g., by informing a human
operator to execute system manipulation or there is no interaction at all, such as where the Digital Twin is
interpreted as a dataset recorded from the twinned system.
Each publication contributes to exactly one of these classes.
RQ-2.6 - Optimization Facet. Many Digital Twins seem to optimize either themselves, their real-world counterpart,
or both. However, not all Digital Twins strive to optimize. With this facet, we aim to investigate whether the optimiza-
tion is considered in the development of Digital Twins and whether the Digital Twin optimizes itself, or its real-world
counterpart. In our classification schema we, therefore, distinguished as follows:
•Digital Twin Optimization incorporates Digital Twins which optimize themselves without aiming to influence
their twinned entity directly with this optimization. In [329] for example, the Digital Twin collects data from a
health monitoring system to optimize its structural health model of the real-world counterpart.
•Counterpart Optimization subsumes Digital Twins aiming to only optimizing their real-world counterpart. For
instance, in [137] a Digital Twin is initialized with real tech parameters and then used to optimize the real-world
counterpart, without updating the Digital Twins simulation model.
•Digital Twin &Counterpart Optimization subsumes Digital Twins that not only optimize their counterpart
alone, but also use information from their counterpart to optimize themselves. In [250], for example, the Digital
Twin is used for state estimation in non-linear electro-mechanical systems and optimizes not only the electro-
mechanical system but also the Digital Twin itself.
Of course, we also encountered multiple publications where counterpart or self-optimization was explicitly not
the purpose of the Digital Twin concept, as they focused on visualization [59] or monitoring [90] alone. Moreover,
some authors decided not to mention the possibility that the Digital Twin performs such optimization at all as, e.g.,
in [198]. However, as we cannot differentiate between publications where optimization is thought of as irrelevant for
the purpose of Digital Twins and publications where the optimization was just not relevant for the published aspect,
we decided to subsume these papers in an additional category.
RQ-2.7 - Digital Twin Parts. A Digital Twin usually is a logical unit, which is composed of different parts. For
example, we can distinguish between data, services, a virtual models and physical entities [406]. To understand how
Digital Twins are developed and which components are necessary for software to become a Digital Twin, we collected
information about Digital Twin parts.
•Data describes live data about the physical entity [11], historical data about the physical entity [26], or data from
other data sources that provide contextual information about the application scenario of the physical entity [95].
•Hardware Components captures Digital Twins, which also contain physical components, such as equipment [350].

•Models describes software artifacts that are classified as models according to Stachowiak [404], i.e., they have
a purpose, perform abstraction and have a physical entity. Frequent examples were simulations [3] and CAD
models [33]. We further classified models according to the aspects they describe. Thus we identified
–structure of the physical twin, e.g., inner components,
–behavior of the physical twin, e.g., interaction with its environment,
–appearance of the physical twin, e.g., material information, and dimensions,
–constraints of the physical twin, e.g., physical laws.
•Software Components characterizes custom [152] and external [272] software services that are described as part
of the Digital Twin and cannot be classified as models.
Where publications ambiguously define which components they consider as part of the Digital Twin, we decided
in favor of including these components as parts of the Digital Twin. In cases where no categorization was possible,
we discussed the papers between us until we could reach an agreement. Many Digital Twins consist of multiple
components, which can be assigned to different facets. These publications contribute to several facets.
RQ-3.1 - Implementation Technique Facet Digital twins are created using various implementation techniques de-
pending on their purpose, lifetime, and more. With RQ-3.1, we analyze the different facets in which they are realized,
including various models, programming languages, or simulations. Thus, we identified the following classes.
•CAD/3D Models describing the geometric representation of a physical component [1].
•Data and Databases covering collecting and analyzing operation data [328] as well as different data formats
[352].
•General Purpose Languages, such as Java [144], C++ [76], and Matlab [44].
•Mathematical/Physical Models, such as finite element [128] and multi-physics models [309].
•Model-Driven Engineering (MDE), such as UML or SysML models [289], language workbenches [87], and
AutomationML [120].
•Simulation and Analysis, such as Simulink [121], Verosim [188], or AnyLogic [299].
The selection is not disjoint as contributions may use particular techniques as a foundation for implementing
Digital Twins. For instance, [128] combines geometric data with physical models, which contribute to the construction
of the system.
RQ-3.2 - Tooling Facet. Engineering Digital Twins of different counterparts and for different purposes efficiently de-
mands corresponding tool support. With this facet, we aim to uncover which kind of tools are used in the development
and operations of Digital Twins. Our classification schema regarding tools applied to the engineering or operations of
Digital Twins comprise tools that were mentioned 7 or more times by the publications of our corpus:
•Artificial Intelligence Software, such as Apache MXNet [95], the IBM Watson software development kit [218],
or TensorFlow [256].
•Communication Software, including ROS [236], OPC UA [148].
•Computer-Aided Design (CAD) and 3D modeling, such as SolidWorks [7], Siemens NX [45], or Autodesk
Revit [129].
•MDD Software, such as AutomationML [325], Modelica [159], or SysML [176].
•Data Management Software, such as Apache SOLR [344], SQL databases [100], or SAP HANA [214].

•Process Management Software, including ChemSiemens10 Tecnomatix [160], UniSim Design [53], or in-house
developmed solutions [355].
•Product Lifecycle Management Software, such as Siemens PLM [45].
•Programming Languages, including Python [124], Java [144], and others.
•Simulation Software, such as Abaqus [279], Gazebo [57], the MAYA simulation framework [25], or Verosim [50].
•Visualization Software, such as Unity [104], OpenCV [36], or APIs for augmented reality devices [60].
•Other Software, including various programming languages [294], specific self-developed toolsets [316], APIs
for communication [3], or website development tools [248].
Some contributions use the same software, such as MDE software or the various programming languages for
multiple purposes and employ a wide variety of software to engineer or operate Digital Twins. Consequently, our
classification schema for the tooling facet allows for more entries than the number of publications included in the
corpus.
RQ-3.3 - Digital Twin Engineering Process Facet. Digital Twins can be developed before the twinned system,
together with it, or after it. Developing the Digital Twin before the twinned system can facilitate design space explo-
ration by frontloading of the twinned system as the Digital Twin might be used as substitute to explore properties of
the twinned system at higher levels of abstraction. Developing the Digital Twin together with the observed system
enables optimizing their interaction by, e.g., joint design-space exploration of both, the Digital Twin and the twinned
system. Developing the Digital Twin after the twinned system enables adding advanced functionality to existing sys-
tems and makes these accessible for analyses typically related to Digital Twins, such as behavior prediction. To better
understand whether Digital Twins are generally developed together with their counterparts or detached from them,
we have grouped the publications of our study accordingly. Our classification scheme of the development process of
Digital Twins, thus, includes the following categories:
•Joint Engineering incorporates publications where the engineering and the evolution of Digital Twins are inter-
twined with the engineering of their counterparts.
•Subsequent Engineering incorporates publications where the engineering of the Digital Twin i.e., the process
itself, succeeds the engineering of their counterpart. Here, the counterpart to be twinned, or previous versions
of it, already exists, and that this information about the counterpart can be leveraged for its twinning.
•Explorative Engineering means that the development of the Digital Twin frontloads the development of their
counterpart. Here, Digital Twins are developed from scratch without including information about their existing
counterpart. Instead, they can be used to explore the properties of their counterpart.
We encountered some publications reporting that Digital Twins can be engineered in both fashions, either in a
joint process together with their counterpart or in their own process independent of the engineering process of their
counterpart. Of those publications that reported that a Digital Twin has its own engineering process, only some
clarified if the Digital Twin was to be developed before or after the system. If a publication did not report on the
engineering process of Digital Twins, or if that information was not derivable from the purpose of Digital Twins, then
we regarded corresponding information as not available.
RQ-3.4 - Quality Assurance Facet. The Digital Twin can be understood as a precise design-time model, e.g., used for
design space exploration of the system under development, prediction of its future behavior, and general frontloading,
or as a software system observing another system at the other system’s runtime. Consequently, different means of
quality assurance need to be employed to produce, operate, and maintain a high-quality Digital Twin. Concerning
RQ-3.4, we aim to understand the state of quality assurance for Digital Twins. Our classification schema of Digital
Twin quality assurance comprises the following dimensions:

•Consistency monitoring, e.g., by monitoring the differences between Digital Twin predictions and data obtained
from the twinned system at its runtime [136, 76].
•Simulation, used at system design-time [5, 147].
•Testing other than simulation [244], also employed at design time.
•Other verification, such as model-checking, also applied at design time [27, 247].
As the boundaries between simulation, testing, and other verification are not used strictly through our corpus, we
followed the terminology employed by the respective authors.
RQ-3.5 - Requirements Facet. Like any engineered system, Digital Twins are likely to have certain requirements
to meet. So we addressed the question of whether Digital Twins have requirements of their own and considered the
requirements that are typically placed on Digital Twins. To this end, we identified the following facets based on the
publications in our corpus:
•Real-time capability requires from the Digital Twin that it provides its services or responses within a specified
time constraint as e.g., required in [5, 214].
•Digital Twin reaction matches real-world behavior subsumes various verification and validation requirements
aiming to ensure that the behavior of the observed system meets the reaction of the Digital Twin [336].
•Reusability requires the Digital Twin to be reuseable in different contexts or for closely related systems [336,
231].
As no other facet was mentioned more than once in the publications of our corpus, we decided not to further
investigate these possible facets.
RQ-4.1 Digital Twin Host Facet. A Digital Twin may operate in some context, such as in some cloud, on an edge
device, or directly on the twinned system. To understand where Digital Twins operate we classify the publications in
our corpus according to the various hosts described:
•Cloud incorporates publications where the Digital Twin is deployed in some cloud, either named or none-
specific [19].
•Fog means that the Digital Twin is deployed on another device than its counterpart but still resides in a local
network [294].
•Edge when the Digital Twin is deployed on the same device as its counterpart [44].
This facet also includes data regarding Digital Twins host of unspecified provenance with respect to the hosting
alternatives above but on the kind of system hosting the Digital Twin:
•Data Management System if the host of the Digital Twin is a database or some other data-centric applica-
tion [152].
•Simulator incorporates Digital Twins that are employed as part of a simulation, such as 3D simulation mod-
els [148].
•Virtual Reality for Digital Twins deployed in a virtual reality [304].
Reported findings describe the hosts’ location, such as in the cloud, on an edge device, or directly on the twinned
system; or describe the host’s kind, including data management systems, simulations, and virtual reality. As such, we
encountered publications with multiple findings in this facet, e.g., if the Digital Twin was deployed on some database
in the cloud. But also findings with multiple reports of the host’s location where possible, in cases where Digital Twin
could be deployed either in the clouded or some local system.

RQ-4.2 Digital Twin Counterpart Connection Facet. As Digital Twins seem to be connected with their real world
counterpart, we investigated in the context of RQ-4.2 which technologies are used to connect the Digital Twin with
its real world counterpart. To understand these connections we classify the used technologies as follows. First, we
identified a set of hardware or technology related connections that were often described in our corpus:
•Local Networks connect Digital Twins by establishing BUS systems [5], Ethernet [16] or WiFi [122].
•Short distance wireless communication subsumes non-LAN short distance communications, such as RFID [10]
or Bluetooth [216].
•Data connect Digital Twins with their counterparts by its data as for example described in [83] which use a
database.
•Server/Cloud/Proxy in this connection type the Digital Twin is connected with the real world counterpart based
on a remote server such as a cloud application [292].
On the other hand, we also found connection descriptions that solely focus on the used communication protocol:
•Industrial protocols, such as MTConnect [252] or OPC UA [235].
•Internet protocols, such as TCP/IP [4], UDP [177], or HTTP [89].
•IoT Protocols, such as [90].
Of course, not all publications of our corpus strictly differentiate between the communication technology and the
protocol. Thus, multiple selections were possible.
RQ-5.1 - Decision Making Facet To influence the observed entity, a Digital Twin should be able to make decisions
based on its counterpart’s current state and condition. Thus, we distinguish in RQ-5.1 between the following classes
of decision-making functions to better understand the nature of these decisions.
•Data Mining, such as big data methods [146], or data cleansing techniques [53].
•Machine Learning, such as artificial neural networks [89], or deep learning [95].
•Reasoning techniques, further distinguished into
–Case-based Reasoning, e.g., [151].
–Symbolic Reasoning, such as [50].
–Stochastic Reasoning, for instance [22].
–Other Numeric Reasoning, classifying remaining methods such as [136].
•Simulation, such as finite element analysis [8], virtual testbeds [72], or rigid body dynamics [74].
As some contributions employ multiple decision-making functions for their Digital Twins or use combinations of
different techniques, these classes are generally not disjoint. For instance, in [97], the Digital Twin uses a data mining
to process gathered data and furthermore simulates values that could not be obtained from the physical counterpart.
If a publication did not explicitly specify any corresponding functions, we classified the Digital Twin to offer no
decision-making functionality.
RQ-5.2 - Digital Twin Input and Events Facet. Digital Twins rely on specifications that define how the Digital Twin
should behave in different situations, e.g., when context changes occur, when external inputs are given, or when the
equivalent acts in a certain way. Inputs for Digital Twins have different sources, as humans that explicitly control the
Digital Twins’ actions or models that specify the Digital Twin’s behavior. Digital Twins also react to events that occur
in their operating context or in the physical entity that they represent. We introduce the input facet to classify the input
data that the Digital Twin relies on upon and differentiate between different input data sources as follows:

•Machine Data specifies all data captured by sensors [6] or emitted by machines [15].
•Models and Simulations includes data that is provided by simulations [25] or specified through models [106].
•User Specifications characterizes all Digital Twins that a configured via direct user input, e.g., through a user
interface [16] or motion capturing [179].
For the papers where a description was given, this facet could generally be identified unambiguously. However, in
some cases, it was not reported how the DT was configured. We did not assign a facet to these papers and classified
these as non-assignable.
RQ-5.3 - Digital Twin Output Facet. Digital twins that fulfill some kind of purpose other than modeling their
counterpart often interact with their environment through inputs and outputs. Outputs can take different forms and
have various intentions. To better understand what kind of outputs Digital Twins can produce, we have grouped the
contributions according to the following schema:
•Observations, if the Digital Twin represents the current state of the twinned system, i.e., monitoring data.
•Prescriptions, for Digital Twins emitting instructions send by the digital twin about changes that should be
applied, incorporating parameter configuration changes, detailed control commands, and planning data.
•Predictions, when the Digital Twin produces predictions or estimations about the system behavior, failures, and
life expectancy, i.e., what can be.
•Other Data, when Digital Twins output data but the nature of this data is not further specified.
•Visualization &3D Models, when the output of Digital Twins are changes to a UI or 3D models for visualization
of their counterpart.
As these categories overlap, publications may be assigned multiple times.
Generally, contributions reported that the output of Digital Twins can represent the current state of the system,
contain predictions or estimations, or describe changes and modifications that should be applied. In some cases, the
output of Digital Twins was described as less detailed, only stating that Digital Twins emit some kind of analysis result
or some kind of data. Also, some contributions reported visualization of Digital Twins through a UI or 3D models as
an output of Digital Twins. If a publication did not explicitly report outputs for Digital Twins, then we regarded the
corresponding information is not available.
RQ-6.1 - Digital Twin Evaluation Maturity Facet. Digital Twins are expected to improve our understanding and
use of systems. To understand how mature the research results within our corpus are, we classify they contributions
according to the technology readiness level (TRL) [381] of their evaluations or case studies. As publications rarely can
provide fully detailed evaluations due to, e.g., page limitations or confidentiality considerations, a precise estimation
of evaluation maturity is rarely possible as well. Hence, we classify evaluation maturity as follows:
•Proof of Concept (TRL 1-3), includes evaluations in which at least basic principles of the research can be
observed and at most an experimental proof of concept is reported.
•Technology (TRL 4-6), includes evaluations where technology is at least evaluated in a laboratory context and
at most in a relevant environment.
•System (TRL 7-9), includes evaluations in which at least a system prototype is demonstrated in an operation
environment.
Each publication was assigned a single TRL.

4. Vertical Analysis
During the vertical analysis, we aim to provide quantitative results for all research questions, where an answer for
this research question without considering other research questions was possible. In the following subsections, we
present the results of this analysis for our questions. As not all research questions could be answered unambiguously
or with sufficient significance based on the publications in our corpus, observations and potential insights to these
research questions are briefly revisited in the discussion.
4.1. Contribution Type
The papers in our corpus address various topics and represent different types of contributions. Accordingly, we
classify the publications in our study by their type of contribution [386]. The classified contribution types are disjoint,
and thus, each publication was classified to exactly one contribution type. Publications suitable to more than one
contribution type were classified to the most suitable.
Overall, the majority of contributions address methods (223, 62.64 %) or concepts (98, 27.53 %), whereas con-
tributions mainly addressing tools (19, 5.34 %), analyses (13, 3.65 %), or metrics (3, 0.84 %) are considered far less
often. The distribution of contribution types is shown in Figure 3.
0
50
100
150
200
250
Analyses Conc epts Methods Metrics Tools
Figure 3: The contribution types of included publications
The major focus on concepts and methods and the lack of analyses may be a symptom of the still young field
of research on Digital Twins. Research and industry have not advanced enough, and exhaustive solutions employing
Digital Twins exist whose effect could be analyzed in detail. Surprising is the lack of tooling, which would be needed
to employ Digital Twin solutions. While some contributions on metrics for Digital Twins exist, most of these are
employed in a broader research concern, with only a few publications focusing on metrics in particular. Focusing
research on tooling for Digital Twins realizing the presented concepts and methods could move research on Digital
Twins forward in the future.
4.2. Research Type
Besides the contribution type, we also analyzed the publications research type and classified the publications
according to the schema presented in Section 3.4. Again, each paper was classified to exactly one research type, that
is to the most suitable research type if a paper was eligible for more than one category.
We found that solution proposals are the most common research type in Digital Twin research, which make up 233
(65.45 %) of all publications in our corpus. Other research types are far less common. Out of the 356 publications in
our corpus, vision papers only contribute 59 (16.57 %), experience reports only 37 (10.39 %), evaluation reports only
16 (4.49 %), and validation papers only 11 (3.09 %) publications to the overall corpus. The distribution of research
types is shown in Figure 4. It is similar to the distribution of the classification of contribution types in the sense that
two-thirds of the publications also contribute to a single class.

0
50
100
150
200
250
Solution Vision Experience Validation Evaluation
Figure 4: The research types of included publications
A similar distribution of research and contribution types might reflect constructive research on Digital Twins,
which is reinforced in particular by the large number of publications describing solutions. Despite the many papers
describing methods or concepts, there are only a few publications that focus primarily on validation. One of the
upcoming goals in Digital Twin research should be to validate and evaluate existing approaches and solutions in
detail.
4.3. RQ-1.1 - Digital Twin Application Domains
While manufacturing and Industry 4.0 might come to mind first, when considering the application domains of
Digital Twins, there are plenty of other domains Digital Twin research is applied to. By classifying the studies
according to the schema presented in the corresponding facet, we identified eight domains that research on Digital
Twins focuses on. The domains, together with the number of publications addressing these, are illustrated in Figure 5.
All 356 (100 %) publications of our corpus relate to a specific domain or contribute to generic research on Digital
Twins. The large majority of research on Digital Twins focuses on (C) manufacturing (252, 70.79 %) and on (J) Infor-
mation and Communication (47, 13.2 %). The remaining (16.01 %) publications are almost equally split into further
domains and generic Digital Twin research. Of the other domains, (D) energy (17, 4.77 %) and (F) construction are
(12, 3.37 %) are more often addressed than (Q) human health (8, 2.25 %), (B) mining (9, 2.53 %), or (H) transporta-
tion and storage (6, 1.68 %). We also found publications conducting research on Digital Twins for (P) education (4,
1.12 %) and (A) agriculture (1, 0.28 %). For the other twelve classes of economic areas, we did not find research on
Digital Twins.
This especially holds for areas where digitalization and automation might not be as advanced as in manufacturing,
such as (E) water supply and sewage, (G) wholesale, (I) accommodation and food service activities, (L) real estate
activities, or (R) arts and entertainment. This maybe be because they do not need a Digital Twin, use different ter-
minology, or we have simply missed publications as they are not indexed by our databases or the people involved do
not publish their results in scientific literature. Many of the areas not addressed by publications in our corpus also are
areas in which human actions and decisions are central to creation of added value, such as wholesale, accommoda-
tion, financial activities, real estate activities, public administration, arts and entertainment. The lack of publications
addressing these might be a symptom of properly capturing Digital Twins of human actors, which is in line with the
small number of Digital Twins of beings as observed regarding RQ-2.1.
4.4. RQ-1.2 - Digital Twin Purpose
One question any application of new concepts, methods, or tools are faced with is their purpose. Thus, we
analyzed which purposes Digital Twins fulfill. Based on the classification scheme presented in the corresponding
facet, we aimed to answer RQ-2 and concluded how many publications mention a purpose and how these purposes
distribute over the purpose facets. Of our corpus (356, 100 %), 347 (97.47 %) publications make the purpose of the

Figure 5: Application domains of Digital Twins
Digital Twin explicit. Consequently, a majority mention the purpose of the Digital Twin explicitly, which distribute
over these purposes as illustrated in Figure 6. Moreover, we found out that most Digital Twins have more than one
purpose (196, 55.06 %).
CPS Behavior
Optimization (134)
CPS Monitoring
(105)
CPS Validation
(79)
CPS Behavior
Prediction (65)
CPS Data
Processing /
Integration /
Persistence (44)
Visualization &
Representation (43)
Design Space
Exploration (30)
CPS
Maintenance (24)
CPS ReuseTeachingEnterprise Decision Making
(12) (3) (3)
Figure 6: Purposes of Digital Twins
The four biggest clusters we identified are CPS Behavior Optimization (134, 37.64 %), CPS Monitoring (105,
29.49 %), CPS Validation (79, 22.19 %), and CPS Behavior Prediction (65, 18.26 %). With almost equal numbers
CPS Data Processing, Integration, Persistence (44, 12.36 %), Visualization & Representation (43, 12.08 %). Of the
other purposes, Design Space Exploration (30, 8.43 %) is mentioned more often than CPS Maintenance (24, 6.74 %),
which is still mentioned twice as often as Enterprise Decision Making (12, 3.37 %). Almost not mentioned are Digital
Twins for CPS Reuse (3, 0.84 %) and Teaching (3, 0.84 %).
These numbers show that most Digital Twins are either used for behavior prediction or optimization, which com-
bined make a total of 199 (55.9 %) publications. Another huge application area seems to be CPS Monitoring and
Visualization, which together make 148 (41.57 %) of the corpus. This shows that Digital Twins used today seem most
likely to optimize, monitor, or visualize their physical counterpart.

4.5. RQ-2.1 - Digital Twin Counterparts
To uncover which kinds of counterparts Digital Twins are used with, we analyzed the publications in our corpus
for this aspect. We found that mostly all publications made explicit what the counterpart of the presented Digital Twin
concepts is. Out of the 356 (100 %) publications, a total of 350 (98.31 %) publications make the counterparts of their
Digital Twins explicit. Overall, 31 (8.99 %) publications present research in which the Digital Twin supports more
than one counterpart, for instance, when Digital Twins for the production system and the produced product [247] or
a production process on an individual system [138] are considered. Overall, 384 counterparts are reported by the
publications included in our survey, as illustrated in Figure 7.
Individual
System (210)
Systems-of-
Systems (104)
Product (19)
Process (17)
Other (16)
Biological
Being (18)
Figure 7: The predominant counterparts of Digital Twins are individual systems and systems of systems.
The predominant counterparts of Digital Twins are individual systems (210, 54.69 %) and systems of systems
(104, 27.08 %), which make a total of 314 (81.77 %) of the Digital Twin counterparts identified in our corpus. Digital
Twins for beings, processes, products, and other counterparts are significantly less common and make only a total of
70 (18.23 %) publications to the counterparts of Digital Twins.
Finding most Digital Twins relating to counterparts that are individual systems and systems-of-systems is not
unexpected. However, the latter entails questions regarding the communication and (de)composition of Digital Twins
that are further investigated in the context of RQ-2.8. Especially when individual systems can flexibly enter or
leave system-of-systems structures, such as within smart manufacturing, automated convoys, or distributed Internet
of Things systems, the interfaces of Digital Twins, their interactions, and means for flexible composition need to be
understood.
The low number of publications contributing Digital Twins of products is unexpected as the smart product of
lot-size 1 is one of the driving visions of Industry 4.0, and Industry 4.0 is one of the main disciplines driving research
on Digital Twins. However, in line with the increasing number of publications on Digital Twins and the finding that
the digital representation of assets still is one of the prime research topics, at least in modeling for Industry 4.0 [361]
suggests that before the product can be twinned, first the assets and the processes relating to its production must be
considered. Yet, there also is a small number of publications on Digital Twins relating to processes that are not tied to
one or more systems directly.
The overwhelming focus on contributions to engineering Digital Twins for systems consequently indicates that
research still is in a very early stage of understanding the systematic engineering of Digital Twins, means to reuse parts
of Digital Twins for Digital Twins of different counterparts, and suggests that established reuse techniques from soft-
ware engineering, such as encapsulation, type-based substitution, product lines are not as common for Digital Twins
yet. We assume the latter is due to the different perspectives on Digital Twins as (design-time) models, (run-time)
systems, or something in-between and the heterogeneous implementation techniques that are employed accordingly.
Research on heterogeneous modeling [391] and software language engineering [384, 362] can contribute to closing
the gaps between the different technological spaces [390] and applying established reuse techniques to Digital Twins
systematically.

4.6. RQ-2.4 - Digital Twin Lifecycle
To understand how Digital Twins are applied to the different lifecycle phases of their counterparts, we classified
the publications accordingly. As designed Digital Twins consider the ideal design of their physical entities, thus not
taking into account minor derivations that may occur during the construction of the counterpart. As-manufactured
Digital Twins do consider these derivations, while as-operated Digital Twins also include usage data which may
inflict the physical counterpart’s behavior or appearance. Figure 8 shows the distribution of described Digital Twins
as a Venn diagram. In our corpus of 356 (100 %), all but 17 (4.77 %) publications made explicit which lifecycle the
presented Digital Twin represents. A total of 60 (16.85 %) publications describe Digital Twins for more than one
lifecycle, for instance, when Digital Twins for the design and manufacturing were combined [210]. Overall, 29 (8.15
%) publications also presented Digital Twins that were used across all lifecycle phases.
As Designed
52
As
Manufactured
15
As Operated
212
618
7
29
Figure 8: Lifecycle phases of Digital Twins
Most Digital Twins (266, 74.72 %) represented the operation lifecycle phase of the physical entity. This might
indicate that Digital Twins are often employed for simulating the physical entity’s behavior, e.g., if it is not built
yet, or to test new application scenarios before they are realized. Also, when Digital Twins fulfill informative and
representing requirements, they also integrate runtime sensor data to mirror the entity’s state. It should also be pointed
out that only 57 (16.01 %) publications consider the manufacturing of the physical counterpart, and only 15 (4.21
%) publications focus the manufacturing exclusively. Many of the Digital Twins described in the literature represent
CPS with a long lifecycle, e.g., production machines. In such systems, sensors are often retrofitted, and Digital
Twins are developed while the machine is already operating (brownfield Digital Twin development), to represent
their counterparts as-operated, which explains the high number of publications reporting on Digital Twins for this
lifecycle phase. However, representing the design of future systems by a Digital Twin can be beneficial to evaluate,
e.g., different variants before the system is realized. Therefore an increase in design-time Digital Twins is expected in
the future.
Another possible explanation for the derivation between frequency of as-designed and as-operated could be that
the physical twin is designed once, but then multiple instances conforming to this design are produced. Thus, leading
to only one as-designed Digital Twin but multiple as-operated Digital Twins where the as-operated Digital Twins
represent different instances conforming to the same design. To enable co-evolution of physical objects and their
Digital Twins, future research also should be conducted on the transformation from as-designed Digital Twins to
as-manufactured Digital Twins and to as-operated Digital Twins.
Combinations of the different lifecycle phases are generally not researched thoroughly. Especially, transitions
between Digital Twins (i) as-designed and as-manufactured; (ii) and as-manufactured and as-operated; and (iii) as-

operated back to as-designed yield promising potentials for a pervasive model-driven DevOps of Digital Twins that
saves development time in the future.
4.7. RQ-2.6 - Digital Twin Optimization
With RQ-2.6, we investigate whether Digital Twins are used for optimization and whether the Digital Twins
optimize their counterpart, themselves, or both. To this end, we classified Digital Twin optimizations mentioned in
the publications of our corpus according to the classification schema presented in Section 3.4.
Overall, we found that out of 356 (100 %) publications, 193 (54.21 %) publications explicitly perform optimiza-
tion, whereas 163 (45.79 %) publications do not mention or consider Digital Twin-based optimizations. Of the 193
publications making the Digital Twin optimization explicit, only 6 (1.68 %) publications present Digital Twins that
only optimize the Digital Twin. Most publications present Digital Twins that optimize their counterpart (138, 38.76
%) or optimize both (49, 13.76 %).
Observed System
138
Twin Only
6
49
Both
Figure 9: Regarding optimization, Digital Twins largely focus on the observed system.
These numbers show that most of the publications on Digital Twins of our corpus describe Digital Twins that
optimize their twinned counterpart. Moreover, there is a clear trend to only optimize the virtual counterpart without
adapting or optimizing the Digital Twin itself. As there is also a smaller proportion of papers that either describe the
optimization of the Digital Twin alone or the Digital Twin and the observed system, it can be followed that further
research on methods to optimize the Digital Twin parallel to the observed system might be required. In this context
self-optimizing Digital Twins might benefit from models at runtime or self adaptive sources [6].
4.8. RQ-2.7 - Digital Twin Parts
With research question RQ-2.7, we aim to find out the essential components that are part of Digital Twins. We
found that out of 356 (100 %) all but 33 (9.27 %) publications made explicit what they consider part of the reported
Digital Twin. Of the publications making the parts explicit, a total of 144 (40.45 %) publications describe Digital
Twins composed of more than one of the facets of parts, while none of the publications report on a Digital Twin that
is composed of elements from all facets.
Models (274) Data live/historic (129)
Physical
Components
(49)
Software
Components
(39)
Figure 10: Distribution of Digital Twin constituents with models as the predominant factor.
Hardware components are also named as components of the Digital Twin. Since the Digital Twin is a digital
object, the number of 49 (13.76 %) papers that also name hardware components seems surprisingly high. A possible

explanation are cyber-physical components whose functionality is realized by a combination of hardware and software
components. This makes the boundary blurry and since software and hardware are delivered together, it is less
recognizable for the user. In many applications, a combination of a Digital Twin together with a physical model is
utilized to provide the user with haptic feedback. This is especially the case for Digital Twins used in training, e.g.,
for medical professionals [280].
Most publications (274, 76.97 %) mention models as parts of the Digital Twin. The terms “model” and “Digital
Twin” are even used synonymously [2]. These models mainly describe the physical counterpart’s constraints (133,
37.36 %) and its appearance (94, 26.4 %). Only a few publications (18, 5.06 %) apply models for describing data
structures and only 30 (8.43 %) publications describe Digital Twin behavior through models. The most frequently
reported model types were simulations, physical models, and geometric models. This is consistent with the fact that
many described Digital Twins come from the engineering field, where these types of models are highly prevalent [321].
Digital Twins often offer services, e.g., to evaluate system states [214] or to influence system behavior [215].
Thus, it is not surprising that (39, 10.95 %) publications mention software as a part of the Digital Twin.
Since Digital Twins often monitor CPSs at runtime, data is also mentioned as part of the Digital Twin frequently
(129, 36.24 %) as well. To fulfill their representative purpose Digital Twins need information about the underlying
system, which makes the data a reasonable part. Therefore, intelligent data processing and storage could become
an important functionality of Digital Twins in the future, enabling them to remain up-to-date representation of the
twinnned system despite further growing data volumes.
4.9. RQ-3.1 - Implementation
Multiple publications not only elaborate on the conceptual foundations of Digital Twins but also provide a detailed
explanation about used techniques for their implementation. Thus, RQ-3.1 analyses different facets of realizations.
Our goal is to identify key technologies or methodologies to implement Digital Twins. Overall, 191 publications
contain information on implementation details. Thus, the following classification results refer to the total number of
publications that actually contribute to the research question. As different technologies can be used in combination
for realization, the following statistics are not disjoint.
Overall, 71 (37.17%) Digital Twin concepts are implemented using CAD or 3D models. Furthermore, 31 (16.23%)
of the papers follow data-driven approaches, such as standardized data formats (e.g., JSON, XML) or complete
database systems. General-purpose programming languages make a total of 65 (34.03%) publications. 59 (30.89%)
approaches are realized via mathematical or physical models, and 38 (19.9%) use a model-based or model-driven
approach. Finally, 42 (21.99%) papers consider simulations or similar analyses when implementing these twins.
Figure 11 presents the corresponding distribution.
CAD / 3D Models (71)
General Purpose
Language (65)
Mathematical /
Physical Models
(59)
Simulation /
Analysis (42)
MDE (38) Data /
Database (31)
Figure 11: Distribution of implementation techniques for Digital Twins

In summary, most of the considered publications use 3D and mathematics-based models as well as general-purpose
languages. Simulations, model-based, and data-driven approaches are also widely applied for implementing Digital
Twins. Most publications also describe a composite approach of several technologies to realize Digital Twins. Fur-
thermore, 8.71% of all publications describe the appliance of types or another kind of reuse. This relatively low
percentage shows that many Digital Twins are still created in a purpose-driven way, without relying on a consistent
foundation.
CAD or, more general, 3D models being the most prominent way of implementing Digital Twins is consistent
with the findings of RQ-3.2 and could indicate that a replica comprising the physical characteristics is often required.
Furthermore, the use of general-purpose languages is significantly high. This suggests that for many aspects of a
real-world entity or process, suitable tools do not yet exist, such that many features have to be implemented manually
on an individual basis. Additionally, Digital Twins are often based on physical models, model-based techniques,
simulations, and slightly less on data. Overall, this distribution seems to indicate that currently, these twins primarily
rely on model-driven or analytic approaches rather than purely data-driven techniques.
4.10. RQ-3.2 - Digital Twin Tooling
The purpose of RQ-3.2 is to understand which kinds of software tools are applied to the engineering and operations
of Digital Twins. To this effect, we classified the tools mentioned in the publications of our corpus according to the
classification schema presented in the corresponding facet.
Overall, we found that only 186 (52.25 %) publications of 356 publications make the tools employed explicit.
Various publications to Digital Twin architectures or infrastructures present concepts or methods that are unrelated
to specific tools. Out of the 186 publications making the employed tools explicit, the most popular category of
tools is simulation tools (70, 19.66 %), which includes the Maya Simulation Framework [25] Siemens PLC Sim
Advanced [156], Simumatik 3D [148], Verosim [50], and more. The second most often category of tools focuses
on computer-aided design and manufacturing (CAX) tools (53, 14.89 %), which includes Autodesk Revit [129],
CATIA [33], Delima 3D Experience [263], Siemens NX [45], SOLIDWORKS [106], and similar tools.
Tools for process management (37, 10.39 %), communication (24, 6.74 %), data management (32, 8.99 %),
visualization (35, 9.83 %), as well as the direct use of programming languages (34, 9.55 %), and model-driven
development tools (33, 9.27 %), are less common than simulation and CAX. Moreover, there also is a large number
of publication using various other software (36, 10.11 %), which includes website development for Digital Twin
representation with Apache Kepler [248], interfacing specific robot APIs [141, 136], data modeling with Microsoft
Excel [160], or specific programming environments [220]. Overall, the main categories of tools employed in the
engineering and operations of Digital Twins are as illustrated in Figure 12.
Simulation
Tools (70)
CAX Tools (53)
Process
Management
Tools (37)
MDD Tools (33)
A.I. Tools
(8)
PLM
Tools (7)
Other Tools
(36)
Programming
Languages
(34)
Data Management
Tools (32)
Visualization
Tools (35)
Communication
Tools (24)
Figure 12: Simulation tools and CAX tools are used most often to develop Digital Twins.
The most prominent tools to engineer Digital Twins are simulation tools and CAX tools. This might indicate
that some notions of Digital Twins indeed aim for a sufficiently precise replica of the twinned system that can be

subjected to experiments as a substitute for the twinned system itself. Moreover, this might entail that Digital Twins
are predominantly researched in domains being used to describing systems for a physical-geometrical perspective (cf.
RQ-1.1). As both simulation tools and CAX tools traditionally are employed to engineer systems, i.e., prior to the
deployment of the system under development and its operations, this furthermore might suggest that there is a strong
focus on Digital Twins used at design time of the twinned system. On the other hand, the widespread use of process
management software, communication software, and data management software indicates that there also is extensive
interest in observing and possibly optimizing the behavior of the twinned system at runtime.
The horizontal analysis on the use of tools relative to the domain (RQ-1.1), its purpose (RQ-1.2), and lifecycle
(RQ-2.4) of the respective Digital Twins discusses this.
The number of process management tools and data management tools, ranging from traditional databases to data
analysis tools might portend that Digital Twins also are about better understanding the twinned system and its opera-
tions in its context. Yet, in contrast to RQ-2.1, according to which twinning system-of-systems is important, the low
number of communication tools employed to engineer and operate Digital Twins might suggest that from a Digital
Twin perspective, Digital Twins of systems-of-systems are generally considered a single, monolithic system instead.
This might be due to the lack of support for composing Digital Twins (cf. RQ-2.7).
With one interpretation of Digital Twins being that these are models of the twinned systems–which appears to
be the predominant perspective on Digital Twins for the researchers employing simulation tools and CAX tools–
the lack of applications of MDE tools is surprising. Again, this might be due to the large number of publications
focusing on Digital Twins in manufacturing included in our corpus. Also, the number of papers leveraging general
programming languages is relatively low. This might suggest that Digital Twins often are engineered and operated
by reusing existing software, such as specific tools for simulation, data management, or visualization. If research on
Digital Twins does not require new software, this might suggest that Digital Twins are not a new paradigm or kind of
software per-se, but the combination of existing paradigms, methods, and tools for a new purpose. This also is in line
with the observed lack of special software solely focusing on Digital Twin engineering or operations. The horizontal
analysis of the use of tools (RQ-3.2) relative to the purpose (RQ-1.2) highlights this.
Finally, we found the lack of research employing artificial intelligence tools, including machine learning, knowl-
edge representation, and planning, surprising.
4.11. RQ-3.3 - Digital Twin Development Processes
Analyzing whether Digital Twins are developed together with their counterpart or in an independent process in
RQ-3.3, we identified that only 232 (65.17 %) make the engineering process of the Digital Twin explicit. Various
publications address the usage of Digital Twins or broader concepts and do not address the engineering of Digital
Twins or their counterparts.
The publications that made the engineering of Digital Twins explicit could be categorized into two categories,
those which describe an engineering process intertwined with the counterpart and those that describe that Digital
Twins are developed in a separate process. Furthermore, we identified whether the development of Digital Twins in-
corporated knowledge about the manufactured system or not if they are not developed together with their counterpart.
Of the 232 (65.17 %) publications making the development process of Digital Twins explicit, a total of 7 (1.97 %)
of 356 publications describe processes for Digital Twins both together with the development of the counterpart and
also independent of it. Overall, 29 (8.15 %) publications report that the Digital Twin is developed together with the
system, and 196 (55.06 %) that it is not. Of the latter, 14 (3.93 %) publications report that the development of Digital
Twins is independent of any manufactured counterpart, whereas 26 (7.3 %) report that the development of a Digital
Twin incorporates knowledge about the manufactured counterpart.
It is striking that Digital Twins are primarily developed in a separate development process, i.e., not in a joint
engineering process with the actual system. This could be a result of historical developments or because the Digital
Twin and its counterpart are primarily regarded as separate entities. Moreover, this might entail that Digital Twins are
mostly researched in domains where there already exists strong engineering processes independent of Digital Twins.
Integrating Digital Twins and the information they provide into a complex engineering process could present greater
challenges and require a corresponding shift in mindset. After all, some articles describe the development of Digital
Twins as anticipating the development of their counterparts. This suggests that information gained from Digital Twins
is already flowing into the development of their counterparts, for example, for design space exploration. The next

Isolated Engineering (196)
Generic Standalone
Engineering (156)
Joint Engineering (29)
Both (7)
Subsequent
Engineering
(26)
Explorative
Engineering
(14)
Figure 13: Distribution of the reported development process of Digital Twins
step here would probably be to develop Digital Twins and counterparts together in order to allow information to flow
iteratively into the development process and thus to be able to react to changes in the development. However, it is
also interesting that in some cases, the Digital Twin is developed after the actual counterpart. This could occur, for
example, in cases where the counterpart is retrofitted with a Digital Twin that then interacts with the system at runtime,
for example, to control or influence the system. However, we found that for the actual engineering of and with Digital
Twins, there is a lack of research.
4.12. RQ-4.2 - Digital Twin Connections
We investigate how Digital Twins are connected to their counterpart. To this end, we classified Digital Twin con-
nections mentioned in the publications of our corpus according to the classification schema presented in Section 3.4.
We found that 77 (21.63 %) explicitly connect the Digital Twin with their real-world counterpart. Out of these
publications, 32 (8.99 %) explicitly name the technology they used for this connection. The technologies, together
with their occurrences, are illustrated in Figure 14.
Local Networks (15) Short Distance
Communication
(8)
Data / Database /
Data Format (7)
Server / Cloud / Proxy (2)
RFID (3)
Bluetooth (5)
Figure 14: Digital Twin communication technologies
Most of the connected Digital Twins are connected with their counterparts via a local networks (15, 4.21 %),
or short distance wireless communication (8, 2.25 %), such as Bluetooth (5, 1.4 %) and RFID (3, 0.84 %). Other
publications mention that the Digital Twin is connected to the twinned system through data access via some database or

data format (7, 1.97 %) and via some cloud or server (2, 0.56 %). Furthermore, we classified the used communication
scheme as also described in Section 3.4. The results are illustrated in Figure 15.
Local Control
Systems (21)
IoT Protocols (11)
Internet
Protocols (8)
REST (5)
OPC UA (15)
MQTT (5)
MT Connect (6)
Other (6)
Figure 15: Digital Twin communication protocols
Most of the publications used protocols from industrial control system environments (21, 5.9 %). Of these pub-
lications, the majority (15, 4.21 %) use OPC UA and MTConnect (6, 1.68 %) to connect the Digital Twin with it’s
real-world counterpart. Some publications mention MQTT (5, 1.4 %) and other IoT protocols (11, 3.09 %). Further-
more, internet protocols (8, 2.25 %) are used in some cases to connect the Digital Twin with its real-world counterpart.
From these numbers, it becomes clear that Internet technologies and protocols are the predominant means to
connect Digital Twins with their counterpart. Moreover, it is easy to see that technologies and networks from IoT
applications are also an important part of Digital Twin development and connection. The large number of industrial
control system communication protocols and IoT protocol also meets our observation from RQ-1 that Digital Twins
are mostly used in the manufacturing domain in the context of industrial control systems. However, we can see
from the relatively small number of publications that communication between the Digital Twin and its counterpart is
currently not the main focus of research.
4.13. RQ-5.1 - Decision Making Functions
The notion of Digital Twins often comes with the intention of optimizing systems or process optimization. There-
fore, RQ-5.1 investigates the possibility of decision making functions. The goal was to explore the most prominent
decision methods used for Digital Twins. In general, it is noteworthy that 234 (65.73%) publications do not describe
any explicit decision making functionality. Therefore, the following statistics apply to the remaining 122 publications.
As certain publications use several decision making functions, the categorization is not disjointed. Overall, 55
(45.08%) Digital Twin concepts use some kind of reasoning. Further investigation shows that only a few case-based
reasoning is performed (four publications in total) compared to symbolic, stochastic, or other numerical reasoning
methods (14 to 19 papers each). Other numerical reasoning represents the subset of reasoning methods that do not fit
into the remaining three categories. Furthermore, 41 (33.61%) of the presented Digital Twins use machine learning
techniques, and 31 (25.41%) rely on simulation. Data Mining techniques have the least impact, with only 12 (9.84%)
publications reporting on this topic. Figure 16 presents the corresponding distribution.
In summary, only one-third of the publications describe decision making in combination with Digital Twins.
Methods of reasoning, simulation, and machine learning seem to have made significant advances. Data Mining
techniques are severely underrepresented. This is an interesting fact indicating that most decision making processes
rely on analyzing near real-time data and do not perform exhaustive computations on historical data. This might be
due to the lack of historical data and change in the future accordingly.
4.14. RQ-5.2 - Digital Twin Inputs and Events
We aim to understand how Digital Twins gain information about their counterparts and the operating context and
to which events and external inputs they react. We found that all but 129 (36.24 %) publications made explicit on

Reasoning
(55)
Machine
Learning (41)
Simulation (31)
Data Mining (12)
Symbolic
Reasoning
(19)
Other Numeric
Reasoning (18)
Stochastic
Reasoning
(14)
CBR (4)
Figure 16: Distribution of decision making techniques
which input the presented Digital Twin relies. Of the publications making the Digital Twin’s inputs explicit, a total of
45 (12.64 %) described Digital Twins processing multiple types of inputs. For example, several Digital Twins react
to changes of machine data but also support reconfiguration [12], where a Digital Twin of a 3D printing machine
automates experiments to detect parameters for achieving desired product attributes. For this purpose, the Digital
Twin analyzes sensor data and also provided 3D specifications of the produced part. Most Digital Twins (177, 49.72
Machine data
177
70
User Specification
Models & Simulation
38
Figure 17: Inputs of Digital Twins
%) react to machine data as input. This aligns with the large number of Digital Twins from manufacturing and might
simply stress that Digital Twins are often applied for automation in manufacturing. From these numbers, it is easy to
follow that many Digital Twins tend to react to the data they receive or measure from their observed systems, from the
system’s users, or the systems environment. In addition, some Digital Twins appear to react to damage or fault events,
which is consistent with our observation from RQ-1.2 that Digital Twins can be used for maintenance purposes.
4.15. RQ-5.3 - Digital Twin Output
We want to understand whether Digital Twins produce outputs and, if so, what kind of outputs they produce.
To this end, we identified publications that reported that Digital Twins emit some output or affect the environment
they are deployed in. We classified the kind of these outputs according to the classification schema presented in the
corresponding facet.
Of the examined publications, 227 (63.76 %) made explicit that Digital Twins emit some kind of output. In cases
where Digital Twins do not provide outputs, they could represent structural models instead of software that provides
analyses. In some cases, Digital Twins performed analyses. However, it was not stated what happens with the analyses
results or how these influence the environment or the counterpart.
Out of the 227 publications making outputs of Digital Twins explicit, a total of 66 (29.07 %) broadly stated that
Digital Twins produce some kind of analyses result or emit some kind of data, but did not report any specifics on
these outputs. This was, e.g., the case in publications that presented a high-level concept of Digital Twins that could
fulfill varying purposes. Interestingly, 33 (14.54 %) papers reported that Digital Twins had as output some kind
of effect on visualization by updating information shown in user interfaces or even updating producing whole 3D
models. A total of 44 (19.38 %) publications reported that the output of Digital Twins represents the current state of

the counterpart, including information about material or energy consumption, defect information, or current system
behavior. Instructions or modifications, i.e., what should change (76, 33.48 %), or predictions and estimations (42,
18.5 %) are also common outputs of Digital Twins.
Plans (76) Observations
(44)
Predictions
(42)
Visualization (33)
Other Data (66)
Actions and
Commands (27)
Plan (22)
Configuration (14)
Gen. Optimization (13)
Predictive System
Behavior (14)
Defect
Estimation (11)
Expected
Lifespan (7)
Other
Predictions
(6)
Fault/
Quality
Estimation
(4)
System
Behavior (21)
Defect
Information
(8)
System
Effectors
(5)
Other
Monitor
Data (4)
Concern (1)
Material
Consumption
(5)
Figure 18: Distribution of the reported outputs of Digital Twins
Digital Twins can produce various outputs and consequently affect their environment, respectively counterpart in
different ways. Most prominent are outputs that provide controlling data or instruction to the counterpart, ranging from
changes of the parameterization to elaborated planning. These outputs are intuitive as they describe a strong interaction
between Digital Twin and counterpart. The Digital Twin serves in particular as controller of the counterpart. With
pure monitoring approaches, which also make up a substantial part of the examined publications, the question arises
why Digital Twins are needed here. Monitoring approaches are probably intended to record the current state of the
system as accurately as possible and make this information available to other systems. More sophisticated Digital
Twins can not only examine the current state of the system but also make predictions, such as analyzing expected
lifetime or predicting failure probabilities. Such Digital Twins could be used to monitor safety-critical systems in
particular.
4.16. Further Insights
Not all research questions could be answered reliably by the publications included in our corpus. For these
questions, this section presents our observations.
RQ-2.2 - Multiple Digital Twins. In our survey, we investigated whether the described Digital Twins are unique or if
the described physical entity may have multiple Digital Twins. The majority of papers (251, 70.5 %) did not explicitly
state how many Digital Twins are supported by their approach. Overall, only 63 (17.7 %) publications explicitly
excluded the possibility of multiple Digital Twins while 42 (11.8 %) publications supported the idea of multiple
Digital Twins.
RQ-2.3 - Lifetime. Throughout the mapping study, we also tracked whether the described Digital Twins were operated
at designtime or at runtime of the physical counterpart. Most publications (282, 79.21 %) report on runtime Digital
Twins and 97 (27.25 %) publications describe Digital Twins that are operated at design time of the physical twin.

RQ-2.5 - Interaction Facet. We also investigated whether current research considers the Digital Twin to interact with
a twinned system. In our corpus, the majority of publications makes the interaction or the lack of it explicit (277, 77.81
%). Among the publications making interaction explicit, the majority (164, 46.07 %) supports interaction between the
Digital Twin and its counterpart. The other (113, 31.74 %) publications do not support such interaction. This might
be due to the different times a Digital Twin is employed in the lifecycle of its observed system and is discussed in the
horizontal analysis.
RQ-3.4 - Quality Assurance. We also found that only a small number publications (51, 14.32 %) of our corpus
consider quality assurance for Digital Twins at all. Where quality assurance was considered, testing (24, 6.74 %) was
more prominent than simulation (21, 5.9 %). Also, the number of publications considering the online verification of
Digital Twins with their counterparts is vanishingly low (7, 1.97 %). This implies a need for further research regarding
the quality assurance of Digital Twins (1) at design-time of the Digital Twin; (2) during design-time of the twinned
system; and (3) during runtime of both systems. Especially, the fidelity of Digital Twins, i.e., the verification that
these can properly the represent the twinned system at needs further investigation.
RQ-3.5 - Requirements. In addition to quality assurance measures, we also investigated whether own requirements
for Digital Twins and their development are considered in the publications of our corpus. As a result, we found that
only 38 (10.67 %) publications discuss own requirements for Digital Twins. Of these publications, most prominently
the necessity of real-time capability was mentioned (7, 1.97 %) closely followed by the requirement that the behavior
of the Digital Twin must match the behavior of its real-world counterpart (7, 1.97 %). Finally, it was mentioned
that Digital Twins have to be reusable in only (2, 0.56 %) publications, which is almost negligible. The remaining
publications that mentioned Digital Twin requirements, were either the only source in our corpus mentioning this
requirement, or discussed the necessity of Digital Twin requirements without going into the details.
Since apparently only a few authors have investigated the specific requirements of Digital Twins, it is reasonable to
conclude that Digital Twin specific requirements are currently not focused in research. However, the above-mentioned
Digital Twin specific requirements obviously focus on important aspects of Digital Twin development such as real-
time capability or the Digital Twins behavioral relationship to its physical counterparts. Thus, we think that further
research on the requirements for Digital Twins and their implementation may be necessary in future works on Digital
Twins.
RQ-4.1 - Digital Twin Host. We were furthermore interested which systems host Digital Twins. The investigation
showed a wide variety of concept and technologies used, ranging from the the Digital Twin living on the same device
as its counter part to the Digital Twin being deployed in a cloud. Only few publications (3) report that the Digital
Twin lives on the edge of its counterpart. Of the publications that reported that the Digital Twin is deployed further
away from its counterpart, 81 reported that the Digital Twin its deployed in the cloud, 8 reported that the Digital Twin
lives on a specifically named platform, while 11 contributions do not further clarify on what kind of external system
the Digital Twin is deployed. Also,5 publications report on deploying the Digital Twin on the edge of its counterpart.
For another 4 publications, the Digital Twin is managed in a database, and for 2 contributions it is part of a virtual
reality. In summary, the different implementations show the diverse perception of Digital Twins, but it is clear that
network-based systems and management in the cloud are strong pioneers here.
RQ-6.1 - Digital Twin Evaluation Maturity Facet. We found that the majority of publications in our corpus feature
some form of evaluation (271, 76.12 %). Overall, proof-of-concept (TRL 1-3) evaluations are significantly prevalent
(181, 50.84 %). Evaluations featuring technologies employed in a laboratory or relevant environment (TRL 4-6) are
reported less often (77, 21.63 %) and evaluations featuring system prototypes in operation environments (TRL 7-9)
are very rare (13, 3.65 %). As Digital Twin research often aims at the application domains of manufacturing, energy,
or construction, the low number of evaluations featuring systems in their operation environments is comprehensible:
evaluating a research product in a real factory, power plant, or construction site is challenging and costly. Yet, with
Digital Twins aiming to improve productivity, this validation in the field ultimately is necessary to promote Digital
Twin research into industrial practice.

5. Orthogonal Analysis
The orthogonal analysis investigates the potential correlations between related dimensions of our classification
framework. To this end, we juxtapose several dimensions and further group their data to generate interesting findings.
In addition, we investigated other pairs of dimensions which are not explicitly presented in this paper. Based on
these investigations, we present the six most interesting analyses. Further analyses can be performed based on the
replication package our replication package6.
5.1. Digital Twin Purpose (RQ-1.2) vs. Lifetime (RQ-2.3)
Out investigation of RQ-2.3 revealed that most Digital Twins operate at runtime of the twinned system rather than
at its design-time. However, since we suspect a correlation between the purpose of Digital Twins and their lifetime, we
examine this relationship in more detail. Since for both, Digital Twin lifetime and Digital Twin purpose, the identified
categories are non-disjoint and a publication can therefore be assigned more than once, the number of combinations
considered here is larger than the number of publications. That is, the 356 publications of our corpus contribute to
638 combinations of Digital Twin purpose and Digital Twin lifetime, as shown in Figure 19.
In the vertical analysis, we identified that Digital Twins are more often employed at runtime than at design-time of
the twinned system Section 4.4. Following this trend, CPS Behavior Optimization (145 observations, (79.31 %)) and
Monitoring (116 observations) are predominantly performed at runtime. On the other hand, validation is more often
performed at design-time (92 observations, (66.3 %)) than the overall trend suggests. Only design-space exploration
is more often mentioned in publications presenting Digital Twin at design-time of the twinned system than at runtime,
which also meets the intuition that design-space exploration is performed at the design-time of a system.
153
638
456
50
Data
Processing
33
14
3
CPS Reuse
2
1
15
Enterprise
Decision
Making
8
6
34
Design-Space
Exploration
15
18
1
9
N/A
7
1
1 29
78
Behavior
Prediction
60
15
3
95
122
Monitoring
21
6
22
26
Maintenance
3
1
51
Visualization
36
13
2
96
Validation
61
31
4
Teaching
2
3
Runtime
Design-Time
N/A
Digital Twin Purpose (RQ-1.2)
Digital Twin Lifetime (RQ-2.3)
151
Behavior
Optimization
115
30
6 13 1
Figure 19: Digital Twin purposes relative to observed lifetime phases
Interestingly, a high percentage of contributions to design-space exploration of Digital Twins refers to runtime
data for the design-space exploration. As the idea of using runtime data of system under development seems counter-
intuitive, we looked at the publications again to get a better understanding of these cases. By this, it became clear that
most of these publications consider multiple purposes at different times of the Digital Twins lifecycle [152, 179, 351],
use real time data from similar products [129, 190], or perform experiments to get the required real time information
at design-time [67].
6https://zenodo.org/record/6560195

5.2. Digital Twin Purpose (RQ-1.2) vs. Lifecycle Phase (RQ-2.4)
While the vertical investigation of RQ-2.4 finds that most Digital Twins aim to describe, monitor, or control
the twinned system as-operated, this section relates the lifecycle phases of the twinned system that are observed or
represented by Digital Twins to the Digital Twins purposes uncovered through RQ-1.2. Through this analyses we aim
to better understand for which purposes Digital Twins are used with respect to the observed systems lifecycles and
whether there are gaps on this. While some purposes might appear to be obviously related to certain lifecycle phases,
such as that Digital Twins with the purpose of supporting system maintenance might more often be used with the
systems as-operated, other purposes, such as behavior prediction or enterprise decision making are equally suited for
Digital Twins twinning systems as-designed, as-manufactured, or as-operated. Understanding the relation between
Digital Twin purpose and the twinned lifecycle phases of the observed system sheds light the use of Digital Twins and
can guide further research.
Overall, the 356 publications of our corpus contribute research to 799 combinations of purposes and lifecycle
phases. This is due to many publications considering multiple purposes for Digital Twins presented in their research.
For instance, the Digital Twin presented in [8] aims to ease CPS data processing, monitoring, behavior prediction,
and behavior optimization. Hence, this publication contributed four entries to this orthogonal analysis as presented in
Figure 20.
Generally, we have found five times as many publications that focus on Digital Twins of systems as-operated
than on the systems as-manufactured and two times as many that focus on as-designed than on as-manufactured. As
illustrated in Figure 20, these are not distributed evenly over the different purposes. For instance, research on Digital
Twins for CPS behavior optimization and CPS monitoring often focuses on Digital Twins twinning the observed
system as-operated. To this end, Digital Twins often are considered software systems that collect data from the
observed system and process that to represent it to human operators [4, 235, 236] or control the system directly [28,
115, 353]. In contrast, research on Digital Twins for CPS behavior prediction instead focuses on Digital Twins as
manufactured [22, 111, 217]. And research on Digital Twins for CPS validation and CPS design-space exploration
focuses on the systems as-designed [278, 94, 73].
Considering the lifecycle phases globally, it is apparent that behavior optimization is the most important topic for
Digital Twins relating to any of the three phases.
Overall, research on Digital Twins that aims to improve the behavior optimization, behavior prediction, data
processing, monitoring, or validation for Digital Twins that twin their observed system as-operated makes up the five
most important combinations of Digital Twin purposes and lifecycle phases and contribute a total of 351 (43.93 %) to
the 799 combinations of purposes and lifecycle phases (highlighted with bold numbers in Figure 20).
As discussed in the vertical analysis, Digital Twins twinning systems as-operated make up the large majority
of approach in current Digital Twin research. This might indicate that Digital Twins are closely related to real-world
data processing or the consideration of real-world effects, such as hardly foreseeable environmental conditions, system
uses, or highly detailed wear-and-tear. However, this focus of Digital Twin research vanishes where Digital Twins
are used for validation or design-space exploration, where Digital Twins of the observed system as-designed are more
prominent. As design-space exploration and validation typically are activities performed during systems development,
this might indicate a gap between Digital Twins used during systems design for these purposes and the Digital Twins
used during real operations of the developed systems. We suggest investigating this gap as well as means to reduce it,
e.g., the derivation of a Digital Twin for a system as-operated from a Digital Twin of the same system as-designed or
as-manufactured.
5.3. Digital Twin Lifetime (RQ-2.3) vs. Parts (RQ-2.7)
As Digital Twins exist at different times (RQ-2.3) and may also consist of various parts (RQ-2.7), this section
examines the correlation of both corresponding research questions. We investigate which constituents are prominent
as their diversity is very likely concerning the different purposes of a design-time and runtime twin (cf. Section 5.2).
While some parts can obviously be related to a certain lifecycle phase, such as the use of models during design-time,
there are also constituents that may exist to unexpected phases (e.g., historical data during design-time). Understand-
ing the relationship between the lifecycle of Digital Twins and their parts assists their further engineering.
Overall, 591 combinations have emerged from comparing the lifecycle and constituents of Digital Twins. While
428 (72.42%) of the combinations describe a twin at runtime, some interesting correlations still appear.

101
799
434
58
Data
Processing
37
8
6
7
5
CPS Reuse
2
1
1
1
16
Enterprise
Decision
Making
7
5
2
289
44
Design-Space
Exploration
16
16
6
5
1
15
N/A
5
4
2
3
121
154
105
Behavior
Prediction
57
15
16
15
2
89
26
157
Monitoring
20
17
5
20
5
31
Maintenance
3
2
1
58
Visualization
30
14
6
7
1
119
Validation
57
29
17
12
4
Teaching
3
3
As-Operated
As-Designed
Combinations
As-Manufactured
N/A
Digital Twin Purpose (RQ-1.2)
Observed Lifecycle Phase (RQ-2.4)
188
Behavior
Optimization
111
31
22
18
6
Figure 20: Digital Twin purposes relative to observed lifecycle phases
Generally, Digital Twins that exist during the runtime of the observed entity tend to use relatively more hardware
components as during design-time. The same observation can be made for historical and live data. Additionally, this
leads to the intriguing question of how a Digital Twin can access such data in the first place since it exists before the
system under investigation is put into operation. In general, models are often part of Digital Twins as they are involved
in 312 (52.79%) realizations.
The frequent use of hardware components and data at runtime of the system is intuitive since they are effectively
available at this lifecycle phase. Generally, the question arises to what extent hardware can be part of a purely digital
system at all; however, this result can be related to the authors’ interpretation of Digital Twin constituents. For
instance, sensors that are dedicated to produce input for a twin’s computation could be considered part of a Digital
Twin. Thus, some authors seem to include the required hardware infrastructure, while others clearly distinguish
between hardware and software.
The use of hardware, as well as historical and live data at design-time, can have different origins. The Digital
Twin could use hardware prototypes and simulations that produce input data. Furthermore, a twin might be subject to
a bootstrapping process, in which a system is developed from previous versions of a similar system, enabling access
to legacy components and recorded data traces.
The overall outstanding use of models as part of Digital Twins indicates a growing application of model-driven
techniques. Beyond the extensive benefit of models at design-time, however, their predominant use at runtime of a
system is also significant. One explanation might be that models that were already created during development are
also stored in the Digital Twin (e.g., for documentation purposes). Furthermore, this finding might indicate a growing
relevance for models at runtime, applying changeable models in the behavior of the overall system.
5.4. Digital Twin Decision Making (RQ-5.1) vs. Lifetime (RQ-2.3)
This subsection relates the Digital Twin’s lifetime with its decision-making capabilities. Figure 22 shows a map-
ping between the possible lifetimes of a Digital Twin that can either be design-time or runtime and different techniques

591
30
43
Software
Components
3
9
31
312
Models
16
80
216 428
37
N/A
5
9
23
133
52
Hardware
Components
4
7
41
N/A
Design Time
Runtime
Digital Twin Parts (RQ-2.7)
Digital Twin Lifetime (RQ-2.3)
147
Data (live/ historic)
2
28
117
Figure 21: Digital Twin lifetime relative to observed parts
for decision-making that were reported as Digital Twin capabilities. The Digital Twin’s ability to respond to its con-
text seems to be related to its lifetime. Of the publication on Digital Twins investigated in our study, 93 (26.12 %) are
able report decision-making capabilities. Of these, 30 (8.43 %) publications reported on decision making at design
time while 126 (35.39 %) publications applied decision making at runtime. As these numbers indicate, there must be
an overlap between Digital Twins that apply decision making at design-time and those that apply decision making at
runtime. More specifically, all Digital Twins that apply decision making at design time also apply decision making at
runtime.
At design-time, simulation is especially applied for decision making (11, 36.67 %). At runtime, 40 (31.75 %)
Digital Twins applied machine learning, 26 (20.63 %) relied on simulations, and 15 (11.9 %) Digital Twins used
symbolic reasoning.
While it is intuitive that Digital Twins perform adaptations autonomously at runtime, there seems to be a research
gap for Digital Twins that act on their own at design-time. However, Digital Twins that evaluate different designs
at design-time and create an optimal configuration of the designed product could decrease development times in the
future.
5.5. Digital Twin Connection Types (RQ-4.2) vs. Lifetime (RQ-2.3)
The goal of this section is to understand which connection techniques can be applied at which lifetime of the
twinned system. For example, if communication requires spatial proximity between the communicating entities. In
total, 14 (3.93 %) publications reported a Digital Twin that was connected and applied at design time, respectively
114 (32.02 %) publications reported a Digital Twin that was connected and applied at runtime of the counterpart.
This discrepancy is quite intuitive since many Digital Twins encapsulated sensor data (Section 4.8) as part of the
Digital Twin and thus require a connection to the physical counterpart to acquire this data. These runtime Digital
Twins were often connected via local area networks (28, 24.56 %), or support OPC UA (21, 18.42 %) or internet
protocols (15, 13.16 %).
Integrating sensor data or historical data into Digital Twins that are employed at design time, can support learning
from usage information and adapting future versions of the physical twin. Of the design time Digital Twins (6, 42.86
%) were connected via OPC UA, thus integrated runtime data of operating physical things.
Considering Figure 23, the majority of design time Digital Twins do not mention their connection, yet. Thus,
further research in integrating data from operating twins or the envisioned operation context is still relevant.

423
23
4
Case-Based
Reasoning
4
21
Symbolic Reasoning
1
5
15
13
Data Mining
2
11 298
38
Simulation
1
11
26
102
17
Stochastic
Reasoning
3
14
19
72
263
None
172
1
5
46
Machine Learning
40
N/A
Design Time
Runtime
Own Decision Making (RQ-5.1)
Lifetime (RQ-2.3)
21
Other Numeric
Reasoning
1
4
16
Figure 22: Relations between lifetime and decision making
5.6. Digital Twin Implementation Techniques (RQ-3.1) vs. Lifecycle Phase (RQ-2.4)
The orthogonal analysis relating implementation techniques to the lifecycle phases the Digital Twin addresses (cf.
Figure 24) aims to uncover which techniques are best suited for the twinning a system as-designed, as-manufactured,
or as-operated. Consequently, it also might identify gaps in research in form of technologies not applied to specific
lifecycle phases.
Overall, the 356 publications of our corpus contribute research to a total of 678 combinations of implementation
techniques and lifecycle phases. This occurs as many publications combine multiple implementation techniques and
consider several purposes for Digital Twins presented in their research.
Generally, research on Digital Twins as-operated produced two times as many contributions than research on
Digital Twins as-designed and and four times as many on Digital Twins as-manufactured. The high number of publi-
cations not making their implementation explicit (225, 33.18 %) suggests that much research on Digital Twins actually
focuses on conceptual research that cannot be translated into Digital Twins without further information.
Hence, out of the five most popular facet combinations, two belong to the “N/A” column, i.e., where the technique
of implementation is unspecified. This suggests that there are many publications reporting conceptual contributions
to Digital Twin research. In our corpus, these most often are high-level reference models on the similar conceptual
abstraction than RAMI 4.0 [383] that suggest how to organize architectures of Digital Twins without implementa-
tion [12, 97, 107]. The other three most popular facet combinations belong to research on Digital Twins as-operated
while using CAD models, General-Purpose Programming Language (GPL) code, or mathematical models, which
suggests that purely data-driven approachs, MDE models and simulation models are less relevant implementation
techniques for Digital Twins.
For Digital Twins as-designed, CAD and 3D models as well as mathematical models are the primary implemen-
tation techniques, which make up 55 (52.38 %) of the contributions to corresponding Digital Twin research. While
for Digital Twins as-manufactured, the applied implementation techniques are distributed almost evenly, the overall
numbers of contributions to such Digital Twins is too small to generalize.
Overall, the data suggests that CAD and 3D models are overproportional important for developing Digital Twins
as-designed, where they account for 36 (25.17 %) of the overall as-designed contributions. In contrast, for Digital
Twins as-operated, they only make up 45 (12.89 %) of the overall as-operated contributions. Similar observations

419
24
27
OPC UA
21
6
REST
6
1
5
8
IoT Protocols
1
7300
6
MTCOnnect
6
95
7
RFID
7
1
1
12
Specific Data Format
10
1
16
Internet Protocols
15
N/A
Design Time
How Connected (RQ-4.2)
Lifetime (RQ-2.3)
3
Server/ Cloud/
Proxy
1
1
1
1
29
Local Networks
28
1
8
Bluetooth
7
21
81
288
N/A
186
Runtime
2
9
MQTT
7
Figure 23: Relations between Digital Twin lifetime and their connection to the twinned system
hold for simulation implementations, which seem to be more important for Digital Twins as-designed (16, 11.19 %)
than for Digital Twins as-operated (31, 8.88 %).
The different prominence of implementation techniques for Digital Twin research focusing on different lifecycle
phases of the twinned system might suggest a technological gap between Digital Twins used to twin systems as-
designed and Digital Twins used to twin systems as-operated. This also could explain the low number of Digital Twin
research addressing more than one lifecycle phase of the twinned system.
6. Engineering Dimensions of Digital Twins
While reading the included publications, we noted and synthesized a collection of concerns that need to be con-
sidered when engineering and operating different digital twins. We have clustered and arranged these in the feature
models presented in the following. Note that each intermediate feature refers to the research question its subfeatures
where extracted from. Overall, we have identified four dimensions of digital twin engineering and operations:
1. The requirements dimension comprises concerns that define the capabilities of the Digital Twin under develop-
ment. Design choices within this dimension include identifying the Digital Twin’s counterpart, defining whether
there can be one or multiple Digital Twins of the same system, and fixing the phase of the twinned system that
the Digital Twin shall represent. Decisions made along this dimension govern what the Digital Twin under
development will be capable of.
2. The realization dimension comprises concerns about implementation techniques, tools, and methods applied
to engineering Digital Twins. Design choices within this dimension include selecting modeling and program-
ming languages, a development process, and suitable quality assurance techniques. Decisions made along this
dimension essentially govern how the Digital Twin will be developed.
3. The deployment dimension is about bringing the Digital Twin to life and includes concerns about hosting and
connecting the Digital Twin. Design choices include deploying the Digital Twin locally, in the cloud, or in-
between, installing it on a simulator or in a virtual environment, and selecting appropriate means to connect it
to its counterpart(s). This dimension governs where the Digital Twin will exist.

83
678
349
77
47
MDE Model
28
9
4
4
2
225
N/A
131
39
24
22
9 26
143
43
Data (base)
24
7
4
5
3
As-Operated
As-Designed
Combinations
As-Manufactured
N/A
Digital Twin Implementation (RQ-3.1)
Observed Lifecycle Phase (RQ-2.4)
112
CAD / 3D
Models
45
36
16
14
1
89
Handcrafted
GPL Code
43
17
11
10
8
92
Mathematical/
Physical Model
47
19
13
12
1
70
Simulation
31
16
11
10
2
Figure 24: Digital Twin implementation techniques relative to addressed lifecycle phases
4. The operations dimension is about the Digital Twins runtime behavior. It includes concerns about stimuli the
Digital Twin reacts to, interaction with other systems (such as enterprise information systems), and decision
making techniques influencing its behavior. Hence, this dimension governs what the Digital Twin will do.
Developing a Digital Twin involves making choices for each characteristic along these dimensions. To support
this, the feature models presented in the following make these dimensions and their concerns to Digital Twins explicit
and guide Digital Twin engineers and users. Therefore, we considered all research questions that apply to these
dimensions and have categorized them accordingly.
6.1. Requirements Dimension
The requirements dimension covers the conceptual foundation for Digital Twins. These cover the basic con-
stituents and characteristics a Digital Twin must or can have to fulfill its purpose. Figure 25 provides a general
overview of conceptual features: To realize a Digital Twin, there must be some kind of real-world entity in the first
place that acts as its counterpart. However, this requirement does not contradict the actual usage phase of the ob-
served entity. Thus, a Digital Twin may exist before its physical counterpart. Overall, we investigated three types of
counterparts, the first one of which is a living being, considering an individual. Furthermore, the physical twin can be
a dedicated system, for instance, a production machine in a factory. Finally, a Digital Twin can observe a composed
system (i.e., a system of systems), where multiple sub-systems are included. This situation focuses more on an overall
goal than on supervising individual components. Digital Twins for systems of systems often prove to be very mature.
The second requirement on Digital Twins deals with the question of whether a real-world entity may feature
multiple twins. This topic is highly controversial, and there are different approaches. On the one hand, some argue
that by the nature of Digital Twins, there can only be one twin for a physical counterpart, managing all tasks for
fulfilling its purpose. On the other hand, a Digital Twin might have a specialized view on a distinct part of a system,

Multiple
Representation
(RQ-2.2)
Counterpart
(RQ- 2.1)
Representation
Phase
(RQ-2.4)
Usage
Phase
(RQ-2.3)
Asset
Interaction
(RQ-2.5)
Consist of
(RQ-2.7)
Requirements
Being Individual
System
System of
Systems
As
Designed
As
Manufact.
As
Operated
Design
Time Runtime Direct Indirect
Optimization
(RQ-2.6)
Self CPS Data Hardware Model Software
Component
Digital Twins
(RQ-2.8)
Figure 25: Requirement dimensions of Digital Twins in terms of a feature model
thus allowing the coexistence of multiple Digital Twins for a single observed entity. While there are pros and cons to
both views, twin developers should consider this issue from the start to avoid potential conflicts later on.
A Digital Twin may exist at different stages in the lifecycle of a system. During design time, it supports the devel-
opment and during runtime the system’s operation. There may exist twins that cover both. Furthermore, independent
of its stage of existence, a Digital Twin can also represent different lifetime phases of the observed entity. Therefore,
Digital Twins can represent an entity as designed, as manufactured, or as operated. Again, multiple selections are
possible if the twins should cover more than one specific phase.
As there are different concepts on Digital Twins, the approaches also differentiate regarding the interaction with
their real-world counterparts. Some propose that the nature of a Digital Twin always includes direct interaction
between the twin and its asset, while others are content with a pure indirect approach. As the kind of interaction (or
if any exists at all) strongly depends on the Digital Twin’s purpose, different realizations, including a combination of
both attempts, are quite possible here.
Furthermore, a Digital Twin can be part of an optimization process. Our study revealed two main possibilities.
First, a twin could optimize itself, e.g., to improve its own analyses or give a more accurate representation of the
counterpart’s state. Second, the observed system can be optimized directly by automatically taking measures for
specific situations. Generally, also a combination of both approaches or no optimization at all might be feasible,
depending on the goal.
Finally, a Digital Twin must consist of some conceptual constituents. There are multiple different approaches,
including hardware and software components, data, models, or again other Digital Twins. Often, a combined effort of
different approaches is used. While some findings, such as the reliance on some hardware or software, are expected,
there are also further interesting building blocks. For instance, the use of models indicates an increasing notion
towards model-driven approaches (cf. Section 4.9). Another interesting aspect is the involvement of other Digital
Twins. The possibility of composing different twins to cover distinct sub-tasks comes with new possibilities but also
challenges and shows growing sophistication in the development of Digital Twins.
6.2. Realization Dimension
From an engineering perspective, it is important to know how Digital Twins are implemented, which tools are
used for their implementation, and which process is used for the Digital Twin development. For the realization of
Digital Twins, we propose a feature model as described in Figure 26. We describe the properties of these engineering
aspects in the following.
Every Digital Twin has an implementation, which defines how it achieves its purposes. In general, we identified
that Digital Twins describe a counterpart’s geometry, (software) systems, behavior, and general information about the
counterpart. To describe a Digital Twin’s geometry, the CAD/3D Model feature provides a modeling implementation
for geometry description and design. Moreover, the data feature describes an implementation to handle information
about the Digital Twins counterpart. We also noted that behavior descriptions are often implemented as mathematical
(including stochastic models) or physical, as well as simulations. Finally, we identified that the implementation of
concrete systems is often also realized as handcrafted GPL code and MDE models.
In addition to the implementation, we also identified several tools that may be used for the implementation. For
this, we identified A.I. tools that use the information from the counterpart’s data to make predictions and provide

CAD / 3D
Models
Handcrafted
GPL
Code
Mathematical/
Physical
Model
MDE
Model
Data
(base) Simulation
Tools
(RQ-3.2)
A.I.
Tools
Comm.
Tools
Data
Mgmt.
Tools
GPLs
CAX / 3D
Tools
Process
Mgmt.
Tools
PLM
Tools
Visualization
Tools
Simulation
Tools
Process
Explorative
Engineering
Subsequent
Engineering
Joint
Engineering
Consistency
Checking Simulation Testing Consistency Real-Time
Capabilities Reusability
Implementation
(RQ-3.1)
Realization
Development
Process
(RQ-3.3)
Quality
Assurance
(RQ-3.4)
Own
Requirements
(RQ-3.5)
Figure 26: Engineering dimensions of Digital Twins in terms of a feature model
services. To handle the required data, data management tools may be used to engineer Digital Twins. Furthermore,
CAx and 3D Tools may be used to process and provide geometric models and visualization tools for their visualization.
For other simulation purposes, we found out that also general simulation tools are usable in this context. As Digital
Twins are often embedded in a production environment, process management and PLM tools may also be helpful for
Digital Twin engineering. Finally, communication tools and GPLs may also be used.
For Process, we differentiate between different kinds of development processes. Each product and Digital Twin is
either developed jointly or isolated, with the development of the Digital Twin either frontloading the development of
the product (explorative engineering) or following its development (subsequent engineering). While the development
of the Digital Twin with explorative engineering is not bound to restrictions by already existing systems, the subse-
quent engineering of the Digital Twin has to take the constraints given by already existing systems into account but
may also reuse elements and knowledge from the development of these prior systems. In contrast to these approaches,
joint engineering of Digital Twin and counterpart enables to incorporate joint design decisions. Another aspect we
considered under the topic of Process is quality assurance. We identified mainly three kinds of quality assurance that
can be either used alone or together to assure a high quality of Digital Twins. First, consistency checking can be used
to validate the information the Digital Twin uses or produces. Moreover, simulations can be used as a verification
technique. Finally, also, testing as a verification and validation technique is a good method to assure a Digital Twin’s
quality. Apart from quality assurance, we also identified requirements specific to Digital Twins in our feature model.
We identified that consistency requirement, which requires that the Digital Twin’s behavior matches the behavior of
their real-world counterpart, are typical requirements of Digital Twins. Moreover, real-time capabilities of Digital
Twins are often required when the Digital Twin may serve a specific purpose concerning its real-time counterpart, and
therefore the Digital Twin must react to events in real-time. Finally, reuse is an own Digital Twin requirement, as a
new development of a Digital Twin for each physical counterpart is often unnecessary if the Digital Twin or parts of
the Digital Twin are reusable.
6.3. Deployment Dimension
The deployment dimension supports the design decision to bring the Digital Twin into action and is characterized
by its features shown in Figure 27. To this end, this dimension is concerned with two closely related topics, hosting

the Digital Twin in the real world and appropriate means for connecting it to its counterpart. Hosting is furthermore
concerned with where the host is located, which could be the twinned system itself, a local server, or a cloud system.
But hosting is also concerned with the kind of the host, as this may either be a data management system, a simulation,
or even a virtual environment. When deciding about deploying a Digital Twin in the real world, it is also relevant how
the Digital Twin is connected to its counterpart. While decisions about the connection of a Digital Twin are subject to
its host location, various design options still exist. As such, a Digital Twin could be connected through a BUS, some
other kind of local network, or even deploy Internet technology, such as respective protocols.
Hosting
(RQ-4.1)
Connection
(RQ-4.2)
Deployment
Cloud
Platform
Host
Location
Twinned
System
Local
Server
Data
Mgmt.
System
Virtual
Env.
Simulator
BUS Internet
Tech.
Local
Networks
Host
Kind
Figure 27: Deployment dimensions for Digital Twins in terms of a feature model
Decisions about the hosts’ location are alternatives. A Digital Twin does mostly not live simultaneously on a
twinned system and a cloud platform. Decisions here may also be subject to the type of Digital Twin to be employed
and its real-world counterpart. A Digital Twin that governs a whole factory is probably not located on a local machine;
vice versa, a Digital Twin that controls and monitors a single machine, may rather be deployed on the machine itself
than on a cloud platform. Design decisions here should be made with the purpose of the Digital Twin in mind.
In contrast, Digital Twins can support multiple host types. For example, a Digital Twin can be part of a database
management system that also incorporates simulations for value updates or provides simulations as an alternative
service. Furthermore, a Digital Twin deployed in a virtual environment could also function as a simulation of that
twin. Finally, the connection domain provides multiple selectable options. A Digital Twin can be both connected to
its counterpart through a bus and employed in a local area network.
While design decisions about the Digital Twin’s deployment are important, they are mostly subject to other con-
cerns and the available environment. The purpose of the twin affects the host location, which then limits the available
connection options.
6.4. Operation Dimension
The operation dimension classifies the Digital Twin behavior while the Digital Twin is running. It specifically
focuses on interaction with other systems, how the Digital Twin decides on next actions, and which kind of information
it exchanges with peripheral systems. All characteristics in this dimension are optional, which means that they are not
necessarily covered by all Digital Twins.
Horizontal communication encapsulates all Digital Twin communication with the main focus on information ex-
change where none of the involved partners can instruct another one to behave or change in a certain way. We
distinguish between information exchange with PLM systems, which was frequently mentioned during our study,
information exchange with the physical counterpart sharing, e.g., its current state, and even interaction with other
Digital Twins. The feature decision-making specifies how the Digital Twin determines its next actions or the data that
it exchanges. Machine Learning covers all Digital Twins that make predictions or decisions without being explicitly
programmed only by evaluating provided data. The data mining feature characterizes Digital Twins that evaluate data
sets and try to detect patterns. When Digital Twins imitate the physical world to decide on the best action, they have
the simulation feature. For reasoning, we further classified Digital Twins in symbolic reasoning, numeric reasoning,
and case-based reasoning. Digital Twins react to different input data and sources. These are covered by the Inputs

Decision Making
(RQ-5.1)
Outputs
(RQ-5.3)
Operations
Inputs and Events
(RQ-5.2)
Horizontal
Interaction
Other
Digital
Twins
PLM
Systems
(Physical
Counterpart)
Observations
(is)
Plans
(should)
Predictions
(can)
Machine
Data
User
Specification
Models and
Simulation
Results
Machine
Learning
Data
Mining Simulation
Case-Based
Reasoning
Symbolic
Reasoning
Numeric
Reasoning
Control
Command
Environment
Data
IoT Data Positioning &
Movement
Errors Sensor
Reasoning Visualization Other Data
Figure 28: Operating dimensions for Digital Twins in terms of a feature model
and Events feature. Machine data can cover error logs or notifications of machines, sensor data, other IoT data, and
general data about the environment in which the Digital Twin operates, e.g., temperature values in a production lo-
cation. User specifications can either be given as direct control commands that are specified via a user interface, but
some Digital Twins also evaluate human movements and gestures. Models and simulation results are also options for
gaining knowledge about the operating context, the intended behavior, or the physical entities. The output feature
describes Digital Twin outputs on a content level, so this feature does not characterize output formats or communi-
cation channels. Some Digital Twin reflect the current system state (the physical entity as it is), some Digital Twins
plan how the physical entity should act in the future (the physical entity as it should be), and some Digital Twins
predict the future behavior of the physical entity but do not influence it (the physical entity as it can be). Often, Digital
Twins combine several of the described features to fulfill an information need or specifically optimize the underlying
physical entity.
7. Threats to Validity
Our study is subject to threats to validity. In the following, we analyze and classify these according to [396] as
construct, internal, external, and reliability validity. Construct validity directly refers to the study’s overall design, such
as search query or evaluation criteria. External threats restrict the generalizability of a study, while internal validity
refers to the specificity, i.e., factors that influence the conclusions drawn from the readers. Reliability describes the
trustworthiness of the study’s results.
Regarding threats to the construction of this mapping study, there are plenty of similar yet distinct terms for de-
scribing Digital Twins, Digital Shadows, Virtual Twins, etc. While some publications extensively distinguish between
these terms, others use them interchangeably. To ensure an accurate mapping in our study, we have considered these
terms as separate concepts per default. However, if an investigated publication switched the wording while clearly
referring to the Digital Twin, we followed the paper’s intellectual roadmap and considered these as synonyms. Over-
all, this yields an accurate analysis result of the included papers. In contrast, publications using different terms only
(e.g., constantly mentioning virtual copies without including the term of a Digital Twin) could not be recognized in
this study, as it is impossible to extract whether the authors refer to the Digital Twin concept or explicitly distinguish
from it. This topic could be addressed in a future study that explicitly includes all potential synonyms, thus covering
a larger yet less precise scope.
A further threat to construct validity arises from our exclusion criteria during the initial screening of the papers,
as it only considers title, keywords, and abstracts. This procedure could mistakenly exclude potentially relevant
publications. To minimize this effect, we generally included papers for which we were uncertain and only excluded
these in the classification phase when they turned out to be not relevant for our mapping study.
Another threat resulting from the design of our study is based on the classification of publications. In general, the
categorization for several research questions is not disjoint, as a publication could be related to multiple dimensions.
For instance, Digital Twins can use combinations of different techniques for decision making. This causes difficulties
in evaluating the results since some dimensions may be highly interdependent. We designed the classification without

overlapping and used existing classification schemas to minimize the threat and only allow multiple assignments if
necessary.
Since our work is based on a literature study, it is inevitably subject to publication bias. Principally more successes
and positive reports on a topic are published. This complicates assessing the areas that are not positively affected by
Digital Twins or which concepts and methods for constructing them are not applicable. Furthermore, there may be
research and material outside of common research distribution channels, i.e., grey literature, which must be handled
specifically [406]. Further work on the analysis of the current status of Digital Twin research could focus on grey
literature.
Our study is also affected by external validity in terms of generalizability. We selected a rather general search
query to obtain a large corpus of publications. Including only online-available, peer-reviewed, English publications
(excluding short papers) reduces the corpus. This slightly affects generalizability, but at the same time, guarantees the
accessibility of our results and the reliability of the study. As the investigated publications cover different domains
and produce various findings, we cannot generally conclude that results from one problem domain apply to another.
Therefore, our study elaborates on the relationships between the individual clusters to identify similarities as well as
differences in a generalizable fashion.
We have used the search engines of leading scientific databases and libraries, such as SpringerLink, IEEE Xplore,
ACM, WoS, and Scopus, for searching the literature. We intentionally excluded google scholar as a search engine,
as it contains vast amounts of non-peer-reviewed publications (which are excluded during screening in any case).
Furthermore, google scholar does not store any publications such that generally, most relevant publications are found
as long as the related libraries are considered. Although this may negatively affect the external validity, it increases
the reliability of the search results.
Regarding internal validity, the publications differ significantly in the level of detail in which they explain Digital
Twins and their constituents. Authors often do not specify the exact system boundary of the Digital Twin, which
impedes a precise mapping regarding relevant technologies. For instance, it is often obscure whether a cloud system
is an integral part of the Digital Twin or whether the Digital Twin merely uses it for communication. To obtain an
unambiguous mapping, we generally decided these cases in favor of the Digital Twin, attributing these properties and
technologies to its realization. Additionally, controversial issues were discussed among the authors.
A further threat to internal validity is the readers’ different previous knowledge, which may lead to classification
discrepancies,e.g., through experience, more details can be anticipated. To minimize this effect as much as possible,
we have collectively read the first 60 publications to synchronize our mapping.
The conclusions drawn from analyzing the included publications can influence the reproducibility and, thus, the
study’s reliability. As mentioned for internal validity, we analyzed the publications in favor of the Digital Twin to
ensure an unbiased evaluation of the different sources. Another research group might draw slightly different conclu-
sions in particular circumstances. To add transparency and to ensure a reproducible study, we explained the research
method and corresponding design decisions in detail (cf. Section 3).
8. Conclusion
Our survey has shown that Digital Twins are researched in many domains, including agriculture, construction,
education, mining, transportation, and for a variety of purposes. Yet, the large majority of research on Digital Twins
investigates individual (cyber-physical) systems in manufacturing. We could not detect a trend that research on Digital
Twins is catching up in other domains, at least in terms of the number of publications. However, advanced Digital
Twins are already being presented for domains beyond manufacturing. Often, research on Digital Twins focuses on
monitoring the twinned system, as well as optimizing or predicting its behavior. Where research focuses on optimizing
the twinned system, the Digital Twin often acts as an outer control loop that adapts the twinned systems behavior, i.e.,
both systems, the twin and the twinned system, form a larger, self-adaptive system point of view. Such often emit
actions, commands, or plans that directly or indirectly (e.g., via another CPS management system) control the CPS’s
behavior. Consequently, research on Digital Twins as-designed, as-operated, or Digital Twins addressing multiple
lifecycle phases, is less common. Furthermore, current research also focuses on Digital Twins that are developed after
the twinned system. Rarely, the Digital Twin and the twinned system are engineered together.
We also found relatively few research on combining AI methods with Digital Twins. Instead, to engineer and
operate Digital Twins, a large variety of tools, e.g., for simulation, CAX, process management, visualization, data

management, and model-driven development, are used. The produced Digital Twins consist of models, complex
subsystems (e.g., databases or dashboards), plain GPL code, and sometimes even (mostly for augmented reality com-
ponents) hardware parts.
Through our survey, we also have identified and organized central design decisions common to engineering Dig-
ital Twins. These include (i) requirements on the number of twinned counterparts, when the Digital Twins should be
used and which lifecycle stage of the twinned system it should represent; (ii) realization decisions regarding imple-
mentation technologies, tools, and process; and (iii) deployment decisions on the Digital Twins hosting location and
its connections to the twinned system. The feature models detailing these represent the state-of-the-art decisions to
consider when engineering Digital Twins. We expect future Digital Twins development to contribute further decisions
to the presented feature models. Yet, they can serve researchers and practitioners as a guidance when considering
Digital Twins.
Based on our observations, we identified seven challenges for the future of Digital Twin research:
1. Domain-specific Digital Twins (RQ-1.1). The large body of Digital Twin research focuses on a single domain,
primarily manufacturing, yet other domains employ technologies that can serve as an excellent foundation for
further Digital Twin research.
2. Composable Digital Twins (RQ-2.8). Most Digital Twins found in our survey are build from scratch. The
reliable combination and composition of Digital Twins is essential for their effective (re)use. Different methods
to support these processes are required. For instance, integrating the Digital Twin of a motor into the Digital
Twin of a car may require another composition method than integrating the Digital Twin of a manufacturing
device into the Digital Twin of a factory. For instance, building information modeling based on IFC (ISO
16739) in architecture and construction supports the integration of various concerns of Digital Twins and can
be employed for many of the purposes found in our study.
3. Standardization (RQ-3.1). Literature yields a wide continuum of systems considered Digital Twins by the
authors from various domains. These range from high-fidelity simulation models to model-less software sys-
tems operating on the twinned systems and various combinations in-between. A future, in which Digital Twins
(e.g., using Digital Twins as contract parts between OEMs and suppliers) can be exchanged, combined, and
integrated, requires a common understanding of the concept. Currently, there is an ISO standard for Digital
Twins in manufacturing7in development that might at least harmonize the understanding of Digital Twins in
that domain. Whether this standard will be compatible with the understanding in other domains needs to be
evaluated and technological implementations on, e.g., exchange interfaces for Digital Twins, need to follow
then.
4. Tool support (RQ-3.2). While we have identified a large variety of tools employed to engineer and operate
Digital Twins, we found very few tools specifically tailored to Digital Twins. While there are some tools
mentioned in literature, such as Amazon Greengrass8, Eclipse Vorto9, or Microsoft’s Digital Twin Definition
Language10, these largely focus on data structure modeling and data exchange for Digital Twins but do not
cover the full spectrum of modeling concerns.
5. Modeling support (RQ-3.2). Abstraction is the key to understanding and improving CPSs. Consequently,
models are essential to Digital Twins. This is not limited to software engineering models, but includes CAD
models, mathematical models, physical models, simulation models, and many more. However, modeling meth-
ods developed by software engineers are also used by professionals without formal software engineering train-
ing. Therefore, software engineering must provide methods to integrate, analyze, and transform models used in
research and practice so that they can be used without software engineering background.
7https://www.iso.org/standard/75066.html
8https://aws.amazon.com/de/greengrass/
9https://www.eclipse.org/vorto/
10https://www.aka.ms/dtdl

6. Quality assurance and requirements (RQ-3.4). Digital Twins are subject to common expectation, such as
to high-fidelity representation of the twinned system. Yet, we found few research on quality assurance and
requirements for Digital Twins. Hence, it currently is hardly possible for a Digital Twin to fail in fulfilling
requirements on it. For instance, it is left to investigate how much the fidelity of a Digital Twin may degrade
before its not a (useful) Digital Twin anymore. While the feature models presented in this paper can be a starting
point for exploring such requirements, these also need to build on a common understanding of the concept of
Digital Twins in general.
7. Tool selection support (RQ-3.2). An incredible variety of methods and technologies are used in the develop-
ment of Digital Twins. Identifying which methods and technologies are suitable for which challenges, require-
ments, and Digital Twins goals would facilitate advancing the state-of-the-art in Digital Twins. To this end, the
employed methods and tools used in engineering Digital Twins need to be cross-tabulated against the purposes
the these Digital Twins. Such research could result in a design catalog of technologies to achieve certain effects
with Digital Twins.
To improve our insights into the software engineering for and use of digital twins, future studies on the topic should
consider the evolution of concerns, tools, and methods across time. Moreover, with Digital Twins increasingly being
deployed in various industries, considering including patents or gray literature from industry might yield valuable
insights as well.
Overall, the study presented in this paper sheds light on the state-of-the-art in Digital Twins and on the concerns
related to engineering and operating these for future research to build upon our results and for practitioners to guide
their work.
Primary Studies
[P1] F. Biesinger, D. Meike, B. Kraß, M. Weyrich, A Case Study for a Digital Twin of Body-in-White Production Systems General Concept for
Automated Updating of Planning Projects in the Digital Factory, in: 23rd International Conference on Emerging Technologies and Factory
Automation (ETFA), 2018.
[P2] B. A. Talkhestani, N. Jazdi, W. Schlögl, M. Weyrich, A concept in synchronization of virtual production system with real factory based on
anchor-point method, Procedia CIRP 67 (2018) 13–17 (2018).
[P3] T. Kubota, C. Liu, K. Mubarok, X. Xu, A cyber-physical machine tool framework based on STEP-NC, in: Proceedings of the 48th
International Conference on Computers and Industrial Engineering (CIE), 2018.
[P4] C. Liu, H. Vengayil, Y. Lu, X. Xu, A Cyber-Physical Machine Tools Platform using OPC UA and MTConnect, Journal of Manufacturing
Systems 51 (2019) 61–74 (2019).
[P5] P. Jain, J. Poon, J. Singh, C. Spanos, S. R. Sanders, S. K. Panda, A Digital Twin Approach for Fault Diagnosis in Distributed Photovoltaic
Systems, IEEE Transactions on Power Electronics 35 (2020) 940–956 (2020).
[P6] A. Padovano, F. Longo, L. Nicoletti, G. Mirabelli, A Digital Twin based Service Oriented Application for a 4.0 Knowledge Navigation in
the Smart Factory, IFAC-PapersOnLine 51 (2018) 631–636 (2018).
[P7] W. Ellgass, N. Holt, H. Saldana-Lemus, J. Richmond, A. Vatankhah Barenji, G. Gonzalez-Badillo, A Digital Twin Concept for Manufac-
turing Systems, in: ASME International Mechanical Engineering Congress and Exposition, 2018.
[P8] S. Zambal, C. Eitzinger, M. Clarke, J. Klintworth, P.-Y. Mechin, A digital twin for composite parts manufacturing : Effects of defects
analysis based on manufacturing data, in: 16th International Conference on Industrial Informatics (INDIN), 2018.
[P9] P. Pileggi, J. Verriet, J. Broekhuijsen, C. van Leeuwen, W. Wijbrandi, M. Konsman, A Digital Twin for Cyber-Physical Energy Systems,
in: 7th Workshop on Modeling and Simulation of Cyber-Physical Energy Systems (MSCPES), 2019.
[P10] K. Kannan, N. Arunachalam, A Digital Twin for Grinding Wheel: An Information Sharing Platform for Sustainable Grinding Process,
Journal of Manufacturing Science and Engineering 141 (2018).
[P11] F. Biesinger, D. Meike, B. Kraß, M. Weyrich, A digital twin for production planning based on cyber-physical systems: A Case Study for a
Cyber-Physical System-Based Creation of a Digital Twin, Procedia CIRP 79 (2019) 355–360 (2019).
[P12] T. Mukherjee, T. DebRoy, A digital twin for rapid qualification of 3D printed metallic components, Applied Materials Today 14 (2019).
[P13] A. Detzner, M. Eigner, et al., A digital twin for root cause analysis and product quality monitoring, in: 15th International Design Confer-
ence, 2018.
[P14] J. Sleuters, Y. Li, J. Verriet, M. Velikova, R. Doornbos, A digital twin method for automated behavior analysis of large-scale distributed
IoT systems, in: 14th Annual Conference System of Systems Engineering (SoSE), 2019.
[P15] Y. Lu, X. Xu, A digital twin reference model for smart manufacturing, in: 48th International Conference on Computers and Industrial
Engineering, 2018.
[P16] H. Zhang, Q. Liu, X. Chen, D. Zhang, J. Leng, A Digital Twin-Based Approach for Designing and Multi-Objective Optimization of Hollow
Glass Production Line, IEEE Access 5 (2017) 26901–26911 (2017).

[P17] K. Josifovska, E. Yigitbas, G. Engels, A digital twin-based multi-modal ui adaptation framework for assistance systems in industry 4.0, in:
Human-Computer Interaction. Design Practice in Contemporary Societies, 2019.
[P18] V. Damjanovic-Behrendt, A Digital Twin-based Privacy Enhancement Mechanism for the Automotive Industry, in: 2018 International
Conference on Intelligent Systems (IS), 2018.
[P19] X. Sun, J. Bao, J. Li, Y. Zhang, S. Liu, B. Zhou, A digital twin-driven approach for the assembly-commissioning of high precision products,
Robotics and Computer-Integrated Manufacturing 61 (2020) 101839 (2020).
[P20] Y. Xu, Y. Sun, X. Liu, Y. Zheng, A Digital-Twin-Assisted Fault Diagnosis Using Deep Transfer Learning, IEEE Access 7 (2019) 19990–
19999 (2019).
[P21] M. Dietz, B. Putz, G. Pernul, A distributed ledger approach to digital twin secure data sharing, in: Data and Applications Security and
Privacy XXXIII, 2019.
[P22] C. Li, S. Mahadevan, Y. Ling, L. Wang, S. Choze, A dynamic Bayesian network approach for digital twin, in: 19th AIAA Non-
Deterministic Approaches Conference, 2017.
[P23] D. Shangguan, L. Chen, J. Ding, A Hierarchical Digital Twin Model Framework for Dynamic Cyber-Physical System Design, in: Proceed-
ings of the 5th International Conference on Mechatronics and Robotics Engineering, 2019.
[P24] K. Dröder, P. Bobka, T. Germann, F. Gabriel, F. Dietrich, A Machine Learning-Enhanced Digital Twin Approach for Human-Robot-
Collaboration, Procedia CIRP 76 (2018) 187–192 (2018).
[P25] M. Ciavotta, M. Alge, S. Menato, D. Rovere, P. Pedrazzoli, A microservice-based middleware for the digital factory, Procedia Manufac-
turing 11 (2017) 931–938 (2017).
[P26] N. Lauzeral, D. Borzacchiello, M. Kugler, D. George, Y. Rémond, A. Hostettler, F. Chinesta, A model order reduction approach to
create patient-specific mechanical models of human liver in computational medicine applications, Computer methods and programs in
biomedicine 170 (2019) 95–106 (2019).
[P27] K. Bakliwal, M. H. Dhada, A. S. Palau, A. K. Parlikad, B. K. Lad, A Multi Agent System architecture to implement Collaborative Learning
for social industrial assets, IFAC-PapersOnLine 51 (2018) 1237–1242 (2018).
[P28] G. Avventuroso, M. Silvestri, P. Pedrazzoli., A Networked Production System to Implement Virtual Enterprise and Product Lifecycle
Information Loops, IFAC-PapersOnLine 50 (2017) 7964–7969 (2017).
[P29] T. Okita, T. Kawabata, H. Murayama, N. Nishino, M. Aichi, A new concept of digital twin of artifact systems: synthesizing monitoring/in-
spections, physical/numerical models, and social system models, Procedia CIRP 79 (2019) 667 – 672 (2019).
[P30] Y. Liu, L. Zhang, Y. Yang, L. Zhou, L. Ren, F. Wang, R. Liu, Z. Pang, M. J. Deen, A Novel Cloud-Based Framework for the Elderly
Healthcare Services Using Digital Twin, IEEE Access 7 (2019) 49088–49101 (2019).
[P31] S. Kumar, R. Madhumathi, P. R. Chelliah, L. Tao, S. Wang, A novel digital twin-centric approach for driver intention prediction and traffic
congestion avoidance, journal of Reliable Intelligent Environments 4 (2018).
[P32] R. Allemang, M. Spottswood, T. Eason, A principal component analysis (pca) decomposition based validation metric for use with full field
measurement situations, in: Model Validation and Uncertainty Quantification, volume 3, 2014.
[P33] J.-L. Gregorio, C. Lartigue, F. Thiébaut, H. Falgarone, A Reverse-Engineering Approach for the Management of Product Geometrical
Variations During Assembly, in: Advances on Mechanics, Design Engineering and Manufacturing II, Springer, 2019.
[P34] K. Sivalingam, M. Sepulveda, M. Spring, P. Davies, A Review and Methodology Development for Remaining Useful Life Prediction
of Offshore Fixed and Floating Wind turbine Power Converter with Digital Twin Technology Perspective, in: 2018 2nd International
Conference on Green Energy and Applications (ICGEA), 2018.
[P35] E. L. Grange, A Roadmap for Adopting a Digital Lifecycle Approach to Offshore Oil and Gas Production, in: OTC Offshore Technology
Conference, 2018.
[P36] N. K. Chakshu, J. Carson, I. Sazonov, P. Nithiarasu, A semi-active human digital twin model for detecting severity of carotid stenoses
from head vibration—A coupled computational mechanics and computer vision method, International journal for numerical methods in
biomedical engineering 35 (2019) e3180 (2019).
[P37] T. Gabor, L. Belzner, M. Kiermeier, M. T. Beck, A. Neitz, A Simulation-Based Architecture for Smart Cyber-Physical Systems, in: 2016
IEEE International Conference on Autonomic Computing (ICAC), 2016.
[P38] A. Redelinghuys, A. Basson, K. Kruger, A six-layer digital twin architecture for a manufacturing cell, in: International Workshop on
Service Orientation in Holonic and Multi-Agent Manufacturing, 2018.
[P39] M. Eckhart, A. Ekelhart, A specification-based state replication approach for digital twins, in: Proceedings of the 2018 Workshop on
Cyber-Physical Systems Security and Privacy, 2018.
[P40] P. Zheng, T.-J. Lin, C.-H. Chen, X. Xu, A systematic design approach for service innovation of smart product-service systems, journal of
Cleaner Production 201 (2018) 657–667 (2018).
[P41] R. Wagener, M. Scurria, T. Bein, About a digital twin for the fatigue approach of additively manufactured components, in: TMS 2019
148th Annual Meeting & Exhibition Supplemental Proceedings, 2019.
[P42] R. Rosen, G. von Wichert, G. Lo, K. D. Bettenhausen, About The Importance of Autonomy and Digital Twins for the Future of Manufac-
turing, IFAC-PapersOnLine 48 (2015) 567–572 (2015).
[P43] S. Zobel-Roos, A. Schmidt, F. Mestmäcker, M. Mouellef, M. Huter, L. Uhlenbrock, M. Kornecki, L. Lohmann, R. Ditz, J. Strube, Accel-
erating Biologics Manufacturing by Modeling or: Is Approval under the QbD and PAT Approaches Demanded by authorities Acceptable
without a Digital-Twin?, Processes 7 (2019).
[P44] G. Saini, P. Ashok, E. van Oort, M. R. Isbell, Accelerating Well Construction Using a Digital Twin Demonstrated on Unconventional Well
Data in North America, in: Unconventional Resources Technology Conference, Houston, Texas, 23-25 July 2018, 2018.
[P45] S. Anand, O. Ghalsasi, B. Zhang, A. Goel, S. Reddy, S. Joshi, G. Morris, Additive Manufacturing Simulation Tools in Education, in: 2018
World Engineering Education Forum-Global Engineering Deans Council (WEEF-GEDC), 2018.
[P46] M. F. Milazzoa, P. Bragattob, G. Ancionea, G. Sciontia, Ageing assessment and management at major-hazard industries, Chemical Engi-
neering 67 (2018).
[P47] J. Hatakeyama, D. Seal, D. Farr, S. Haase, Systems Engineering “V” in a Model-Based Engineering Environment: Is it still relevant?, in:

AIAA SPACE and Astronautics Forum and Exposition, 2018.
[P48] Y. Zheng, S. Yang, H. Cheng, An application framework of digital twin and its case study, journal of Ambient Intelligence and Humanized
Computing 10 (2019) 1141–1153 (2019).
[P49] R. Tharma, R. Winter, M. Eigner, et al., An approach for the implementation of the digital twin in the automotive wiring harness field, in:
DS 92: Proceedings of the DESIGN 2018 15th International Design Conference, 2018.
[P50] N. Wantia, J. Roßmann, An online task planning framework reducing execution times in industrial environments, in: 2017 4th International
Conference on Industrial Engineering and Applications (ICIEA), 2017.
[P51] V. Damjanovic-Behrendt, W. Behrendt, An open source approach to the design and implementation of Digital Twins for Smart Manufac-
turing, International journal of Computer Integrated Manufacturing 32 (2019) 366–384 (2019).
[P52] F. Yacob, D. Semere, E. Nordgren, Anomaly detection in Skin Model Shapes using machine learning classifiers, The International journal
of Advanced Manufacturing Technology 105 (2019).
[P53] N. Yusupbekov, F. Abdurasulov, F. Adilov, A. Ivanyan, Application of Cloud Technologies for Optimization of Complex Processes of
Industrial Enterprises, in: International Conference on Theory and Applications of Fuzzy Systems and Soft Computing, 2018.
[P54] E. Sujova, H. ˇ
Cierna, I. ˙
Zabi´
nska, Application of digitization procedures of production in practice, Management Systems in Production
Engineering 27 (2019) 23–28 (2019).
[P55] Y. Tan, W. Yang, K. Yoshida, S. Takakuwa, Application of IoT-Aided Simulation for a Cyber-Physical System, in: Proceedings of the 2018
Winter Simulation Conference, 2018.
[P56] Y. Tan, W. Yang, K. Yoshida, S. Takakuwa, Application of IoT-Aided Simulation to Manufacturing Systems in Cyber-Physical System,
Machines 7 (2019).
[P57] D. Mejia, A. Moreno, O. Ruiz-Salguero, I. Barandiaran, Appraisal of open software for finite element simulation of 2D metal sheet laser
cut, International journal on Interactive Design and Manufacturing (IJIDeM) 11 (2017) 547–558 (2017).
[P58] A. Kychkin, A. Deryabin, O. Vikentyeva, L. Shestakova, Architecture of Compressor Equipment Monitoring and Control Cyber-Physical
System Based on Influxdata Platform, in: 2019 International Conference on Industrial Engineering, Applications and Manufacturing
(ICIEAM), 2019.
[P59] A. Blaga, L. Tamas, Augmented reality for digital manufacturing, in: 2018 26th Mediterranean Conference on Control and Automation
(MED), 2018.
[P60] S. Utzig, R. Kaps, S. M. Azeem, A. Gerndt, Augmented reality for remote collaboration in aircraft maintenance tasks, in: 2019 IEEE
Aerospace Conference, 2019.
[P61] J. Hlady, M. Glanzer, L. Fugate, Automated Creation of the Pipeline Digital Twin During Construction: Improvement to Construction
Quality and Pipeline Integrity, in: International Pipeline Conference, 2018.
[P62] G. S. Martínez, S. Sierla, T. Karhela, V. Vyatkin, Automatic Generation of a Simulation-Based Digital Twin of an Industrial Process Plant,
in: IECON 2018 - 44th Annual Conference of the IEEE Industrial Electronics Society, 2018.
[P63] Y. Kado, K. Katagiri, Autonomous Distributed Power Network Consisting of Triple Active Bridge Converters, in: 2018 Energy and
Sustainability for Small Developing Economies (ES2DE), 2018.
[P64] A. A. Nazarenko, L. M. Camarinha-Matos, Basis for an approach to design collaborative cyber-physical systems, in: Technological
Innovation for Industry and Service Systems, 2019.
[P65] C. Mandolla, A. M. Petruzzelli, G. Percoco, A. Urbinati, Building a digital twin for additive manufacturing through the exploitation of
blockchain: A case analysis of the aircraft industry, Computers in Industry 109 (2019) 134–152 (2019).
[P66] G. Knapp, T. Mukherjee, J. Zuback, H. Wei, T. Palmer, A. De, T. DebRoy, Building blocks for a digital twin of additive manufacturing,
Acta Materialia 135 (2017) 390–399 (2017).
[P67] T. Debroy, W. Zhang, J. Turner, S. S. Babu, Building digital twins of 3D printing machines, Scripta Materialia 135 (2017) 119–124 (2017).
[P68] K. M. Alam, A. El Saddik, C2PS: A Digital Twin Architecture Reference Model for the Cloud-Based Cyber-Physical Systems, IEEE
Access 5 (2017) 2050–2062 (2017).
[P69] M. Trancossi, M. Cannistraro, J. Pascoa, Can constructal law and exergy analysis produce a robust design method that couples with industry
4.0 paradigms? The case of a container house, Mathematical Modelling of Engineering Problems 5 (2018) 303–312 (2018).
[P70] M. Eisenträger, S. Adler, M. Kennel, S. Möser, Changeability in Engineering, in: 2018 IEEE International Conference on Engineering,
Technology and Innovation (ICE/ITMC), 2018.
[P71] B. Gockel, A. Tudor, M. Brandyberry, R. Penmetsa, E. Tuegel, Challenges with structural life forecasting using realistic mission profiles,
in: 53rd AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference 20th AIAA/ASME/AHS Adaptive
Structures Conference 14th AIAA, 2012.
[P72] M. Di Maio, G.-D. Kapos, N. Klusmann, L. Atorf, U. Dahmen, M. Schluse, J. Rossmann, Closed-Loop Systems Engineering (CLOSE):
Integrating Experimentable Digital Twins with the Model-Driven Engineering Process., in: 2018 IEEE International Systems Engineering
Symposium (ISSE), 2018.
[P73] M. Weiss, V. Tuncay, S. Richter, J. Broz, Comprehensive Simulation and Connected Intelligence in Thermal Management Systems, MTZ
worldwide 78 (2017) 36–41 (2017).
[P74] W. V. Mars, J. D. Suter, M. Bauman, Computing Remaining Fatigue Life Under Incrementally Updated Loading Histories, in: WCX World
Congress Experience, 2018.
[P75] B. A. Talkhestani, N. Jazdi, W. Schloegl, M. Weyrich, Consistency check to synchronize the Digital Twin of manufacturing automation
based on anchor points, Procedia CIRP 72 (2018) 159–164 (2018).
[P76] S.-J. Song, Y.-G. Jang, Construction of Digital Twin Geotechnical Resistance Model for Liquefaction Risk Evaluation, in: 2nd International
Symposium on Computer Science and Intelligent Control, 2018.
[P77] K. Rambow-Hoeschele, A. Nagl, D. K. Harrison, B. M. Wood, K. Bozem, K. Braun, P. Hoch, Creation of a digital business model builder,
in: IEEE International Conference on Engineering, Technology and Innovation (ICE/ITMC), 2018.
[P78] E. Frontoni, J. Loncarski, R. Pierdicca, M. Bernardini, M. Sasso, Cyber physical systems for industry 4.0: Towards real time virtual reality
in smart manufacturing, in: Augmented Reality, Virtual Reality, and Computer Graphics, 2018.

[P79] A. R. Al-Ali, R. Gupta, A. A. Nabulsi, Cyber physical systems role in manufacturing technologies, AIP Conference Proceedings 1957
(2018) 050007 (2018).
[P80] F. Arafsha, F. Laamarti, A. El Saddik, Cyber-Physical System Framework for Measurement and Analysis of Physical Activities, Electronics
8 (2019).
[P81] M. Armendia, T. Fuertjes, D. Plakhotnik, J. Sossenheimer, D. Flum, Cyber-Physical System to Improve Machining Process Performance,
in: Twin-Control, Springer, 2019.
[P82] R. Bohlin, J. Hagmar, K. Bengtsson, L. Lindkvist, J. S. Carlson, R. Söderberg, Data Flow and Communication Framework Supporting
Digital Twin for Geometry Assurance, in: International Mechanical Engineering Congress and Exposition, 2017.
[P83] S. Yun, J.-H. Park, W.-T. Kim, Data-centric middleware based digital twin platform for dependable cyber-physical systems, in: 2017 Ninth
International Conference on Ubiquitous and Future Networks (ICUFN), 2017.
[P84] N. Stojanovic, D. Milenovic, Data-driven Digital Twin approach for process optimization: an industry use case, in: 2018 IEEE International
Conference on Big Data (Big Data), 2018.
[P85] R. He, G. Chen, C. Dong, S. Sun, X. Shen, Data-driven digital twin technology for optimized control in process systems, ISA transactions
95 (2019) 221–234 (2019).
[P86] A. Coraddu, L. Oneto, F. Baldi, F. Cipollini, M. Atlar, S. Savio, Data-driven ship digital twin for estimating the speed loss caused by the
marine fouling, Ocean Engineering 186 (2019).
[P87] F. Oquendo, Dealing with Uncertainty in Software Architecture on the Internet-of-Things with Digital Twins, in: International Conference
on Computational Science and Its Applications, 2019.
[P88] R. Dong, C. She, W. Hardjawana, Y. Li, B. Vucetic, Deep learning for hybrid 5G services in mobile edge computing systems: Learn from
a digital twin, IEEE Transactions on Wireless Communications 18 (2019) 4692–4707 (2019).
[P89] K. Ding, F. T. Chan, X. Zhang, G. Zhou, F. Zhang, Defining a Digital Twin-based Cyber-Physical Production System for autonomous
manufacturing in smart shop floors, International journal of Production Research 57 (2019) 6315–6334 (2019).
[P90] J. M. Eyre, T. J. Dodd, C. Freeman, R. Lanyon-Hogg, A. J. Lockwood, R. W. Scott, Demonstration of an industrial framework for an
implementation of a process digital twin, in: ASME 2018 International Mechanical Engineering Congress and Exposition, 2018.
[P91] R. Bitton, T. Gluck, O. Stan, M. Inokuchi, Y. Ohta, Y. Yamada, T. Yagyu, Y. Elovici, A. Shabtai, Deriving a Cost-Effective Digital Twin of
an ICS to Facilitate Security Evaluation, in: Computer Security, 2018.
[P92] K. T. Park, Y. W. Nam, H. S. Lee, S. J. Im, S. D. Noh, J. Y. Son, H. Kim, Design and implementation of a digital twin application for a
connected micro smart factory, International journal of Computer Integrated Manufacturing 32 (2019).
[P93] B. Jeon, S.-H. Suh, Design Considerations and Architecture for Cooperative Smart Factory: MAPE/BD Approach, Procedia Manufacturing
26 (2018) 1094–1106 (2018).
[P94] C. Marty, J. Yu, G. Martin, R. Bornoff, A. Poppe, D. Fournier, E. Morard, Design flow for the development of optimized LED luminaires
using multi-domain compact model simulations, in: 2018 24rd International Workshop on Thermal Investigations of ICs and Systems
(THERMINIC), 2018.
[P95] M. Uzun, M. U. Demirezen, E. Koyuncu, G. Inalhan, Design of a hybrid digital-twin flight performance model through machine learning,
in: 2019 IEEE Aerospace Conference, 2019.
[P96] P. P. Pereverzev, A. V. Akintseva, M. K. Alsigar, D. V. Ardashev, Designing optimal automatic cycles of round grinding based on the
synthesis of digital twin technologies and dynamic programming method, Mechanical Sciences 10 (2019) 331–341 (2019).
[P97] D. Renzi, D. Maniar, S. McNeill, C. Del Vecchio, et al., Developing a digital twin for floating production systems integrity management,
in: OTC Brasil, 2017.
[P98] D. Shcherba, A. Tarasov, A. I. Borovkov, Developing of phenomenological damage model for automotive low-carbon structural steel for
using in validation of euroncap frontal impact, Materials Physics and Mechanics 40 (2018) 246–253 (2018).
[P99] R. A. Khakimov, N. S. Shcherbo, Development and creation a model of a digital twin of the cubepart rectification installation for the
separation of a binary water-alcohol mixture, IOP Conference Series: Materials Science and Engineering 450 (2018).
[P100] A. Carbonari, L. Messi, B. Naticchia, M. Vaccarini, M. Pirani, Development of a BIM-based holonic system for real-time monitoring of
building operational efficiency, Frontiers of Engineering Management 7 (2020) 89–103 (2020).
[P101] C.-S. Shim, N.-S. Dang, S. Lon, C.-H. Jeon, Development of a bridge maintenance system for prestressed concrete bridges using 3D digital
twin model, Structure and Infrastructure Engineering 15 (2019) 1319–1332 (2019).
[P102] Y. Umeda, J. Ota, F. Kojima, M. Saito, H. Matsuzawa, T. Sukekawa, A. Takeuchi, K. Makida, S. Shirafuji, Development of an education
program for digital manufacturing system engineers based on ‘Digital Triplet’ concept, Procedia Manufacturing 31 (2019) 363–369 (2019).
[P103] J. Gomez-Escalonilla, D. Garijo, O. Valencia, I. Rivero, Development of efficient high-fidelity solutions for virtual fatigue testing, in: ICAF
2019 – Structural Integrity in the Age of Additive Manufacturing, 2020.
[P104] X. Chen, E. Kang, S. Shiraishi, V. M. Preciado, Z. Jiang, Digital behavioral twins for safe connected cars, in: Proceedings of the 21th
ACM/IEEE International Conference on Model Driven Engineering Languages and Systems, 2018.
[P105] F. Ameri, R. Sabbagh, Digital factories for capability modeling and visualization, in: Advances in Production Management Systems.
Initiatives for a Sustainable World, 2016.
[P106] A. Thomas, D. A. Guerra-Zubiaga, J. Cohran, Digital Factory: Simulation Enhancing Production and Engineering Process, in: ASME
International Mechanical Engineering Congress and Exposition, 2018.
[P107] C. Block, B. Kuhlenkötter, Digital factory implementation approach starting from the macroscopic perspective with an example for holistic
planning in assembly systems, in: Advances in Production Research, 2019.
[P108] W. N. Dawes, N. Meah, A. Kudryavtsev, R. Evans, M. Hunt, P. Tiller, Digital Geometry to Support a Gas Turbine Digital Twin, in: AIAA
Scitech 2019 Forum, 2019.
[P109] F. Lima, C. N. De Carvalho, M. B. Acardi, E. G. Dos Santos, G. B. De Miranda, R. F. Maia, A. A. Massote, Digital manufacturing tools in
the simulation of collaborative robots: towards industry 4.0, Brazilian journal of Operations & Production Management 16 (2019) 261–280
(2019).
[P110] G. Shao, D. Kibira, Digital Manufacturing: Requirements and Challenges for Implementing Digital Surrogates, in: 2018 Winter Simulation

Conference (WSC), 2018.
[P111] J. V. Zweber, R. Kolonay, P. Kobryn, E. J. Tuegel, Digital thread and Twin for systems engineering: pre-MDD through TMRR, in: 55th
AIAA Aerospace Sciences Meeting, American Institute of Aeronautics and Astronautics, 2017.
[P112] C. Geselschap, G. Meskers, R. V. Dijk, I. V. Winsen, Digital Twin - Engineering with the Human Factor in the Loop, in: Offshore
Technology Conference, 2019.
[P113] S. Kaewunruen, Q. Lian, Digital twin aided sustainability-based lifecycle management for railway turnout systems, journal of Cleaner
Production 228 (2019) 1537–1551 (2019).
[P114] D. Iglesias, P. Bunting, S. Esquembri, J. Hollocombe, S. Silburn, L. Vitton-Mea, I. Balboa, A. Huber, G. Matthews, V. Riccardo, F. Rimini,
D. Valcarcel, Digital twin applications for the JET divertor, Fusion Engineering and Design 125 (2017) 71–76 (2017).
[P115] D. Kostenko, N. Kudryashov, M. Maystrishin, V. Onufriev, V. Potekhin, A. Vasiliev, DIGITAL TWIN APPLICATIONS: DIAGNOSTICS,
OPTIMISATION AND PREDICTION., Annals of DAAAM & Proceedings 29 (2018).
[P116] K. A. Kurniadi, S. Lee, K. Ryu, Digital twin approach for solving reconfiguration planning problems in rms, in: Advances in Production
Management Systems. Smart Manufacturing for Industry 4.0, 2018.
[P117] B. Brenner, V. Hummel, Digital Twin as Enabler for an Innovative Digital Shopfloor Management System in the ESB Logistics Learning
Factory at Reutlingen - University, Procedia Manufacturing 9 (2017) 198–205 (2017).
[P118] V. Kuts, T. Otto, T. Tähemaa, Y. Bondarenko, Digital twin based synchronised control and simulation of the industrial robotic cell using
virtual reality, journal of Machine Engineering 19 (2019).
[P119] Y. Liau, H. Lee, K. Ryu, Digital Twin concept for smart injection molding, IOP Conference Series: Materials Science and Engineering
324 (2018) 012077 (2018).
[P120] G. N. Schroeder, C. Steinmetz, C. E. Pereira, D. B. Espindola, Digital Twin Data Modeling with AutomationML and a Communication
Methodology for Data Exchange, IFAC-PapersOnLine 49 (2016) 12 – 17 (2016).
[P121] I. Raineri, F. La Mura, H. Giberti, Digital Twin Development of HexaFloat, a 6DoF PKM for HIL Tests, in: The International Conference
of IFToMM ITALY, 2018.
[P122] W. Luo, T. Hu, C. Zhang, Y. Wei, Digital twin for CNC machine tool: modeling and using strategy, journal of Ambient Intelligence and
Humanized Computing 10 (2019) 1129–1140 (2019).
[P123] L. Pivano, D. T. Nguyen, K. Bruun Ludvigsen, Digital Twin for Drilling Operations – Towards Cloud-Based Operational Planning, in:
Offshore Technology Conference, 2019.
[P124] N. Karanjkar, A. Joglekar, S. Mohanty, V. Prabhu, D. Raghunath, R. Sundaresan, Digital twin for energy optimization in an SMT-PCB
assembly line, in: 2018 IEEE International Conference on Internet of Things and Intelligence System (IOTAIS), 2018.
[P125] K. Vijayakumar, C. Dhanasekaran, R. Pugazhenthi, S. Sivaganesan, Digital Twin for factory system simulation, International journal of
Recent Technology and Engineering 8 (2019) 63–68 (2019).
[P126] A. Khan, M. Dahl, P. Falkman, M. Fabian, Digital Twin for Legacy Systems: Simulation Model Testing and Validation, in: 2018 IEEE
14th International Conference on Automation Science and Engineering (CASE), 2018.
[P127] Q. Qiao, J. Wang, L. Ye, R. X. Gao, Digital Twin for machining tool condition prediction, Procedia CIRP 81 (2019) 1388–1393 (2019).
[P128] J. Wang, L. Ye, R. X. Gao, C. Li, L. Zhang, Digital Twin for rotating machinery fault diagnosis in smart manufacturing, International
journal of Production Research 57 (2019) 3920–3934 (2019).
[P129] S. Kaewunruen, N. Xu, Digital twin for sustainability evaluation of railway station buildings, Frontiers in Built Environment 4 (2018) 77
(2018).
[P130] X. Fei, Q. Fengchen, S. Bing, F. Yuzhu, Digital Twin of Solid Rocket Motor, Problem and Challenge, in: 2018 11th International Sympo-
sium on Computational Intelligence and Design (ISCID), 2018.
[P131] S. KOSENKOV, V. Y. TURCHANINOV, I. Korovin, D. Y. IVANOV, Digital Twin of the Oil Well, Based on Data Mining Technologies,
in: 2nd International Conference on Modeling, Simulation and Optimization Technologies and Applications (MSOTA), 2018.
[P132] Q. Qi, F. Tao, Y. Zuo, D. Zhao, Digital Twin Service towards Smart Manufacturing, Procedia CIRP 72 (2018) 237–242 (2018).
[P133] F. Tao, M. Zhang, Digital Twin Shop-Floor: A New Shop-Floor Paradigm Towards Smart Manufacturing, IEEE Access 5 (2017) 20418–
20427 (2017).
[P134] R. Anderl, S. Haag, K. Schützer, E. Zancul, Digital twin technology–An approach for Industrie 4.0 vertical and horizontal lifecycle
integration, it-Information Technology 60 (2018) 125–132 (2018).
[P135] T. Brewer, D. Knight, G. Noiray, H. Naik, Digital Twin Technology in the Field Reclaims Offshore Resources, in: Offshore Technology
Conference, 2019.
[P136] K. Yan, W. Xu, B. Yao, Z. Zhou, D. T. Pham, Digital twin-based energy modeling of industrial robots, in: Asian Simulation Conference,
2018.
[P137] R. H. Guerra, R. Quiza, A. Villalonga, J. Arenas, F. Castaño, Digital Twin-Based Optimization for Ultraprecision Motion Systems With
Backlash and Friction, IEEE Access 7 (2019) 93462–93472 (2019).
[P138] J. Liu, H. Zhou, G. Tian, X. Liu, X. Jing, Digital twin-based process reuse and evaluation approach for smart process planning, The
International journal of Advanced Manufacturing Technology 100 (2019) 1619–1634 (2019).
[P139] C. Zhuang, J. Liu, H. Xiong, Digital twin-based smart production management and control framework for the complex product assembly
shop-floor, The International journal of Advanced Manufacturing Technology 96 (2018) 1149–1163 (2018).
[P140] X. V. Wang, L. Wang, Digital twin-based WEEE recycling, recovery and remanufacturing in the background of Industry 4.0, International
journal of Production Research 57 (2019) 3892–3902 (2019).
[P141] S. Meng, S. Tang, Y. Zhu, C. Chen, Digital twin-driven control method for robotic automatic assembly system, in: IOP Conference Series:
Materials Science and Engineering, 2019.
[P142] H. Zhang, G. Zhang, Q. Yan, Digital twin-driven cyber-physical production system towards smart shop-floor, journal of Ambient Intelli-
gence and Humanized Computing 10 (2019) 4439–4453 (2019).
[P143] R. Zhao, D. Yan, Q. Liu, J. Leng, J. Wan, X. Chen, X. Zhang, Digital Twin-Driven Cyber-Physical System for Autonomously Controlling
of Micro Punching System, IEEE Access 7 (2019) 9459–9469 (2019).

[P144] J. Leng, H. Zhang, D. Yan, Q. Liu, X. Chen, D. Zhang, Digital twin-driven manufacturing cyber-physical system for parallel controlling of
smart workshop, journal of ambient intelligence and humanized computing 10 (2019) 1155–1166 (2019).
[P145] F. Tao, F. Sui, A. Liu, Q. Qi, M. Zhang, B. Song, Z. Guo, S. C.-Y. Lu, A. Y. C. Nee, Digital twin-driven product design framework,
International journal of Production Research 57 (2019) 3935–3953 (2019).
[P146] F. Tao, J. Cheng, Q. Qi, M. Zhang, H. Zhang, F. Sui, Digital twin-driven product design, manufacturing and service with big data, The
International journal of Advanced Manufacturing Technology 94 (2018) 3563–3576 (2018).
[P147] Q. Liu, H. Zhang, J. Leng, X. Chen, Digital twin-driven rapid individualised designing of automated flow-shop manufacturing system,
International journal of Production Research 57 (2019) 3903–3919 (2019).
[P148] M. Ayani, M. Ganebäck, A. H. Ng, Digital Twin: Applying emulation for machine reconditioning, Procedia CIRP 72 (2018) 243–248
(2018).
[P149] R. Lu, I. Brilakis, Digital twinning of existing reinforced concrete bridges from labelled point clusters, Automation in Construction 105
(2019) 102837 (2019).
[P150] F. Tao, Q. Qi, L. Wang, A. Nee, Digital Twins and Cyber–Physical Systems toward Smart Manufacturing and Industry 4.0: Correlation
and Comparison, Engineering 5 (2019) 653–661 (2019).
[P151] J. Kaivo-oja, O. Kuusi, M. S. Knudsen, T. Lauraeus, Digital Twins Approach and Future Knowledge Management Challenges: Where We
Shall Need System Integration, Synergy Analyses and Synergy Measurements?, in: International Conference on Knowledge Management
in Organizations, 2019.
[P152] L. Rauch, M. Pietrzyk, Digital twins as a modern approach to design of industrial processes, journal of Machine Engineering 19 (2019).
[P153] W. Kuehn, Digital twins for decision making in complex production and logistic enterprises, International journal of Design & Nature and
Ecodynamics 13 (2018) 260–271 (2018).
[P154] H. Brandtstaedter, C. Ludwig, L. Hübner, E. Tsouchnika, A. Jungiewicz, U. Wever, DIGITAL TWINS FOR LARGE ELECTRIC DRIVE
TRAINS, in: 2018 Petroleum and Chemical Industry Conference Europe (PCIC Europe), 2018.
[P155] K. Bruynseels, F. Santoni de Sio, J. van den Hoven, Digital twins in health care: ethical implications of an emerging engineering paradigm,
Frontiers in genetics 9 (2018) 31 (2018).
[P156] H.-D. Wuttke, K. Henke, R. Hutschenreuter, Digital Twins in Remote Labs, in: International Conference on Remote Engineering and
Virtual Instrumentation, 2019.
[P157] A. A. Malik, A. Bilberg, Digital twins of human robot collaboration in a production setting, Procedia Manufacturing 17 (2018) 278–285
(2018).
[P158] F. Jaensch, A. Csiszar, C. Scheifele, A. Verl, Digital Twins of Manufacturing Systems as a Base for Machine Learning, in: 2018 25th
International Conference on Mechatronics and Machine Vision in Practice (M2VIP), 2018.
[P159] A. A. Malozemov, V. N. Bondar, V. V. Egorov, G. A. Malozemov, Digital twins technology for internal combustion engines development,
in: 2018 Global Smart Industry Conference (GloSIC), 2018.
[P160] F. Caputo, A. Greco, M. Fera, R. Macchiaroli, Digital twins to enhance the integration of ergonomics in the workplace design, International
journal of Industrial Ergonomics 71 (2019) 20–31 (2019).
[P161] T. Cichon, J. Roßmann, Digital Twins: Assisting and Supporting Cooperation in Human-Robot Teams, in: 2018 15th International Con-
ference on Control, Automation, Robotics and Vision (ICARCV), 2018.
[P162] A. El Saddik, Digital twins: The convergence of multimedia technologies, IEEE multimedia 25 (2018) 87–92 (2018).
[P163] A. A. Malik, A. Bilberg, Digital twins of human robot collaboration in a production setting, Procedia Manufacturing 17 (2018) 278–285
(2018).
[P164] Y. Shi, J. Xu, W. Du, Discussion on the New Operation Management Mode of Hydraulic Engineering Based on the Digital Twin Technique,
in: journal of Physics: Conference Series, 2019.
[P165] D. Nadhan, M. G. Mayani, R. Rommetveit, Drilling with Digital Twins, in: IADC/SPE Asia Pacific Drilling Technology Conference and
Exhibition, 2018.
[P166] J. Liu, H. Zhou, X. Liu, G. Tian, M. Wu, L. Cao, W. Wang, Dynamic Evaluation Method of Machining Process Planning Based on Digital
Twin, IEEE Access (2019).
[P167] H. Zhang, G. Zhang, Q. Yan, Dynamic resource allocation optimization for digital twin-driven smart shopfloor, in: 2018 IEEE 15th
International Conference on Networking, Sensing and Control (ICNSC), 2018.
[P168] G. Wiegand, C. Mai, Y. Liu, H. Hußmann, Early Take-Over Preparation in Stereoscopic 3D, in: 10th International Conference on Auto-
motive User Interfaces and Interactive Vehicular Applications, 2018.
[P169] A. Peuhkurinen, T. Mikkonen, Embedding web apps in mixed reality, in: 2018 Third International Conference on Fog and Mobile Edge
Computing (FMEC), 2018.
[P170] O. J. Pinon, D. J. Siedlak, D. Mavris, Enabling the Digital Factory through the Integration of Data-Driven and Simulation Models, in: 31st
Congress of the International Council of the Aeronautical Sciences (ICAS), 2018.
[P171] J. Kraft, S. Kuntzagk, Engine Fleet-Management: The Use of Digital Twins From a MRO Perspective, in: Turbo Expo: Power for Land,
Sea, and Air, 2017.
[P172] M. Zhang, Y. Zuo, F. Tao, Equipment energy consumption management in digital twin shop-floor: A framework and potential applications,
in: 2018 IEEE 15th International Conference on Networking, Sensing and Control (ICNSC), 2018.
[P173] D. Dupláková, M. Flimel, J. Duplák, M. Hatala, S. Radchenko, F. Botko, Ergonomic rationalization of lighting in the working environment.
Part I.: Proposal of rationalization algorithm for lighting redesign, International journal of Industrial Ergonomics 71 (2019) 92–102 (2019).
[P174] Z. Tian, S. Gregson, Examination of the Effectiveness of Mode Orthogonalisation and Filtering for Scattering Suppression in Antenna
Measurements Through Computational Electromagnetic Simulation, in: 2019 13th European Conference on Antennas and Propagation
(EuCAP), 2019.
[P175] U. Dahmen, J. Rossmann, Experimentable Digital Twins for a Modeling and Simulation-based Engineering Approach, in: 2018 IEEE
International Systems Engineering Symposium (ISSE), 2018.
[P176] M. Schluse, M. Priggemeyer, L. Atorf, J. Rossmann, Experimentable digital twins—Streamlining simulation-based systems engineering

for industry 4.0, IEEE Transactions on Industrial Informatics 14 (2018) 1722–1731 (2018).
[P177] V. Kuts, G. E. Modoni, W. Terkaj, T. Tähemaa, M. Sacco, T. Otto, Exploiting factory telemetry to support virtual reality simulation in
robotics cell, in: Augmented Reality, Virtual Reality, and Computer Graphics, 2017.
[P178] A. Bekker, Exploring the blue skies potential of digital twin technology for a polar supply and research vessel, in: Proceedings of the 13th
International Marine Design Conference Marine Design XIII (IMDC 2018), 2018.
[P179] M. Peruzzini, F. Grandi, M. Pellicciari, Exploring the potential of Operator 4.0 interface and monitoring, Computers & Industrial Engi-
neering 139 (2020) 105600 (2020).
[P180] I. Verner, M. Reitman, D. Cuperman, T. Yan, E. Finkelstein, T. Romm, Exposing robot learning to students in augmented reality experience,
in: Smart Industry & Smart Education, 2019.
[P181] F. Laborie, O. C. Røed, G. Engdahl, A. Camp, Extracting Value from Data Using an Industrial Data Platform to Provide a Foundational
Digital Twin, in: Offshore Technology Conference, 2019.
[P182] V. Zhidchenko, I. Malysheva, H. Handroos, A. Kovartsev, Faster than real-time simulation of mobile crane dynamics using digital twin
concept, journal of Physics: Conference Series 1096 (2018).
[P183] V. Zaccaria, M. Stenfelt, I. Aslanidou, K. G. Kyprianidis, Fleet monitoring and diagnostics framework based on digital twin of aero-engines,
in: Turbo Expo: Power for Land, Sea, and Air, 2018.
[P184] A. V. Gurjanov, D. A. Zakoldaev, A. V. Shukalov, I. O. Zharinov, Formation principles of digital twins of Cyber-Physical Systems in the
smart factories of Industry 4.0, in: IOP Conference Series: Materials Science and Engineering, 2019.
[P185] J. Ríos, F. M. Morate, M. Oliva, J. C. Hernández, Framework to support the aircraft digital counterpart concept with an industrial design
view, International journal of Agile Systems and Management 9 (2016) 212–231 (2016).
[P186] M. Schluse, J. Rossmann, From simulation to experimentable digital twins: Simulation-based development and operation of complex
technical systems, in: 2016 IEEE International Symposium on Systems Engineering (ISSE), 2016.
[P187] Y. He, J. Guo, X. Zheng, From surveillance to digital twin: Challenges and recent advances of signal processing for industrial internet of
things, IEEE Signal Processing Magazine 35 (2018) 120–129 (2018).
[P188] G. Grinshpun, T. Cichon, D. Dipika, J. Rossmann, From Virtual Testbeds to Real Lightweight Robots: Development and deployment of
control algorithms for soft robots, with particular reference to, in: Proceedings of ISR 2016: 47st International Symposium on Robotics,
2016.
[P189] J. Goldmeer, W. York, P. Glaser, Fuel and Combustion System Capabilities of GE’s F and HA Class Gas Turbines, in: Turbo Expo: Power
for Land, Sea, and Air, 2017.
[P190] B. Schleich, K. Wärmefjord, R. Söderberg, S. Wartzack, Geometrical Variations Management 4.0: towards next Generation Geometry
Assurance, Procedia CIRP 75 (2018) 3–10 (2018).
[P191] A. Fricke, H. Asche, Geospatial database for the generation of multidimensional virtual city models dedicated to urban analysis and
decision-making, in: Computational Science and Its Applications – ICCSA 2019, 2019.
[P192] Z. Forgo, A. Hypki, B. Kuhlenkoetter, Gesture based robot programming using ROS platform, in: ISR 2018; 50th International Symposium
on Robotics, 2018.
[P193] G. Horváth, G. Erd ˝
os, Gesture Control of Cyber Physical Systems, Procedia CIRP 63 (2017) 184–188 (2017).
[P194] C. Dufour, Z. Soghomonian, W. Li, Hardware-in-the-loop testing of modern on-board power systems using digital twins, in: 2018 Interna-
tional Symposium on Power Electronics, Electrical Drives, Automation and Motion (SPEEDAM), 2018.
[P195] A. K. Ghosh, S. U. AMM, A. Kubo, Hidden Markov model-based digital twin construction for futuristic manufacturing systems, Artificial
Intelligence for Engineering Design, Analysis and Manufacturing: AI EDAM 33 (2019) 317–331 (2019).
[P196] Y. Wang, M. Dittmann, R. Anderl, Holistic scenarios by using platform technologies for small batch-sized production, in: Advances in
Neuroergonomics and Cognitive Engineering, 2019.
[P197] T. Stachowski, H. Kjeilen, Holistic ship design—How to utilise a digital twin in concept design through basic and detailed design, in:
International Conference on Computer Applications in Shipbuilding, 2017.
[P198] D. Morais, M. Waldie, et al., How to implement tech in shipbuilding: Charting the course to success, in: SNAME Maritime Convention,
2018.
[P199] C. Savur, S. Kumar, S. Arora, T. Hazbar, F. Sahin, HRC-SoS: Human robot collaboration experimentation platform as system of systems,
in: 2019 14th Annual Conference System of Systems Engineering (SoSE), 2019.
[P200] T. Petkovi´
c, D. Puljiz, I. Markovi´
c, B. Hein, Human intention estimation based on hidden Markov model motion validation for safe flexible
robotized warehouses, Robotics and Computer-Integrated Manufacturing 57 (2019) 182–196 (2019).
[P201] E. Y. Song, M. Burns, A. Pandey, T. Roth, IEEE 1451 Smart Sensor Digital Twin Federation for IoT/CPS Research, in: IEEE Sensors
Applications Symposium (SAS), 2019.
[P202] P. Dziurzanski, J. Swan, L. S. Indrusiak, J. M. Ramos, Implementing Digital Twins of Smart Factories with Interval Algebra, in: IEEE
International Conference on Industrial Technology (ICIT), 2019.
[P203] S. Yun, J.-h. Park, H.-s. Kim, W.-T. Kim, Importance-aware SDN Control Mechanism for Real-time Data Distribution Services, in: 2018
International Conference on Information and Communication Technology Convergence (ICTC), 2018.
[P204] P. Pereverzev, A. Akintseva, M. Alsigar, Improvement of the quality of designed cylindrical grinding cycle with traverse feeding based on
the use of digital twin options, in: MATEC Web of Conferences, 2018.
[P205] H. Y. Jeon, C. Justin, D. N. Mavris, Improving Prediction Capability of Quadcopter Through Digital Twin, in: AIAA Scitech 2019 Forum,
2019.
[P206] N. Teslya, Industrial socio-cyberphysical system’s consumables tokenization for smart contracts in blockchain, in: Business Information
Systems Workshops, 2019.
[P207] J. Vatn, Industry 4.0 and real-time synchronization of operation and maintenance, in: Safety and Reliability-Safe Societies in a Changing
World-Proceedings of the 28th International European Safety and Reliability Conference, ESREL 2018, 2018.
[P208] J. Reitz, M. Schluse, J. Roßmann, Industry 4.0 beyond the factory: An application to forestry, in: Tagungsband des 4. Kongresses Montage
Handhabung Industrieroboter, 2019.

[P209] C. Moussa, K. Ai-Haddad, B. Kedjar, A. Merkhouf, Insights into Digital Twin Based on Finite Element Simulation of a Large Hydro
Generator, in: IECON 2018 - 44th Annual Conference of the IEEE Industrial Electronics Society, 2018.
[P210] K. Wärmefjord, R. Söderberg, L. Lindkvist, B. Lindau, J. S. Carlson, Inspection data to support a digital twin for geometry assurance, in:
ASME international mechanical engineering congress and exposition, 2017.
[P211] W. D. Lin, Y. H. Low, Y. T. Chong, C. L. Teo, Integrated Cyber Physical Simulation Modelling Environment for Manufacturing 4.0, in:
2018 IEEE International Conference on Industrial Engineering and Engineering Management (IEEM), 2018.
[P212] M. Kunath, H. Winkler, Integrating the Digital Twin of the manufacturing system into a decision support system for improving the order
management process, Procedia CIRP 72 (2018) 225–231 (2018).
[P213] I. Graessler, A. Poehler, Integration of a digital twin as human representation in a scheduling procedure of a cyber-physical production
system, in: 2017 IEEE International Conference on Industrial Engineering and Engineering Management (IEEM), 2017.
[P214] H. Pargmann, D. Euhausen, R. Faber, Intelligent big data processing for wind farm monitoring and analysis based on cloud-technologies
and digital twins: A quantitative approach, in: 2018 IEEE 3rd International Conference on Cloud Computing and Big Data Analysis
(ICCCBDA), 2018.
[P215] I. Graessler, A. Poehler, Intelligent control of an assembly station by integration of a digital twin for employees into the decentralized
control system, Procedia Manufacturing 24 (2018) 185–189 (2018).
[P216] M. Krajcovic, P. Grznar, M. Fusko, R. Skokan, Intelligent Logistics for Intelligent Production Systems, Communications - Scientific letters
of the University of Zilina 20 (2018) 16–23 (2018).
[P217] E. Kosicka, E. Kozłowski, D. Mazurkiewicz, Intelligent systems of forecasting the failure of machinery park and supporting fulfilment of
orders of spare parts, in: Intelligent Systems in Production Engineering and Maintenance, 2018.
[P218] A. Dingli, F. Haddod, Interacting with Intelligent Digital Twins, in: International Conference on Human-Computer Interaction, 2019.
[P219] L. Atorf, J. Roßmann, Interactive Analysis and Visualization of Digital Twins in High-Dimensional State Spaces, in: 2018 15th Interna-
tional Conference on Control, Automation, Robotics and Vision (ICARCV), 2018.
[P220] M. Priggemeyer, D. Losch, J. Roβmann, Interactive Calibration and Visual Programming of Reconfigurable Robotic Workcells, in: 2018
IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM), 2018.
[P221] C. Steinmetz, A. Rettberg, F. G. C. Ribeiro, G. Schroeder, C. E. Pereira, Internet of Things Ontology for Digital Twin in Cyber Physical
Systems, in: VIII Brazilian Symposium on Computing Systems Engineering (SBESC), 2018.
[P222] M. Bartelt, B. Kuhlenkötter, Involving the Manufacturing System within its Planning Phase, in: 2018 IEEE International Conference on
Industrial Engineering and Engineering Management (IEEM), 2018.
[P223] T. D. West, M. Blackburn, Is Digital Thread/Digital Twin Affordable? A Systemic Assessment of the Cost of DoD’s Latest Manhattan
Project, Procedia Computer Science 114 (2017) 47–56 (2017).
[P224] H. Zipper, F. Auris, A. Strahilov, M. Paul, Keeping the digital twin up-to-date — Process monitoring to identify changes in a plant, in:
IEEE International Conference on Industrial Technology (ICIT), 2018.
[P225] M. Ringsquandl, S. Lamparter, R. Lepratti, P. Kröger, Knowledge fusion of manufacturing operations data using representation learning,
in: Advances in Production Management Systems. The Path to Intelligent, Collaborative and Sustainable Manufacturing, 2017.
[P226] G. Zhou, C. Zhang, Z. Li, K. Ding, C. Wang, Knowledge-driven digital twin manufacturing cell towards intelligent manufacturing, Inter-
national journal of Production Research 58 (2020) 1034–1051 (2020).
[P227] A. M. Madni, C. C. Madni, S. D. Lucero, Leveraging Digital Twin Technology in Model-Based Systems Engineering, Systems 7 (2019).
[P228] J. David, A. Lobov, M. Lanz, Leveraging Digital Twins for Assisted Learning of Flexible Manufacturing Systems, in: 16th International
Conference on Industrial Informatics (INDIN), 2018.
[P229] S. R. Islavath, D. Deb, H. Kumar, Life cycle analysis and damage prediction of a longwall powered support using 3D numerical modelling
techniques, Arabian journal of Geosciences 12 (2019) 1–15 (2019).
[P230] C. E. Kaed, V. Danilchenko, F. Delpech, J. Brodeur, A. Radisson, Linking an Asset and a Domain Specific Ontology for a Simple Asset
TimeSeries Application, in: 2018 IEEE International Conference on Big Data (Big Data), 2018.
[P231] G. Martin, C. Marty, R. Bornoff, A. Poppe, G. Onushkin, M. Rencz, J. Yu, Luminaire Digital Design Flow with Multi-Domain Digital
Twins of LEDs, Energies 12 (2019).
[P232] C. Weber, J. Königsberger, L. Kassner, B. Mitschang, M2DDM – A Maturity Model for Data-Driven Manufacturing, Procedia CIRP 63
(2017) 173–178 (2017).
[P233] S. Settemsdal, Machine Learning and Artificial Intelligence as a Complement to Condition Monitoring in a Predictive Maintenance Setting,
in: SPE Oil and Gas India Conference and Exhibition, 2019.
[P234] Q. Min, Y. Lu, Z. Liu, C. Su, B. Wang, Machine Learning based Digital Twin Framework for Production Optimization in Petrochemical
Industry, International journal of Information Management 49 (2019) 502–519 (2019).
[P235] C. Liu, X. Hong, Z. Zhu, X. Xu, Machine tool digital twin: modelling methodology and applications, in: 48th International Conference on
Computers and Industrial Engineering (CIE), 2018.
[P236] K. Ponomarev, N. Kudryashov, N. Popelnukha, V. Potekhin, Main Principals and Issues of Digital Twin Development for Complex Tech-
nological Processes, Annals of DAAAM & Proceedings 28 (2017).
[P237] S. Waschull, J. C. Wortmann, J. A. C. Bokhorst, Manufacturing execution systems: The next level of automated control or of shop-floor
support?, in: Advances in Production Management Systems. Smart Manufacturing for Industry 4.0, 2018.
[P238] J. Kim, H. Kim, Y. Ham, Mapping local vulnerabilities into a 3D city model through social sensing and the CAVE system toward digital
twin city, in: Computing in Civil Engineering 2019: Smart Cities, Sustainability, and Resilience, American Society of Civil Engineers,
2019.
[P239] C. L. Popa, C. E. Cotet, D. Popescu, M. F. Solea, S. G. ¸Sa¸scîm, T. Dobrescu, Material flow design and simulation for a glass panel recycling
installation, Waste Management & Research 36 (2018) 653–660 (2018).
[P240] P. Janda, Mechatronic concept of heavy machine tools, in: International DAAAM Symposium, 2018.
[P241] S. Gordon, A. Ryan, S. Loughlin, Meeting The Needs of Industry in Smart Manufacture – The Definition of a New Profession and a Case
Study in Providing the Required Skillset, Procedia Manufacturing 17 (2018) 262–269 (2018).

[P242] M. P. Papazoglou, Metaprogramming Environment for Industry 4.0, in: 2018 Sixth International Conference on Enterprise Systems (ES),
2018.
[P243] P. Aivaliotis, K. Georgoulias, Z. Arkouli, S. Makris, Methodology for enabling Digital Twin using advanced physics-based modelling in
predictive maintenance, Procedia CIRP 81 (2019) 417–422 (2019).
[P244] C. Constantinescu, D. Popescu, O. Todorovic, O. Virlan, V. Tinca, Methodology of Realising the Digital Twins of Exoskeleton-Centered
Workplaces, Acta Technica Napocensis-Series: Applied Mathematics, Mechanics, and Engineering 61 (2018).
[P245] Naplekov, Ilya, Zheleznikov, Ivan, Pashchenko, Dmitry, Kobysheva, Polina, Moskvitina, Anna, Mustafin, Ravil, Gnutikova, Maria, Mulla-
galieva, Alina, Uzlov, Pavel, Methods of computational modeling of coronary heart vessels for its digital twin, MATEC Web of Conferences
172 (2018).
[P246] M. Lohtander, E. Garcia, M. Lanz, J. Volotinen, J. Ratava, J. Kaakkunen, Micro Manufacturing Unit – Creating Digital Twin Objects with
Common Engineering Software, Procedia Manufacturing 17 (2018) 468–475 (2018).
[P247] M. Lohtander, N. Ahonen, M. Lanz, J. Ratava, J. Kaakkunen, Micro manufacturing unit and the corresponding 3D-model for the digital
twin, Procedia manufacturing 25 (2018) 55–61 (2018).
[P248] G. Radchenko, A. Alaasam, A. Tchernykh, Micro-workflows: Kafka and kepler fusion to support digital twins of industrial processes, in:
2018 IEEE/ACM International Conference on Utility and Cloud Computing Companion (UCC Companion), 2018.
[P249] E. Abisset-Chavanne, J. L. Duval, E. Cueto, F. Chinesta, Model and system learners, optimal process constructors and kinetic theory-based
goal-oriented design: a new paradigm in materials and processes informatics, in: AIP Conference Proceedings, Vol. 1960, 2018.
[P250] M. Gonzalez, O. Salgado, J. Croes, B. Pluymers, W. Desmet, Model-based state estimation for the diagnosis of multiple faults in non-linear
electro-mechanical systems, in: International Conference on Condition Monitoring of Machinery in Non-Stationary Operation, 2018.
[P251] A. S. Ullah, Modeling and simulation of complex manufacturing phenomena using sensor signals from the perspective of Industry 4.0,
Advanced Engineering Informatics 39 (2019) 1–13 (2019).
[P252] L. Hu, N.-T. Nguyen, W. Tao, M. C. Leu, X. F. Liu, M. R. Shahriar, S. M. N. Al Sunny, Modeling of Cloud-Based Digital Twins for Smart
Manufacturing with MT Connect, Procedia Manufacturing 26 (2018) 1193–1203 (2018).
[P253] Q. Qi, D. Zhao, T. W. Liao, F. Tao, Modeling of Cyber-Physical Systems and Digital Twin Based on Edge Computing, Fog Computing and
Cloud Computing Towards Smart Manufacturing, in: ASME 2018 13th International Manufacturing Science and Engineering Conference,
2018, p. 7.
[P254] Q. Zhang, X. Zhang, W. Xu, A. Liu, Z. Zhou, D. T. Pham, Modeling of digital twin workshop based on perception data, in: Intelligent
Robotics and Applications, 2017.
[P255] J. Zenisek, J. Wolfartsberger, C. Sievi, M. Affenzeller, Modeling sensor networks for predictive maintenance, in: On the Move to Mean-
ingful Internet Systems: OTM 2018 Workshops, 2019.
[P256] J. Um, J. Popper, M. Ruskowski, Modular augmented reality platform for smart operator in production environment, in: 2018 IEEE
Industrial Cyber-Physical Systems (ICPS), 2018.
[P257] J. Guo, N. Zhao, L. Sun, S. Zhang, Modular based flexible digital twin for factory design, journal of Ambient Intelligence and Humanized
Computing 10 (2019) 1189–1200 (2019).
[P258] M. Vathoopan, M. Johny, A. Zoitl, A. Knoll, Modular Fault Ascription and Corrective Maintenance Using a Digital Twin, IFAC-
PapersOnLine 51 (2018) 1041–1046 (2018).
[P259] M. R. Shahriar, S. M. N. A. Sunny, X. Liu, M. C. Leu, L. Hu, N.-T. Nguyen, MTComm Based Virtualization and Integration of Physical
Machine Operations with Digital-Twins in Cyber-Physical Manufacturing Cloud, in: 5th IEEE International Conference on Cyber Security
and Cloud Computing (CSCloud)/4th IEEE International Conference on Edge Computing and Scalable Cloud (EdgeCom), 2018.
[P260] A. Poppe, G. Farkas, L. Gaál, G. Hantos, J. Hegedüs, M. Rencz, Multi-Domain Modelling of LEDs for Supporting Virtual Prototyping of
Luminaires, Energies 12 (2019).
[P261] T. Delbrügger, M. Meißner, A. Wirtz, D. Biermann, J. Myrzik, J. Rossmann, P. Wiederkehr, Multi-level simulation concept for multidis-
ciplinary analysis and optimization of production systems, The International journal of Advanced Manufacturing Technology 103 (2019)
3993–4012 (2019).
[P262] D. Hauf, S. Süß, A. Strahilov, J. Franke, Multifunctional use of functional mock-up units for application in production engineering, in:
2017 IEEE 15th International Conference on Industrial Informatics (INDIN), 2017.
[P263] N. Demkovich, E. Yablochnikov, G. Abaev, Multiscale modeling and simulation for industrial cyber-physical systems, in: 2018 IEEE
Industrial Cyber-Physical Systems (ICPS), 2018.
[P264] K. Worden, E. J. Cross, P. Gardner, R. J. Barthorpe, D. J. Wagg, On digital twins, mirrors and virtualisations, in: Model Validation and
Uncertainty Quantification, volume 3, 2020.
[P265] D. J. Wagg, P. Gardner, R. J. Barthorpe, K. Worden, On key technologies for realising digital twins for structural dynamics applications,
in: Model Validation and Uncertainty Quantification, volume 3, 2020.
[P266] M. Joordens, M. Jamshidi, On The Development of Robot Fish Swarms in Virtual Reality with Digital Twins, in: 13th Annual Conference
on System of Systems Engineering (SoSE), 2018.
[P267] M. Rückert, S. Merkelbach, R. Alt, K. Schmitz, Online life cycle assessment for fluid power manufacturing systems – challenges and
opportunities, in: Advances in Production Management Systems. Smart Manufacturing for Industry 4.0, 2018.
[P268] N. Teslya, I. Ryabchikov, Ontology-based semantic models for industrial iot components representation, in: Proceedings of the Third
International Scientific Conference “Intelligent Information Technologies for Industry” (IITI’18), 2019.
[P269] S. V. Nagar, A. C. Chandrashekar, M. Suvarna, Optimized additive manufacturing technology using digital twins and cyber physical
systems, in: Cyber-physical Systems and Digital Twins, 2020.
[P270] H. Sun, C. Li, X. Fang, H. Gu, Optimized throughput improvement of assembly flow line with digital twin online analytics, in: IEEE
International Conference on Robotics and Biomimetics (ROBIO), 2017.
[P271] F. Yao, A. Keller, M. Ahmad, B. Ahmad, R. Harrison, A. W. Colombo, Optimizing the Scheduling of Autonomous Guided Vehicle in a
Manufacturing Process, in: IEEE 16th International Conference on Industrial Informatics (INDIN), 2018.
[P272] P. D. Urbina Coronado, R. Lynn, W. Louhichi, M. Parto, E. Wescoat, T. Kurfess, Part data integration in the Shop Floor Digital Twin:

Mobile and cloud technologies to enable a manufacturing execution system, journal of Manufacturing Systems 48 (2018) 25–33 (2018).
[P273] A. Karakra, F. Fontanili, E. Lamine, J. Lamothe, A. Taweel, Pervasive computing integrated discrete event simulation for a hospital digital
twin, in: IEEE/ACS 15th International Conference on Computer Systems and Applications (AICCSA), 2018.
[P274] E. Lutters, Pilot production environments driven by digital twins, South African journal of industrial engineering 29 (2018) 40–53 (2018).
[P275] K. Shubenkova, A. Valiev, V. Shepelev, S. Tsiulin, K. H. Reinau, Possibility of Digital Twins Technology for Improving Efficiency of the
Branded Service System, in: Global Smart Industry Conference (GloSIC), 2018.
[P276] A. Mayes, J. Heffernan, L. Jauriqui, R. Livings, E. Biedermann, J. C. Aldrin, B. R. Goodlet, S. Mazdiyasni, Process compensated resonance
testing (PCRT) inversion for material characterization and digital twin calibration, in: AIP Conference Proceedings, 2019.
[P277] N. Desai, S. K. Ananya, L. Bajaj, A. Periwal, S. R. Desai, Process parameter monitoring and control using digital twin, in: Cyber-physical
Systems and Digital Twins, 2020.
[P278] N. Alaei, A. Rouvinen, A. Mikkola, R. Nikkilä, Product processes based on digital twin, in: Commercial Vehicle Technology 2018,
Springer, 2018.
[P279] B. R. Denos, S. G. Kravchenko, R. B. Pipes, B. Pipes, Progressive failure analysis in platelet based composites using CT-measured local
microstructure, in: International SAMPE Technical Conference, 2017.
[P280] H. Laaki, Y. Miche, K. Tammi, Prototyping a Digital Twin for Real Time Remote Control Over Mobile Networks: Application of Remote
Surgery, IEEE Access 7 (2019) 20325–20336 (2019).
[P281] S. R. Chhetri, S. Faezi, A. Canedo, M. A. A. Faruque, QUILT: Quality Inference from Living Digital Twins in IoT-Enabled Manufacturing
Systems, in: Proceedings of the International Conference on Internet of Things Design and Implementation, 2019.
[P282] P. Sharma, D. Knezevic, P. Huynh, G. Malinowski, et al., RB-FEA based digital twin for structural integrity assessment of offshore
structures, in: Offshore Technology Conference, 2018.
[P283] V. Gopinath, A. Srija, C. N. Sravanthi, Re-design of smart homes with digital twins, Journal of Physics: Conference Series 1228 (2019).
[P284] K. Samir, A. Maffei, M. A. Onori, Real-Time asset tracking; a starting point for Digital Twin implementation in Manufacturing, Procedia
CIRP 81 (2019) 719–723 (2019).
[P285] C. Scheifele, A. Verl, O. Riedel, Real-time co-simulation for the virtual commissioning of production systems, Procedia CIRP 79 (2019)
397–402 (2019).
[P286] E. J. Tuegel, A. R. Ingraffea, T. G. Eason, S. M. Spottswood, Reengineering aircraft structural life prediction using a digital twin, Interna-
tional Journal of Aerospace Engineering 2011 (2011).
[P287] J. Boutrot, Y. Giorgiutti, F. Rezende, S. Barras, Reliable and Accurate Determination of Life Extension for Offshore Units, in: OTC
Offshore Technology Conference, 2017.
[P288] S. Balachandar, R. Chinnaiyan, Reliable digital twin for connected footballer, in: International conference on computer networks and
communication technologies, 2019.
[P289] T. Delbrügger, J. Rossmann, Representing adaptation options in experimentable digital twins of production systems, International Journal
of Computer Integrated Manufacturing 32 (2019) 352–365 (2019).
[P290] L.-L. Liu, X. Wan, Z. Gao, X. Li, B. Feng, Research on modelling and optimization of hot rolling scheduling, Journal of Ambient
Intelligence and Humanized Computing 10 (2019) 1201–1216 (2019).
[P291] Y. Lu, X. Xu, Resource virtualization: a core technology for developing cyber-physical production systems, journal of Manufacturing
Systems 47 (2018) 128–140 (2018).
[P292] I. Verner, D. Cuperman, A. Fang, M. Reitman, T. Romm, G. Balikin, Robot online learning through digital twin experiments: A weightlift-
ing project, in: Online Engineering & Internet of Things, 2018.
[P293] D. Preuveneers, W. Joosen, E. Ilie-Zudor, Robust Digital Twin Compositions for Industry 4.0 Smart Manufacturing Systems, in: 2018
IEEE 22nd International Enterprise Distributed Object Computing Workshop (EDOCW), 2018.
[P294] R. Kloibhofer, E. Kristen, S. Jakši´
c, Safety and Security in a Smart Production Environment, in: International Conference on Computer
Safety, Reliability, and Security, 2018.
[P295] Y. Cai, B. Starly, P. Cohen, Y.-S. Lee, Sensor Data and Information Fusion to Construct Digital-twins Virtual Machine Tools for Cyber-
physical Manufacturing, Procedia Manufacturing 10 (2017) 1031–1042 (2017).
[P296] M. Schirmann, M. Collette, J. Gose, Ship motion and fatigue damage estimation via a digital twin, in: Life-Cycle Analysis and Assessment
in Civil Engineering: Towards an Integrated Vision, 2018.
[P297] A. Martins, H. Costelha, C. Neves, Shop Floor Virtualization and Industry 4.0, in: 2019 IEEE International Conference on Autonomous
Robot Systems and Competitions (ICARSC), 2019.
[P298] Ø. Kristoffersen, M. Stanko, A. Hoffmann, Short Term Production Optimization Using a Model of the Peregrino Field, Brazil, in: Offshore
Technology Conference Brasil, 2017.
[P299] L. Damiani, M. Demartini, P. Giribone, M. Maggiani, R. Revetria, F. Tonelli, Simulation and Digital Twin Based Design of a Production
Line: A Case Study, in: Proceedings of the International MultiConference of Engineers and Computer Scientists, Vol. 2, 2018.
[P300] S. Haag, C. Simon, FSimulation Of Horizontal And Vertical Integration In Digital Twins., in: ECMS, 2019.
[P301] T. Cichon, J. Roßmann, Simulation-based user interfaces for digital twins: Pre-, in-, or post-operational analysis and exploration of virtual
testbeds, in: 31st Annual European Simulation and Modelling Conference 2017, ESM 2017, 2017.
[P302] U. Dahmen, J. Roßmann, Simulation-based verification with experimentable digital twins in virtual testbeds, in: Tagungsband des 3.
Kongresses Montage Handhabung Industrieroboter, 2018.
[P303] B. Korth, C. Schwede, M. Zajac, Simulation-ready digital twin for realtime management of logistics systems, in: 2018 IEEE International
Conference on Big Data (Big Data), 2018.
[P304] N. Mohammadi, J. E. Taylor, Smart city digital twins, in: 2017 IEEE Symposium Series on Computational Intelligence (SSCI), 2017.
[P305] T. Ruohomäki, E. Airaksinen, P. Huuska, O. Kesäniemi, M. Martikka, J. Suomisto, Smart city platform enabling digital twin, in: 2018
International Conference on Intelligent Systems (IS), 2018.
[P306] H. Lee, T. Kim, Smart factory use case model based on digital twin, ICIC express letters. Part B, Applications: an international journal of
research and surveys 9 (2018) 931–936 (2018).

[P307] A. Hehr, M. Norfolk, J. Sheridan, M. Davis, W. Leser, P. Leser, J. A. Newman, Spatial strain sensing using embedded fiber optics, JOM 71
(2019) 1528–1534 (2019).
[P308] A. K. Sahoo, U. Majumder, M. W. Nielsen, J. H. Garm, Strength of Shear Web With Circular Hole in Wind Turbine Blades and Using
Digital Twining Concept to Reduce Material Testing, in: Gas Turbine India Conference , 2017.
[P309] B. R. Seshadri, T. Krishnamurthy, Structural health management of damaged aircraft structures using digital twin concept, in: 25th
AIAA/AHS Adaptive Structures Conference, 2017.
[P310] D. J. Knezevic, H. Kang, P. Sharma, G. Malinowski, T. T. Nguyen, Structural Integrity Management of Offshore Structures via RB-FEA
and Fast Full Load Mapping Based Digital Twins, in: 28th International Ocean and Polar Engineering Conference, Vol. All Days, 2018.
[P311] A. Ardanza, A. Moreno, Á. Segura, M. de la Cruz, D. Aguinaga, Sustainable and flexible industrial human machine interfaces to support
adaptable applications in the Industry 4.0 paradigm, International Journal of Production Research 57 (2019) 4045–4059 (2019).
[P312] A. Gupta, B. Basu, Sustainable Primary Aluminium Production: Technology Status and Future Opportunities, Transactions of the Indian
Institute of Metals 72 (2019) 2135–2150 (2019).
[P313] R. Beregi, Ádám Szaller, B. Kádár, Synergy of multi-modelling for process control, IFAC-PapersOnLine 51 (2018) 1023–1028 (2018).
[P314] W. T. Gruender, Systems engineering requires digital twins of machine elements, in: CONAT 2016 International Congress of Automotive
and Transport Engineering, 2017.
[P315] E. Tuegel, The airframe digital twin: some challenges to realization, in: 53rd AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dy-
namics and Materials Conference 20th AIAA/ASME/AHS Adaptive Structures Conference 14th AIAA, 2012.
[P316] S. Konstantinov, M. Ahmad, K. Ananthanarayan, R. Harrison, The cyber-physical e-machine manufacturing system: Virtual engineering
for complete lifecycle support, Procedia CIRP 63 (2017) 119–124 (2017).
[P317] P. Sharma, H. Hamedifar, A. Brown, R. Green, et al., The dawn of the new age of the industrial Internet and how it can radically transform
the offshore oil and gas industry, in: Offshore Technology Conference, 2017.
[P318] A. J. Zakrajsek, S. Mall, The Development and use of a Digital Twin Model for Tire Touchdown Health Monitoring, in: 58th Structures,
Structural Dynamics and Materials Conference, 2017.
[P319] N. Nikolakis, K. Alexopoulos, E. Xanthakis, G. Chryssolouris, The digital twin implementation for linking the virtual representation of
human-based production tasks to their physical counterpart in the factory-floor, International Journal of Computer Integrated Manufactur-
ing 32 (2019) 1–12 (2019).
[P320] J. Vachálek, L. Bartalský, O. Rovný, D. Šišmišová, M. Morháˇ
c, M. Lokšík, The digital twin of an industrial production line within the
industry 4.0 concept, in: 2017 21st International Conference on Process Control (PC), 2017.
[P321] E. Glaessgen, D. Stargel, The digital twin paradigm for future NASA and US Air Force vehicles, in: 53rd Structures, Structural Dynamics
and Materials Conference, 2012.
[P322] J. Van Os, The Digital Twin throughout the Lifecycle, in: SNAME Maritime Convention, 2018.
[P323] T. H.-J. Uhlemann, C. Lehmann, R. Steinhilper, The Digital Twin: Realizing the Cyber-Physical Production System for Industry 4.0,
Procedia CIRP 61 (2017) 335–340 (2017).
[P324] J. O. Oyekan, W. Hutabarat, A. Tiwari, R. Grech, M. H. Aung, M. P. Mariani, L. López-Dávalos, T. Ricaud, S. Singh, C. Dupuis, The
effectiveness of virtual environments in developing collaborative strategies between industrial robots and humans, Robotics and Computer-
Integrated Manufacturing 55 (2019) 41–54 (2019).
[P325] J. Bao, D. Guo, J. Li, J. Zhang, The modelling and operations for the digital twin in the context of manufacturing, Enterprise Information
Systems 13 (2019) 534–556 (2019).
[P326] C. Wagner, J. Grothoff, U. Epple, R. Drath, S. Malakuti, S. Grüner, M. Hoffmeister, P. Zimermann, The role of the Industry 4.0 asset
administration shell and the digital twin during the life cycle of a plant, in: 2017 22nd IEEE International Conference on Emerging
Technologies and Factory Automation (ETFA), 2017.
[P327] D. A. Zakoldaev, A. V. Gurjanov, A. V. Shukalov, I. O. Zharinov, The projection of cyber and physical systems digital twins in the stage of
production technological preparation of the Industry 4.0 smart factories, IOP Conference Series: Materials Science and Engineering 537
(2019).
[P328] Z. Liu, N. Meyendorf, N. Mrad, The role of data fusion in predictive maintenance using digital twin, in: AIP Conference Proceedings, Vol.
1949, 2018.
[P329] U. T. Tygesen, M. S. Jepsen, J. Vestermark, N. Dollerup, A. Pedersen, The true digital twin concept for fatigue re-assessment of marine
structures, in: ASME 37th International Conference on Ocean, Offshore and Arctic Engineering, 2018.
[P330] H. Wang, M. Zhou, B. Liu, Tolerance allocation with simulation-based digital twin for CFRP-metal countersunk bolt joint, in: ASME
International Mechanical Engineering Congress and Exposition, 2018.
[P331] E. Essa, M. S. Hossain, A. S. Tolba, H. M. Raafat, S. Elmogy, G. Muahmmad, Toward cognitive support for automated defect detection,
Neural Computing and Applications 32 (2020) 4325–4333 (2020).
[P332] E. G. Kaigom, J. Roßmann, Toward physics-based virtual reality testbeds for intelligent robot manipulators - an eRobotics approach, in:
IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2016.
[P333] D. N. Mavris, M. G. Balchanos, O. J. Pinon-Fischer, W. J. Sung, Towards a digital thread-enabled framework for the analysis and design
of intelligent systems, in: AIAA Information Systems-AIAA Infotech@ Aerospace, 2018.
[P334] D. Ngo, D. A. Guerra-Zubiaga, G. González-Badillo, R. Vatankhah Barenji, Towards a Digital Twin for Cloud Manufacturing: Case Study,
in: ASME International Mechanical Engineering Congress and Exposition, 2018.
[P335] D. Heber, M. Groll, et al., Towards a digital twin: How the blockchain can foster E/E-traceability in consideration of model-based systems
engineering, in: Proceedings of the 21st International Conference on Engineering Design (ICED 17), 2017.
[P336] J. Landahl, M. Panarotto, H. Johannesson, O. Isaksson, J. Lööf, et al., Towards adopting digital twins to support design reuse during
platform concept development, Proceedings of NordDesign (2018).
[P337] A. M. Miller, R. Alvarez, N. Hartman, Towards an extended model-based definition for the digital twin, Computer-Aided Design and
Applications 15 (2018).
[P338] M. Minos-Stensrud, O. H. Haakstad, O. Sakseid, B. Westby, A. Alcocer, Towards Automated 3D reconstruction in SME factories and

Digital Twin Model generation, in: 2018 18th International Conference on Control, Automation and Systems (ICCAS), 2018.
[P339] S. Rabah, A. Assila, E. Khouri, F. Maier, F. Ababsa, V. bourny, P. Maier, F. Mérienne, Towards improving the future of manufacturing
through digital twin and augmented reality technologies, Procedia Manufacturing 17 (2018) 460–467 (2018).
[P340] M. Eckhart, A. Ekelhart, Towards Security-Aware Virtual Environments for Digital Twins, in: Proceedings of the 4th ACM Workshop on
Cyber-Physical System Security, CPSS ’18, 2018.
[P341] V. Stojanovic, M. Trapp, R. Richter, B. Hagedorn, J. Döllner, Towards the generation of digital twins for facility management based on 3D
point clouds, in: Proceeding of the 34th Annual ARCOM Conference, Vol. 2018, 2018.
[P342] M. Borth, E. van Gerwen, Tracking dynamics in concurrent digital twins, in: Complex Systems Design & Management, 2019.
[P343] I. Verner, D. Cuperman, S. Gamer, A. Polishuk, Training Robot Manipulation Skills through Practice with Digital Twin of Baxter, Interna-
tional Journal of Online Biomedical Engineering 15 (2019) 58–70 (2019).
[P344] F. Longo, L. Nicoletti, A. Padovano, Ubiquitous knowledge empowers the Smart Factory: The impacts of a Service-oriented Digital Twin
on enterprises’ performance, Annual Reviews in Control 47 (2019) 221–236 (2019).
[P345] S. Settemsdal, et al., Updated case study: The pursuit of an ultra-low manned platform pays dividends in the north sea, in: Offshore
Technology Conference, 2019.
[P346] K. E. Thoresen, Å. Kyllingstad, S. Hovland, A. Hetland, Using an Advanced Digital Twin to Improve Downhole Pressure Control, in:
SPE/IADC Drilling Conference and Exhibition, 2019.
[P347] A. da Silva Barbosa, F. P. Silva, L. R. dos Santos Crestani, R. B. Otto, Virtual assistant to real time training on industrial environment, in:
Transdisciplinary Engineering Methods for Social Innovation of Industry 4.0: Proceedings of the 25th ISPE Inc. International Conference
on Transdisciplinary Engineering, 2018.
[P348] T. Lechler, E. Fischer, M. Metzner, A. Mayr, J. Franke, Virtual Commissioning–Scientific review and exploratory use cases in advanced
production systems, Procedia CIRP 81 (2019) 1125–1130 (2019).
[P349] J. Xie, X. Wang, Z. Yang, S. Hao, Virtual monitoring method for hydraulic supports based on digital twin theory, Mining Technology 128
(2019) 77–87 (2019).
[P350] S. Burrafato, A. Maliardi, P. Ferrara, T. Grasso, E. De Marchi, R. Campaci, et al., Virtual Reality in D & C: New Approaches Towards
Well Digital Twins, in: Offshore Mediterranean Conference and Exhibition, 2019.
[P351] F. Chinesta, E. Cueto, E. Abisset-Chavanne, J. L. Duval, F. El Khaldi, Virtual, digital and hybrid twins: a new paradigm in data-based
engineering and engineered data, Archives of computational methods in engineering 27 (2020) 105–134 (2020).
[P352] A. Moreno, G. Velez, A. Ardanza, I. Barandiaran, Á. R. de Infante, R. Chopitea, Virtualisation process of a sheet metal punching machine
within the Industry 4.0 vision, International journal on Interactive Design and Manufacturing (IJIDeM) 11 (2017) 365–373 (2017).
[P353] I. Halenar, M. Juhas, B. Juhasova, D. Borkin, Virtualization of Production Using Digital Twin Technology, in: 20th International Carpathian
Control Conference (ICCC), 2019, pp. 1–5.
[P354] Z. Zhu, C. Liu, X. Xu, Visualisation of the Digital Twin data in manufacturing by using Augmented Reality, Procedia CIRP 81 (2019)
898–903 (2019).
[P355] G. Baruffaldi, R. Accorsi, R. Manzini, Warehouse management system customization and information availability in 3pl companies,
Industrial Management & Data Systems 80 (2019).
[P356] C. Ley, S. P. Bordas, What makes Data Science different? A discussion involving Statistics2.0 and Computational Sciences, International
Journal of Data Science and Analytics 6 (2018) 167–175 (2018).
Bibliography
[357] P. Bibow, M. Dalibor, C. Hopmann, B. Mainz, B. Rumpe, D. Schmalzing, M. Schmitz, A. Wortmann, Model-driven development of a digital
twin for injection molding, in: S. Dustdar, E. Yu, C. Salinesi, D. Rieu, V. Pant (Eds.), Advanced Information Systems Engineering, Springer
International Publishing, Cham, 2020, pp. 85–100.
[358] T. Bolender, G. Bürvenich, M. Dalibor, B. Rumpe, A. Wortmann, Self-adaptive manufacturing with digital twins, in: 2021 International
Symposium on Software Engineering for Adaptive and Self-Managing Systems (SEAMS), IEEE, 2021, pp. 156–166.
[359] D. Lehner, J. Pfeiffer, E.-F. Tinsel, M. M. Strljic, S. Sint, M. Vierhauser, A. Wortmann, M. Wimmer, Digital twin platforms: Requirements,
capabilities, and future prospects, IEEE Software (01) (2021) 0–0 (2021).
[360] J. C. Kirchhof, B. Rumpe, D. Schmalzing, A. Wortmann, Structurally Evolving Component-Port-Connector Architectures of Centrally
Controlled Systems, in: M. Cordy, M. Acher, D. Beuche, G. Saake (Eds.), International Working Conference on Variability Modelling of
Software-Intensive Systems, ACM, 2020.
[361] A. Wortmann, O. Barais, B. Combemale, M. Wimmer, Modeling Languages in Industry 4.0: an Extended Systematic Mapping Study,
Software and Systems Modeling 19 (1) (2020) 67–94 (January 2020).
[362] K. Hölldobler, B. Rumpe, A. Wortmann, Software Language Engineering in the Large: Towards Composing and Deriving Languages,
Computer Languages, Systems & Structures 54 (2018) 386–405 (2018).
[363] S. Singh, E. Shehab, N. Higgins, K. Fowler, T. Tomiyama, C. Fowler, Challenges of Digital Twin in High Value Manufacturing, in:
Aerospace Systems and Technology Conference, SAE International, 2018.
[364] D. Olivotti, S. Dreyer, B. Lebek, M. H. Breitner, Creating the Foundation for Digital Twins in the Manufacturing Industry: An Integrated
Installed Base Management System, Information Systems and e-Business Management 17 (2019) 89–116 (2019).
[365] M. Ghobakhloo, The Future of Manufacturing Industry: A Strategic Roadmap Toward Industry 4.0, Journal of Manufacturing Technology
Management 29 (2018) 910–936 (2018).
[366] L. F. C. S. Durão, S. Haag, R. Anderl, K. Schützer, E. Zancul, Digital Twin Requirements in the Context of Industry 4.0, in: P. Chiabert,
A. Bouras, F. Noël, J. Ríos (Eds.), Product Lifecycle Management to Support Industry 4.0, 2018.
[367] K. Y. H. Lim, P. Zheng, C.-H. Chen, A state-of-the-art Survey of Digital Twin: Techniques, Engineering Product Lifecycle Management
and Business Innovation Perspectives, Journal of Intelligent Manufacturing 31 (2019) 1–25 (2019).

[368] B. He, K.-J. Bai, Digital Twin-Based Sustainable Intelligent Manufacturing: A Review, Advances in Manufacturing 9 (2020) 1–21 (2020).
[369] E. Negri, L. Fumagalli, M. Macchi, A Review of the Roles of Digital Twin in CPS-based Production Systems, Procedia Manufacturing 11
(2017) 939 – 948 (2017).
[370] M. P. Papazoglou, A. S. Andreou, Smart Connected Digital Factories: Unleashing the Power of Industry 4.0, in: Cloud Computing and
Servic Science, 2019.
[371] P. Mehta, P. Rao, Z. D. Wu, V. Jovanovi´
c, O. Wodo, M. Kuttolamadom, Smart Manufacturing: State-of-the-Art Review in Context of
Conventional and Modern Manufacturing Process Modeling, Monitoring and Control (2018).
[372] G. E. Modoni, E. G. Caldarola, M. Sacco, W. Terkaj, Synchronizing Physical and Digital Factory: Benefits and Technical Challenges,
Procedia CIRP 79 (2019) 472 – 477 (2019).
[373] F. Tao, H. Zhang, A. Liu, A. Y. C. Nee, Digital Twin in Industry: State-of-the-Art, IEEE Transactions on Industrial Informatics 15 (2019)
2405–2415 (2019).
[374] W. Kritzinger, M. Karner, G. Traar, J. Henjes, W. Sihn, Digital Twin in manufacturing: A categorical literature review and classification,
IFAC 51 (2018) 1016 – 1022 (2018).
[375] C. Kan, C. J. Anumba, Digital Twins as the Next Phase of Cyber-Physical Systems in Construction, in: Computing in Civil Engineering
2019: Data, Sensing, and Analytics, American Society of Civil Engineers, 2019, pp. 256–264.
[376] H. van der Valk, H. Haße, F. Möller, M. Arbter, J.-L. Henning, B. Otto, A Taxonomy of Digital Twins, in: 26th Americas Conference on
Information Systems (AMCIS), 2020, p. 10.
[377] R. Minerva, G. M. Lee, N. Crespi, Digital Twin in the IoT Context: A Survey on Technical Features, Scenarios, and Architectural Models,
Proceedings of the IEEE 108 (2020) 1785–1824 (2020).
[378] J. Autiosalo, J. Vepsäläinen, R. Viitala, K. Tammi, A Feature-Based Framework for Structuring Industrial Digital Twins, IEEE Access 8
(2020) 1193–1208 (2020).
[379] A. Fuller, Z. Fan, C. Day, C. Barlow, Digital Twin: Enabling Technologies, Challenges and Open Research, IEEE Access 8 (2020) 108952–
108971 (2020).
[380] L. Barth, M. Ehrat, R. Fuchs, J. Haarmann, Systematization of Digital Twins: Ontology and Conceptual Framework, in: 3rd International
Conference on Information Science and System, 2020.
[381] M. Héder, From NASA to EU: The Evolution of the TRL scale in Public Sector Innovation, The Innovation Journal 22 (2) (2017) 1–23
(2017).
[382] M. Boeker, W. Vach, E. Motschall, Google Scholar as Replacement for Systematic Literature Searches: Good Relative Recall and Precision
are not enough, BMC Medical Research Methodology 13 (2013) 1–12 (2013).
[383] M. Hankel, The Reference Architectural Model Industrie 4.0 (RAMI 4.0), ZVEI 2 (2015) 4–9 (2015).
[384] A. Kleppe, Software Language Engineering: Creating Domain-Specific Languages using Metamodels, Addison-Wesley, 2008.
[385] B. Kitchenham, O. P. Brereton, D. Budgen, M. Turner, J. Bailey, S. Linkman, Systematic Literature Reviews in Software Engineering – A
Systematic Literature Review, Information and Software Technology 51 (2009) 7–15 (2009).
[386] K. Petersen, R. Feldt, S. Mujtaba, M. Mattsson, Systematic Mapping Studies in Software Engineering, in: 12th International Conference on
Evaluation and Assessment in Software Engineering (EASE) 12, 2008.
[387] K. Petersen, S. Vakkalanka, L. Kuzniarz, Guidelines for Conducting Systematic Mapping Studies in Software Engineering: An update,
Information and Software Technology 64 (2015) 1–18 (2015).
[388] Regulation (EC) No 1893/2006 of the European Parliament and of the Council of 20 December 2006 establishing the statistical classification
of economic activities NACE Revision 2 and amending Council Regulation (EEC) No 3037/90 as well as certain EC Regulations on specific
statistical domains (2006).
[389] International Organization for Standardization, ISO/IEC 15288:2015 systems engineering – system life cycle processes, Tech. rep. (2015).
[390] I. Kurtev, J. Bézivin, M. Aksit, Technological Spaces: An Initial Appraisal, in: 4th International Symposium on Distributed Objects and
Applications, 2002.
[391] E. A. Lee, Disciplined heterogeneous modeling, in: International Conference on Model Driven Engineering Languages and Systems, 2010.
[392] B. A. Kitchenham, S. Charters, Guidelines for Performing Systematic Literature Reviews in Software Engineering, Tech. rep., Keele Uni-
versity and Durham University Joint Report (2007).
[393] D. Budgen, M. Turner, P. Brereton, B. A. Kitchenham, Using Mapping Studies in Software Engineering, in: PPIG 20th Annual Workshop,
Vol. 8, 2008, pp. 195–204.
[394] L. M. do Nascimento, D. L. Viana, P. Neto, D. Martins, V. C. Garcia, S. Meira, A Systematic Mapping Study on Domain-Specific Languages,
in: The Seventh International Conference on Software Engineering Advances (ICSEA 2012), 2012, pp. 179–187.
[395] T. Kosar, S. Bohra, M. Mernik, Domain-Specific Languages: A Systematic Mapping Study, Information and Software Technology 71 (2016)
77–91 (2016).
[396] C. Wohlin, P. Runeson, M. Höst, M. C. Ohlsson, B. Regnell, A. Wesslén, Experimentation in Software Engineering, Springer, 2012.
[397] P. Bibow, M. Dalibor, C. Hopmann, B. Mainz, B. Rumpe, D. Schmalzing, M. Schmitz, A. Wortmann, Model-driven development of a digital
twin for injection molding, in: International Conference on Advanced Information Systems Engineering, 2020, pp. 85–100.
[398] D. Jones, C. Snider, A. Nassehi, J. Yon, B. Hicks, Characterising the Digital Twin: A Systematic Literature Review, CIRP Journal of
Manufacturing Science and Technology 29 (2020) 36 – 52 (2020).
[399] Q. Qi, F. Tao, T. Hu, N. Anwer, A. Liu, Y. Wei, L. Wang, A. Nee, Enabling Technologies and Tools for Digital Twin, Journal of Manufac-
turing Systems 58 (2021) 3–21 (2021).
[400] R. Stark, T. Damerau, Digital Twin, Springer, 2019, pp. 1–8.
[401] T. R. Wanasinghe, L. Wroblewski, B. K. Petersen, R. G. Gosine, L. A. James, O. De Silva, G. K. I. Mann, P. J. Warrian, Digital Twin for the
Oil and Gas Industry: Overview, Research Trends, Opportunities, and Challenges, IEEE Access 8 (2020) 104175–104197 (2020).
[402] J. Ríos, F. Mas, M. Oliva, J. Hernandez-Matias, Framework to support the aircraft digital counterpart concept with an industrial design view,
International Journal of Agile Systems and Management 9 (2016) 212–231 (2016).
[403] F. Tao, J. Cheng, Q. Qi, M. Zhang, H. Zhang, F. Sui, Digital Twin-Driven Product Design, Manufacturing and Service with Big Data, The

International Journal of Advanced Manufacturing Technology 94 (2018) 3563–3576 (2018).
[404] H. Stachowiak, Allgemeine Modelltheorie, Springer, 1973.
[405] Y. Lu, K. C. Morris, S. Frechette, Standards Landscape and Directions for Smart Manufacturing Systems, in: IEEE International Conference
on Automation Science and Engineering (CASE), 2015, pp. 998–1005.
[406] Q. Qi, F. Tao, T. Hu, N. Anwer, A. Liu, Y. Wei, L. Wang, A. Nee, Enabling Technologies and Tools for Digital Twin, Journal of Manufac-
turing Systems 58 (2021) 3–21 (2021).