Cited by 812Digital Transformation in IndustryIoT and Edge/Fog ComputingApplied Advanced Technologies

The DOI matches its Crossref record; 261 of the 269 references it cites were found in a registry or an index; 34 of the 163 sources its claims cite were checked for support.

Digital Twin in the IoT Context: A Survey on Technical Features, Scenarios, and Architectural Models

Open access
Minerva, Roberto, Lee, Gyu Myoung, Crespi, Noel·2020·Proceedings of the IEEE

Abstract

Digital twin (DT) is an emerging concept that is gaining attention in various industries. It refers to the ability to clone a physical object (PO) into a software counterpart. The softwarized object, termed logical object, reflects all the important properties and characteristics of the original object within a specific application context. To fully determine the expected properties of the DT, this article surveys the state-of-the-art starting from the original definition within the manufacturing industry. It takes into account related proposals emerging in other fields, namely augmented and virtual reality (e.g., avatars), multiagent systems, and virtualization. This survey thereby allows for the identification of an extensive set of DT features that point to the “softwarization” of POs. To properly consolidate a shared DT definition, a set of foundational properties is identified and proposed as a common ground outlining the essential characteristics (must-haves) of a DT. Once the DT definition has been consolidated, its technical and business value is discussed in terms of applicability and opportunities. Four application scenarios illustrate how the DT concept can be used and how some industries are applying it. The scenarios also lead to a generic DT architectural model. This analysis is then complemented by the identification of software architecture models and guidelines in order to present a general functional framework for the DT. This article, eventually, analyses a set of possible evolution paths for the DT considering its possible usage as a major enabler for the softwarization process.

Overview

AI snapshot

Written from the paper’s full text.

TL;DR

This survey consolidates the Digital Twin concept for IoT by reviewing manufacturing, augmented/virtual reality, multi-agent systems, and virtualization, then proposing foundational properties and an architectural framework. It illustrates applicability through four scenarios and discusses technical, business, and evolution paths.

Key findings

  • The paper identifies a set of characterizing DT properties, including identity, representativeness/contextualization, reflection, replication, entanglement, persistency, memorization, composability, accountability/manageability, augmentation, ownership, servitization, and predictability. It treats reflection/mirroring, virtualization, entanglement, and representativeness/contextualization as foundational for a basic DT.
  • Entanglement is broken into connectivity, promptness, and association, with strong, simple, and weak forms. Replication can involve multiple logical objects per physical object, including a master replica that synchronizes others.
  • Four application scenarios are used: virtual sensors, digital patient, digital city, and cultural heritage. The paper also outlines a value chain with physical object providers, virtualization platform providers, interface/view providers, and service/application providers.
  • A layered DT architecture is proposed, spanning device/edge/cloud infrastructure, object lifecycle management, data collection and analysis, simulation/visualization, and open APIs. Industrial and IoT platform examples include Predix, Microsoft, IBM, SAP, FIWARE, and Eclipse Ditto.

Methods

  • Design: a survey and conceptual consolidation of Digital Twin definitions, technologies, and architectural models, not an empirical study.
  • Data/sample: literature and technical contributions from manufacturing, augmented/virtual reality, multi-agent systems, and virtualization, plus selected academic and industrial platform examples.
  • Analysis: derives a common property set, proposes a consolidated DT definition, and uses four scenarios and a layered reference framework to illustrate implementation requirements and value.

Limitations

  • The paper states that a fully fledged Digital Twin system does not yet exist, that current applications are dominated by prototypes or proprietary solutions, and that many are domain-specific and lack generality or openness.
  • It lists unresolved technical and societal issues, including entanglement, scaling to millions of objects, aggregation, self-management/zero-touch operation, data contextualization, crawling and enrichment, simulation/prediction of large systems, security, privacy, and ethical concerns.
  • Evident from the text: the proposed property set and architecture are not empirically validated or benchmarked, and the four scenarios are illustrative rather than tested deployments.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1109/jproc.2020.2998530

Cites

261 of 261 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

34 of 163 cited sources checked · 3 supported · 0 flagged · accuracy not yet measured

Show all 49 claims

  • Probably not a substantial claim. 297
  • It may state two separate claims as one. 82
311 of 311
  • The paper relies on this without support.

    The Digital Twin concept has been attracting increasing attention for its ability to create a software counterpart of a physical object.

    “The Digital Twin (DT) concept has been attracting increasing attention for its the ability to create a software counterpart of a physical object.”
  • The paper relies on this without support.

    The Digital Twin model has attracted significant interests, both in academia and industry.

    “Since then, the Digital Twin model has attracted significant interests, both in academia and industry.”
  • The paper relies on this without support.

    Usage of DT first grew in the manufacturing environment and later in the community of the Internet of Things (IoT) and Cyber Physical Systems (CPS).

    “Usage of DT first grew in the manufacturing environment and later in the community of the Internet of Things (IoT) and Cyber Physical Systems (CPS).”
  • The paper relies on this without support.

    The definition has moved from an industrial artifact or product to a more generic notion applicable to almost any physical object and, in principle, to intangible objects.

    “One important point is that the definition has moved from an industrial artifact or product to a more generic notion applicable to almost any physical object and, in principle, to intangible objects.”
  • The paper relies on this without support.

    An extensive literature has been produced about DT, but sometimes different words or different facets of the DT have been considered, defined, and stressed.

    “An extensive literature has been produced about DT, but, sometimes, different words or different facets of the DT have been considered, defined, and stressed.”
  • The paper relies on this without support.

    There is an increasing need to develop products or to control processes in an adaptive manner.

    “There is an increasing need to develop products or to control processes in an adaptive manner.”
  • The authors state this as an assumption.

    The possibility that the physical and the logical objects are connected by means of the internet or other specific networks can often be assumed.

    “Thanks to the increasing capabilities of communications, the possibility that the physical and the logical objects are connected by means of the internet or other specific networks can often be assumed.”
  • The paper relies on this without support.

    The different paths have different views and definitions of the Digital Twin, but they are conceptually aligned at the level of replicating physical objects by software.

    “The different paths have different views and definitions of the Digital Twin.They refer to the general idea of replicating, by software, physical objects or products.This is the level at which the DT definitions are 'conceptually' aligned, but they differ in terms of properties and emphasis on specific characteristics.”
  • The authors state this as an assumption.

    The originating object of a Digital Twin is material in nature.

    “For the sake of simplicity, in this paper the originating object is material in nature.”
  • The authors state this as an assumption.

    A physical object can be fully described and characterized by a set of variables and their values in a specific multi-dimensional space.

    “Let us assume that a physical object is fully described and characterized by a set of variables and their values 𝑋 = {𝑥 1 , 𝑥 2 , … , 𝑥 𝑛 } in a specific multi-dimensional space 𝑆 𝑤ℎ𝑒𝑟𝑒 ∀ 𝑥 𝑖 ∈ 𝑆.”
  • The paper relies on this without support.

    Some physical objects may lack the ability to process, store, or communicate data.

    “Depending on the type of (physical) objects, they may or may not be able to process, store and communicate data.”
    “Not all of the physical objects are capable of communication or processing.”
  • The paper relies on this without support.

    A fully-fledged Digital Twin system does not exist yet.

    “Actually, a fullyfledged Digital Twin system does not exist yet, and it is not sure that one will be implemented soon.”
  • The paper relies on this without support.

    The DT approach is sometimes based on a 'closed-loop' approach.

    “The DT approach is sometimes based on a 'closed-loop' approach.”
  • The paper relies on this without support.

    The physical object will often have scarce processing, storage and communication capabilities.

    “The most common one is that the physical object will often have scarce processing, storage and communication capabilities.”
  • The paper relies on this without support.

    One of the major issues, with relevant costs, associated with health care is the monitoring and observation of patients.

    “One of the major issues, with relevant costs, associated with health care is the monitoring and observation of patients.”
  • The authors state this as an assumption.

    The virtualization space, and the supporting platform, is capable of managing the functioning of the logical objects.

    “It is assumed that the virtualization space, and the supporting platform, is capable of managing the functioning of the logical objects.”
  • The paper relies on this without support.

    The relationship between the physical object and the customer and the producer is very important for enabling new business models, and for Servitization in particular.

    “The relationship between the physical object and the customer and the producer is very important for enabling new business models, and for Servitization in particular.”
  • The paper relies on this without support.

    IoT and Smart City applications suffer from application silos, i.e., strong separation of architectures, tools, interfaces, and data between different application domains.

    “Typically, the IoT and the Smart Cities applications suffer from the issue of applications silos, i.e., the strong separation of architectures, tools, interfaces and data between different applications domains, transportation, traffic, cultural heritage, and more.”
  • The paper relies on this without support.

    Smart City platforms and applications require the ability to deal with a multitude of different objects and their instantiations and replications.

    “One strong requirement of a Smart City platform and its applications is the ability to deal with a multitude of different objects and even more with plenty of their instantiations and replications.”
  • The paper relies on this without support.

    Cultural artifacts have physical characteristics and features but also a relevant content or idea associated with them.

    “These kinds of physical objects are not simply objects, they have physical characteristics and features, but they also have a relevant content, an idea, associated to them.”
  • The paper relies on this without support.

    Ownership is important to determine the historical relevance of a cultural artifact.

    “In addition, the ownership is important to determine the historical relevance of the artifact.”
  • The paper relies on this without support.

    Cultural objects are mainly determined by context and perspective of people over time, not only physicality.

    “Cultural Objects are not only determined by their 'physicality' but mainly by the context and the perspective of people over time.”
  • The paper relies on this without support.

    The current offering of IoT applications may cover a large part of the needs and requirements of use cases.

    “The current offering of IoT applications may cover a large part of the needs and requirement of use cases.”
  • The paper relies on this without support.

    AI applications, if data sets are available, could be competitors of Digital Twin implementations.

    “Also AI applications, if data sets are available, could be competitors of DT implementations.”
  • The paper relies on this without support.

    To date, the Digital Twin application domain is dominated by prototypes and/or proprietary solutions, sometimes not fully implemented.

    “It is important to understand that, to date, the application domain is dominated by prototypes and/or by proprietary solutions, sometimes not fully implemented.”
  • The paper relies on this without support.

    Some existing Digital Twin solutions only address specific application domains and lack the needed generality and openness to be widely used.

    “In addition, some of them only address specific application domains and lack the needed generality and openness to be widely used.”
  • The paper relies on this without support.

    Physical objects in IoT often have scarce processing, storage, and communication capabilities.

    “The most common one is that the physical object will often have scarce processing, storage and communication capabilities.”
  • The paper relies on this without support.

    Different communication links can be difficult and costly to maintain.

    “In addition, different communication links can be difficult and costly to maintain.”
28 of 28

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Reflection equation maps every value to the first mirrored value

    Minor

    The first expression equates every xi to x1′, whereas the subsequent expression maps xi to its corresponding xi′. For a set containing distinct values, these cannot both describe the stated reflection mapping.

  • The 1:N relationship is defined in opposite directions

    Minor

    The manufacturing section uses 1:N for multiple physical products sharing one logical object. The identity section uses 1-to-N for multiple logical replicas of one physical asset, reversing what the ratio counts.

2 of 2

Citations

Cites

Flagged first, then by citations

Show all 269 references

8 of the 269 references couldn’t be matched to a known paper.

Research works printed but not linked: 37

  1. 0, DITTO API Definition and Documentation
  2. 2018, Meet Preston and Explore Your Digital Twin in Virtual Reality
  3. edwards, 2017, IBM digital twin: Designing for a connected, software driven world, Digital Twin at the Genius of Things Summit in Boston
  4. van hoof, 2018, Announcing Azure Digital Twins Create Digital Replicas of Spaces and Infrastructure Using Cloud AI and IoT
  5. rødseth, 2018, From digital twin to maritime data space: Transparent ownership and use of ship information, Proc 13th Int Symp Integr Ship’s Inf Syst Mar Traffic Eng Conf (ISIS-MTE), 1
  6. hippold, 2019, How Digital Twins Simplify the IoT
  7. 2017, PREDIX the industrial IoT application platform
  8. 2017, The Promise of a Digital Twin Strategy Best Practices for Designers and Manufacturers of Products and Industrial Equipment
  9. asenjo, 2018, Digital twin—Modeling interrelated devices, Innovations Journal
  10. lammers, 2018, Digital Twin Offers Huge Opportunities for real Estate Life Cycle
  11. erikstad, 2017, Merging physics, big data analytics and simulation for the next-generation digital twins, Proc 10th Symp High-Perform Mar Vehicles (HIPER), 140
  12. wasik, 2013, Programmable World All Our Objects Will Act As One
  13. bottani, 2017, From the cyber-physical system to the digital twin: The process development for behaviour modelling of cyber guided vehicle in M2M logic, Proc 12th Summer School Francesco Turco Ind Syst Eng, 1
  14. pettey, 2017, Prepare for the Impact of Digital Twins
  15. van zütphen, 2018, Digital twins when decency meets temptation
  16. biru, 2015, Towards a definition of the Internet of Things (IoT), Proc IEEE Internet Things Initiative, 1
  17. litvinova, 2018, Collaborating in Virtual Reality by using Digital Twins, Proc ISPIM Innov Symp, 1
  18. castelluccio, 2018, Digital twins invade industry, Strategic Finance, 63
  19. tao, 2017, Digital twin workshop: A new paradigm for future workshop, Comput Integr Manuf Syst, 23, 1
  20. liu, 2018, The role of data fusion in predictive maintenance using Digital Twin, AIP Conf Proc, 1949
  21. hamilton, 2018, Detecting opportunities and challenges for australian rural industries Final report
  22. constantinescu, 2018, Methodology of realizing the digital twin of exoskeleton-centred workspaces, Acta Technica Napocensis Series Applied Mathematics and Mechanics, 29, 3
  23. belkin, 2010, Object-oriented world modelling for autonomous systems, Joint Workshop of Fraunhofer IOSB and Institute for Anthropomatics, 231
  24. 10.1109/SAS.2019.8706111 · DOI
  25. 2018, High Level Architecture Release 4 0
  26. 2016, SmartM2M IoT Standards Landscape
  27. 2009, Overview of Ubiquitous Networking and of Its Support in NGN[S]
  28. kenney, 2016, The rise of the platform economy, Issues Sci Technol, 32, 61
  29. holenstein, 2007, Breaking the Availability Barrier II Achieving Century Uptimes With Active/Active Systems
  30. 2013, Network Function Virtualization Architectural Framework
  31. 2016, GE digital twin—Analytic engine for the digital power plant
  32. rotem-gal-oz, 2006, Fallacies of Distributed Computing Explained
  33. escorsa, 2018, Digital Twin A Glimpse at the Main Patented Developments
  34. bolton, 2018, Gemini principles
  35. pontes, 2018, Visit [Industrial Internet of Things—Digital Twins] Fraunhofer Institute of Optronics
  36. 2017, PREDIX the industrial IoT application platform
  37. luyckx, 2018, Impact of the Digital Twin on the Enterprise Architecture

Cited by

Sorted by citations

Items 401–500 of 804

Metadata

Authors
Minerva, R., Lee, G. M., & Crespi, N.
Affiliations
  • Centre National de la Recherche Scientifique (FR)
  • Liverpool John Moores University (GB)
Venue
Proceedings of the IEEE
Published
2020-10
Volume
108
Issue
10
Pages
1785-1824
ISSN
  • 0018-9219 (ISSN-L)
  • 1558-2256
OpenAlex
W3036600965
Cited by
812
References
269
Citations by year
2026: 119 · 2025: 185 · 2024: 179 · 2023: 150 · 2022: 108 · 2021: 58 · 2020: 13
FWCI
52.45
Type
journal article
Language
en
Locations
  • Proceedings of the IEEE (publishedVersion)
  • Liverpool John Moores University (submittedVersion)
  • HAL (submittedVersion)
Topics
  • Physical Sciences › Engineering › Industrial and Manufacturing Engineering › Digital Transformation in Industry
  • Physical Sciences › Computer Science › Computer Networks and Communications › IoT and Edge/Fog Computing
  • Physical Sciences › Engineering › Industrial and Manufacturing Engineering › Applied Advanced Technologies
Keywords
digital twin, Internet of Things, network softwarization, digital twin definition, digital twin architecture, software architecture models, virtual reality, multi-agent systems, virtualization, manufacturing industry

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-10-11, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-11 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review