digital twin 5g case study

• In-Building Tech
• Telco Cloud
• 5G NR Release 17
• Private Networks
• Telco AI Deep Dive
• Cloud RAN Deep Dive
• Sponsored Channels:
• Keysight 5G Solutions
• Qualcomm 5G Insights
• 6G Technology & Testing

• Advertise/Sponsor
• Editorial Calendar
• 5G Talent Talk
• Wireless Connectivity to Enable Industry 4.0 for the Middleprise
• Well Technically…
• Will 5G Change the World
• Accelerating Industry 4.0 Digitalization
• White Papers
• Analyst Angle
• Network Infrastructure
• Understanding the 5G Advanced and 6G future
• Rohde & Schwarz 6G Technology & Testing
• Telco Cloud & Edge Forum

Digital twins use cases across different industries and a case study

Digital-twin technology is increasingly being adopted in multiple verticals, with the industry close to a maturity point, Dan Isaacs, CTO of the Digital Twin Consortium, said during a presentation at the 5G Manufacturing Forum— available on demand here .

“We’ve seen the evolution of digital twins going across from having its origination in the aerospace and manufacturing sectors, but it’s really accelerating across every major industry,” Isaacs said.

The executive noted that there is an increasing adoption of digital twins in industries such as aerospace, automotive, construction, transportation, fintech, energy, manufacturing, healthcare, pharmaceutical and security.

The executive noted that the Digital Twin Consortium was formed two and a half years ago with Microsoft, Dell, Ansys and LendLease, among the founding members. “And we’ve continued to build out our steering committee across these different areas. So you can see Johnson Controls coming in, you can see also GE Digital and Northrop Grumman. We continue to build out, and we’re actually going to be announcing a major oil and gas company coming in the next in the next coming months. We continue to see the value that digital twins are providing across virtually every industry.”

By the end of this year, the consortium aims to have a regional branch organizer in every major continent of the world, Isaacs said. “We now have associations, consortiums and other groups where we can now build out and extend further the ecosystem of digital twin and enabling technologies (…) This consortium really has ties around the world so we can extend and really drive not just the awareness, but the adoption of digital twins.”

Commenting on the different use cases enabled by digital twin technology in different industries, Isaacs said that some of the key use cases include long-haul Covid-19 management, bio-mimicry in life science, buildings as batteries, health assurance buildings, emergency services, water management, carbon sequestration, wind farm operation, industrial automation, AI real-time quality control, supply chain composability, intelligent transport, smart corridors, fleet charging stations and financial transactions, among others.

Isaacs also highlighted a digital twin use case enabled by Vicomtech, a technological center specializing in artificial intelligence, visual computing and interaction based in Spain.

The executive highlighted that the solution includes an ERP system which sends the order to a Manufacturing Execution System (MES). The MES launches the manufacturing process, including a conveyer that moves parts and a remote-controlled robot. A remote operator decides if the part is valid, viewing a live camera image over 5G, and the robot finishes the manufacturing process. Also, a Data Acquisition Module communicates with the elements of the manufacturing line to store relevant traceability and process data at a High-Performance Computer (HPC). This data feeds an Assets Administration Shell (AAS) based digital representation of the line elements that are the input for the twin. The twin, including the dashboard, is deployed in virtual environment, viewed by AR googles, the executive said.

Issacs explained that the digital twin provides a digital replica of the manufacturing cell, enabling manipulation and quality control inspection of manufacturing cells using virual reality. The digital twin also tracks energy efficiency, performance metrics and maintainability of the manufacturing line.

Other use cases in which digital twins are being implemented are predictive quality and precision control, he added.

“The value that digital twin is providing is looking and understanding and being able to have set points or thresholds where you can have advanced alerts and know that you’re reaching to some point where either the anomalous behavior may occur or there could be some type of failure resulting in unplanned downtime, which cascades and ripples through the entire supply chain,” Isaac said.

Meanwhile, Simon Frumkin, CEO of U.K. firm Freshwave, provided insights on the 5G Integrated Railway Augmented Reality Digital Twin (ARDT) project at St. Pancras International train station in London, in which a consortium of business and organizations created a platform through which rail assets, rolling stock and station control systems are connected via a mobile private network (MPN) to improve efficiency, training and passenger experience.

Funded by Innovate UK and led by Pauley, the project involved close collaboration between HS1, Network Rail High Speed, the University of Sheffield Advanced Manufacturing Research Center, Athonet and Freshwave.

“We delivered to our client, Network Rail, a number of different use cases including fault diagnosis, helping them to understand data, but also to understand passenger flows,” Frumkin said.

Freshwave provided the radio coverage plan, network build services, cloud core hardware and operations support in close partnership with Athonet who provided the network design, 5G core software and network equipment.

“I don’t think there’s any way that a digital twin of this complexity could be delivered on any legacy technology. 5G private network is almost certainly the only type of wireless network it could be,” he added.

• Digital Twin Consortium
• digital twins
• manufacturing

ABOUT AUTHOR

RELATED POSTS

Softbank launches private 5g service for enterprises, good 5g is bad 5g with poor network design – especially in enterprise, simpler systems, faster returns, happier teams – five key trends in private 5g.

Since 1982, RCR Wireless News has been providing wireless and mobile industry news, insights, and analysis to mobile and wireless industry professionals, decision makers, policy makers, analyst and investors.

© 2021-2022 RCR Wireless News

• About RCR Wireless News
• Wireless News Archive

The Digital Twin pp 433–446 Cite as

Digital Twin for 5G Networks

• Marius Corici 4 &
• Thomas Magedanz 4  
• First Online: 03 June 2023

2700 Accesses

1 Citations

The current 5th Generation Mobile Networks (5G) standardization is aiming to significantly raise the applicability of communication networks for a wide variety of use cases spanning from industrial networks, automotive, content acquisition, multimedia broadcasters and eHealth (NGMN Alliance. 5G white paper. Next generation mobile networks, white paper 1, 2015). At the same time, this presumes that a smaller size, dedicated 5G network must be integrated into an existing complex communication infrastructure, specific to the use case. This becomes particularly challenging with a 5G network as it is a highly complex systems by itself with highly complex network management requirements in terms of fault, performance, and security. To address this issue, existing work suggests that the use of Digital Twins (DT) or Asset Administration Shells (AAS) within the industrial domain, to model information about the 5G network and to use this data to plan, evaluate and make decisions on how to optimize the behavior of the system. However, the DT based modelling of 5G systems remains a relative new topic. Within this chapter, we provide a comprehensive overview of how the exiting 5G network management uses a sort of Digital Twin (DT) approach and how a full DT paradigm would optimize the 5G networks. First, the 5G network as a complex system will be described with the specific automation and optimization capabilities as well as underlining its limitations. The additional opportunities for a more flexible DT of the 5G network, due to its softwarization would be further analyzed especially concentrating on the extension of the DT model towards an even more complexity as well as towards the new opportunities of dynamic resource scheduling as representative elements for the 5G network management functionality. A short analysis on the impact of the network between the DT and the 5G system will be provided to understand the impact of the network characteristics such as delay, capacity, and packet loss on the functioning of the system. To conclude, the presented considerations can act as robust enablers for future 6G networks including multiple self-reconfiguration mechanisms. A short set of considerations are made on the governance of the multiple decision points and potential ways to implement such multi-decision models.

• Software networks
• Network management
• System modelling

This is a preview of subscription content, log in via an institution .

Buying options

• Available as PDF
• Read on any device
• Instant download
• Own it forever
• Available as EPUB and PDF
• Durable hardcover edition
• Dispatched in 3 to 5 business days
• Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Boschert, S., & Rosen, R. (2016). Digital twin—The simulation aspect. In Mechatronic futures (pp. 59–74). Springer.

Google Scholar  

Leinwand, A., & Fang, K. (1993). Network management: A practical perspective . Addison-Wesley Longman Publishing.

MATH   Google Scholar  

Ziegler, V., Wild, T., Uusitalo, M., Flinck, H., Räisänen, V., & Hätönen, K. (2019). Stratification of 5G evolution and beyond 5G. In 2019 IEEE 2nd 5G World Forum (5GWF) (pp. 329–334). IEEE.

O-RAN Architecture Description 4.0 – March 2021 (O-RAN.WG1.O-RAN-Architecture-Description-v04.00). https://www.o-ran.org/specifications

3GPP TS 23.501. System architecture for the 5G System (5GS) . v17.0.0, 2021-03-30, www.3gpp.org

3GPP TS 23.502. Procedures for the 5G System (5GS) . v17.0.0, v17.0.0, 2021-03-31, www.3gpp.org

Oxford Dictionary. https://www.lexico.com/definition/Complexity . Last visited on 2021-09-18.

Casti, J. https://www.britannica.com/science/complexity-scientific-theory/Surprise-generating-mechanisms . Last visited in 2021-09-18.

Casti, J. https://core.ac.uk/download/pdf/52944519.pdf . Last visited on 2021-09-18.

Ciavaglia, L., Ghamri-Doudane, S., Smirnov, M., Demestichas, P., Stavroulaki, V.-A., Bantouna, A., & Sayrac, B. (2012). Unifying management of future networks with trust. Bell Labs Technical Journal, 17 (3), 193–212.

Article   Google Scholar  

Wilson, G. (2021). http://blog.fogcreek.com/eight-fallacies-of-distributed-computing-tech-talk/based on Peter Deutsch, http://nighthacks.com/jag/res/Fallacies.html last visited on 2021-09-18 .

Corici, M., & Magedanz, T. (2021, November). “One layer to rule them all”: Data Layer-oriented 6G networks. In Shaping future 6G networks: Needs, impacts and technologies . (pp. 221–233). Wiley – IEEE.

Macal, C. M. (2016). Everything you need to know about agent-based modelling and simulation. Journal of Simulation, 10 (2), 144–156.

Hammersley, J. (2013). Monte Carlo methods . Springer.

Download references

Author information

Authors and affiliations.

Software-Based Networks Division of Fraunhofer FOKUS Institute/Technische Universität Berlin, Berlin, Germany

Marius Corici & Thomas Magedanz

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Marius Corici .

Editor information

Editors and affiliations.

Telecom SudParis, Institut Polytechnique de Paris, Palaiseau, France

Noel Crespi

OpenTechWorks Inc., Wayne, PA, USA

Adam T. Drobot

Roberto Minerva

Rights and permissions

Reprints and permissions

Copyright information

© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG

About this chapter

Cite this chapter.

Corici, M., Magedanz, T. (2023). Digital Twin for 5G Networks. In: Crespi, N., Drobot, A.T., Minerva, R. (eds) The Digital Twin. Springer, Cham. https://doi.org/10.1007/978-3-031-21343-4_16

Download citation

DOI : https://doi.org/10.1007/978-3-031-21343-4_16

Published : 03 June 2023

Publisher Name : Springer, Cham

Print ISBN : 978-3-031-21342-7

Online ISBN : 978-3-031-21343-4

eBook Packages : Computer Science Computer Science (R0)

Share this chapter

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Publish with us

Policies and ethics

• Find a journal
• Track your research

• Search for: Toggle Search

Ericsson Builds Digital Twins for 5G Networks in NVIDIA Omniverse

If a virtual tree falls and hits a 5G antenna in Stockholm, does it make a sound in Chicago?

Thanks to Ericsson, the answer is yes. Everything from the locations of trees to the height and composition of buildings is crucial because they impact 5G wireless signals in networks serving smartphones, tablets and millions of other internet-connected devices.

The Stockholm-based maker of telecommunications equipment is combining decades of radio network simulation expertise with NVIDIA Omniverse Enterprise , a real-time virtual world simulation and collaboration platform for 3D workflows. In NVIDIA Omniverse, Ericsson is building city-scale digital twins to help accurately simulate the interplay between 5G cells and the environment for maximum performance and coverage.

The rollout of 5G networks is creating new challenges, and with over 15 million microcells and towers to be deployed globally by network operators over the next five years, it’s no small undertaking.

The 5G Challenge 

5G enables a multitude of new use cases ranging from IoT and manufacturing to self-driving cars and telehealth. Networks serving these use cases operate in vastly different environments. New types of devices will enter the networks, and the number of devices will grow by orders of magnitude in the next few years. These factors make the design and development of 5G products and networks very complex.

Without a digital-twin approach, the interaction between radio transmitters, the environment, and humans and devices that are on the move had to be understood with less detail. And many features had to be field tested only after the networks were already built.

“Before Omniverse, coverage and capacity of networks was analyzed by simplifying many aspects of the complex interactions, such as the physical phenomena and mobility aspects,” said Germán Ceballos, a researcher at Ericsson. “Now we’ll be able to simulate network deployments and features in a highly detailed scale using Omniverse.”

Creating an end-to-end city-scale digital twin results in faster development cycles, better network optimization and ultimately better, swifter networks because it delivers fast insights into what products to install where, Ceballos says.

GPU-Powered Insights

Ericsson and NVIDIA entered a strategic partnership in 2019 with the goal of cross-pollinating technologies and challenges between their respective domains.

The 3D community has historically been fragmented along several proprietary and competing formats and tool chains. This creates unacceptable degrees of lock-in and limited abilities to extend to new simulation use cases.

The NVIDIA Omniverse platform provides key technologies that allow Ericsson to accurately model network performance across dynamic environmental elements. NVIDIA RTX-accelerated real-time ray tracing allows researchers to see precise representations of signal quality at every point in the city, in real time, which wasn’t possible before. This means Ericsson can experiment with its telecom products such as beam-forming and explore their impact interactively and instantaneously.

Plus, with emerging Omniverse platform capabilities such as Omniverse VR, network engineers could soon put on a virtual-reality headset and explore any part of any model, at 1:1 scale, tuning the parameters, antenna and literally “seeing” the effects – things that aren’t visible in real life.

For Ericsson, the Omniverse digital twin offers universal telecoms insights, faster development cycles and the ability to achieve a cutting-edge network at lower costs.

NVIDIA GTC is taking place online through Nov. 11. Watch NVIDIA CEO Jensen Huang’s GTC keynote address below.

NVIDIA websites use cookies to deliver and improve the website experience. See our cookie policy for further details on how we use cookies and how to change your cookie settings.

Share on Mastodon

• Back to stores
• Find a Verizon Store
• Request a Store Appointment
• Australia (eng)
• Deutschland (de)
• France (fr)
• Great Britain (eng)
• Netherlands (eng)
• Singapore (eng)
• Shop all mobile
• Shop all deals
• Red Hot Deals
• Internet and mobile bundle
• Veteran and military offers
• Smartphones
• Check device compatibility
• Rugged phones
• Certified pre-owned devices
• Basic phones
• Desk and conference phones
• Trade in your device
• Shop all devices
• Shop all plans
• Business Unlimited plans
• Bring your own device
• Device protection plans
• International services
• Mobile hotspots
• Smartwatches
• Public Sector devices
• eSIM devices
• Shop all Business Internet
• Fios Business Internet deals
• Check internet availability
• Fios Business Internet
• 5G Business Internet
• LTE Business Internet
• Internet Dedicated
• Business Broadband
• Jetpacks and hotspots
• Routers & extenders
• Internet speed test
• TechTeam support
• Internet usage calculator
• All solutions
• All products
• Mobile plans
• Phones and devices
• Connected laptops
• Business Internet
• Tech support
• Internet security solution
• Managed network services
• Virtual network services
• Connectivity
• Edge solutions
• Network consulting
• CPE and Subscription Services
• Network and cloud security
• Threat detection and response
• Web security
• Advanced security operations center (SOC) services
• Cyber risk management
• Mobile threat defense
• Mobile device management
• Rapid Response Retainer
• Webex Cloud Communications
• Webex Suite
• Verizon Calling Solutions for Microsoft Teams
• VoIP phone services
• Push to Talk Plus
• Field Force Manager
• Business & Enterprise Messaging
• Cloud contact center
• Voice security
• Customer engagement
• Customer experience consulting
• Voice services
• Voice API solutions
• IoT devices
• IoT networks
• IoT platform and services
• Iot applications
• Connected assets
• Smart cities and communities
• Connected commerce
• Application and device security
• ThingSpace platform
• GPS Fleet Tracking software
• GPS Equipment tracking
• AI dashcam solutions
• Accelerated access
• Cashierless checkout
• Crowd analytics
• Quality assurance for production
• Quality assurance for warehouse
• Future-ready worksites
• Efficient operations
• Connected workforces
• Data security
• Digital intelligence
• Enhanced patient experiences
• Data security and privacy
• Remote care
• Connected vehicles
• Unified technology
• Immersive vehicle experiences
• Connected factories
• Stronger supply chains
• Network automation
• Enterprise connectivity
• Enhanced customer experiences
• Data protection
• Better infrastructure
• Protect critical data
• Personalized customer experiences
• Real-time innovation
• Intelligent venues
• Connected athletes
• Enhance experiences
• Data and IP protection
• Connected smart grids
• Federal Government
• Public Safety
• State & Local Government
• All Industries
• Business internet
• Verizon Mobile for Microsoft Teams
• Private 5G Network
• 5G Edge computing
• 5G Ultra Wideband
• 2023 DBIR for Public Sector
• Improving business intelligence
• Evolving customer experiences
• Enhancing workforce productivity
• Enabling agility via automation
• 5G use cases
• 5G for Public Sector
• Private MEC
• Cloud service providers
• 5G Edge solutions
• Immersive experiences
• Building customer trust
• Contact center technology
• Upgrading IT infrastructure
• Digital transformation services
• Business continuity
• Growing your small business
• Network as a Service
• Social engineering defense
• Understanding security risk
• Data Breach Investigations Report
• Defending from core to edge
• Protecting your small business
• Managing cyber risk
• Reimagining your workplace
• Working remotely
• Securing remote work
• Cloud contact center solutions
• 2023 Payment Security Report
• Why choose us
• Executive Briefing Program
• Looking Glass
• Awards and recognition
• Small Business
• Public Sector
• All events and webinars
• Distance learning
• Edge & why it matters
• Improving CX
• Network-as-a-Service
• New workplace
• Public safety
• Security strategies
• Cybersecurity news and events
• Mobile Security Index report
• Payment Security Report
• Gartner recognition
• Verizon Partner Network
• Fleet management solutions
• Secure fleet vehicle management
• HOPE Agency turns to Verizon
• ABP builds the port of the future
• Cintas pulls off an IT upgrade
• Fujifilm steps up its cybersecurity
• All stories
• All insights

Categories and use cases

• Connected retail experiences
• AR-enabled retail displays
• Enhancing the fan experience with immersive realities
• How 5G Edge can improve Cashierless Checkout technology
• Accelerated Access for Stadiums and Venues with 5G Edge
• AR-enabled collaboration
• Advanced visualization in patient care
• Immersive virtual reality education
• How 5G can Improve Patient Data Analytics in Healthcare
• Autonomous machinery and operations
• 5G-powered drone monitoring
• Semiautonomous transportation
• 5G Edge and Automated Guided Vehicles (AGV)
• 5G-powered connected worksite
• Near real-time supply chain
• How 5G impacts Inventory Management
• How 5G Edge improves Warehouse Quality Assurance

5G-powered digital twin

• Massive IoT
• Near real-time analytics
• How 5G Edge can foster safer crowd movement and management
• 5G-enhanced situational awareness
• Autonomous drone surveillance
• 5G immersive tactical training for first responders

The challenge

Creating accurate and responsive virtual models of physical objects and devices.

Determining what modifications will optimize physical products and systems can be time consuming and costly. Current virtual modeling processes may lack near real-time data inputs needed for generating accurate, ongoing and responsive projections.

Efforts to create more sophisticated models, like digital twins, of real products and systems may need near real-time data and could require costly infrastructure, networks and IoT sensors. 

Forty-six percent of organizations worldwide considered predictive maintenance either critical or very important.

Seventy-six percent of manufacturing respondents said that digital twins will be important or extremely important to their firm’s priorities in the next 12 months..

  https://www.forrester.com/report/Forrester+Infographic+Whats+All+The+Fuss+About+Digital+Twin/-/E-RES160736

Forty-six percent of financial and insurance industry respondents said they were either implementing digital twins or planning to implement in the next 12 months.

Transform predictive maintenance..

Digital twins could transform how businesses optimize and perform predictive maintenance of products, machinery and systems.

5G opportunities

Gain powerful insights from a 5g-powered digital twin..

5G has the latency and throughput to capture near real-time IoT data and support digital twins.

5G could allow you to experiment, test and optimize the digital twin before implementing on the physical operation.

A 5G-enabled digital twin could handle the massive volume of data needed for an accurate analysis in manufacturing operations.

5G could also create increased flexibility with IoT deployment and configuration for a digital twin refinement.

How it works

What a digital twin could do for an aircraft engine design

An engineer sits at a computer monitoring and testing an aircraft engine’s digital twin.

The near real-time updates inform the engineer of the engine’s performance, issues and anomalies.

• The physical engine is located in a manufacturing facility.

The engine has sensors in key locations that could send data in near real time utilizing multi-access edge computing (MEC) via a 5G network, enabling updates to the digital twin.

• The engineer could simulate the engine’s life cycle and material stresses in various real-world environments and conditions virtually, with rapid testing and optimization.

5G and MEC can deliver low latency and high bandwidth to capture data in near real time, helping with performance modeling and optimizations prior to physical implementation.

5G built right

Our 5G Ultra Wideband network is built right to power transformative possibilities for business. Creating a responsive, near real-time digital twin requires low latency and large data volumes that 5G Ultra Wideband and MEC can support.

• Achieving 5G Ultra Wideband’s extraordinary speeds requires a massive fiber infrastructure
• Verizon has made significant investments in fiber densification in major cities around the nation
• Verizon has critical spectrum holdings that include millimeter-wave and C-Band spectrum
• Millimeter-wave spectrum supports 5G Ultra Wideband’s transformative performance and C-band will enable performance and expanded coverage

Small-cell deployment

• Verizon has spent years densifying its 4G LTE network, and its 5G network leverages the same densification
• Verizon has relationships in place with large and small municipalities, enabling its small-cell deployment

Edge computing

• Verizon has network locations that are ideally suited for housing edge-computing resources
• MEC delivers applications closer to the end user while providing access to the tools, power and compute to deploy at scale

This is 5G built right, from the network businesses rely on.

5g nationwide available in 2,700+ cities on most vz 5g devices. 5g ultra wideband (uwb) available only in parts of select cities. verizon 5g access requires a 5g-capable device.  5g uwb access requires a 5g-capable device with select voice/data and 5g uwb plans., future use case not currently available..

Back to categories

Customer Support

If you are already a Verizon customer, we have several options to help you get the support you need.

Enterprise, government and education customers

Fios, internet and phone customers

Small and medium business wireless customers

Let's get started.

Choose your country to view contact details.

• Select Country...
• United States
• Netherlands
• New Zealand
• Switzerland
• United Kingdom

Call for Sales.

1-888-789-1223

Or we'll call you.

Existing customers, sign in to your business account or explore other support options .

• Voice and collaboration
• Contact centers and CX
• Internet of Things
• Fleet management
• Solutions by industry
• Business Devices
• Business Unlimited
• Mobile deals
• Internet deals
• Contact overview
• Request a call
• Wireless store appointment
• Virtual appointment
• Real estate inquiry form
• Support overview
• Register for My Business
• Store locator
• Fios order status
• Tasks without logging in
• Verizon Enterprise Center
• Public Sector support
• Accessibility
• Wholesale and reselling
• Partner network
• Systems Integrator partners
• Technology partners
• For property managers
• Device trade-in
• Wireless return policy
• Refer a business
• Our company
• Corporate responsibility
• Verizon Foundation
• Why Verizon
• Resources and insights
• Report a security vulnerability
• WEA Compatibility
• Service Guide
• facebook-official

• Sign up for email alerts
• Terms & Conditions
• Privacy Policy
• California Privacy Notice
• Health Privacy Notice
• About Our Ads
• Open Internet

• Created with Sketch. Login
• Search Search

Digital twins and 5G in Industry 4.0

This article primarily draws on conversations with key players from Industry 4.0 and academia, held between April and July 2020. Respondents include representatives from a range of industry verticals, with a broad variety of job roles, to gain a snapshot of the existing attitudes and ambitions towards digital twins across both industry and academia.

Digital twins and Industry 4.0

Industry 4.0 hinges on an understanding that digital transformation will create value for enterprises across many verticals. Adoption of new technologies, such as AI, automation, IoT and edge computing, is already starting to help enterprises make more efficient use of their data to optimise processes and find new methods of creating value. Digital transformation requires significant investment and raises questions around when to invest, what to invest in (i.e. which will be the most promising use cases) and how to invest (i.e. which technologies will enable that promised value).

Digital twins is a way to answer such questions by enabling enterprises and developers to predict and understand how to use their existing infrastructure in the most efficient way and how to integrate other technologies to create additional value.

A lack of consensus on digital twins

Consensus as to the meaning of digital twins is elusive, though most agree it is not a static product. Some understand it as a broad philosophy, while others consider digital twins to be an evolution of solutions, where applications change as enterprises progress towards data centricity and transformation.

As a philosophy, digital twins is understood to be based on data-centricity relying on the extraction, management and analysis of data to provide users with insights and predictions on their assets. The following figure illustrates how data is extracted from the physical realm and modelled by the digital twin to enable human or machine interaction: the enterprise can make use of real data from the modelled asset.

Three realms of digital twins, as a philosophy

Source: STL Partners interview programme (April 2020)

See how STL can help you grow your enterprise revenues

Our Growing enterprise solution covers how operators can adopt B2B2X business models to address changing enterprise needs in 5G, private networks, IoT, analytics and more.

As an evolution of solutions, digital twins can be considered to span applications on a four-stage continuum: visualisation, simulation/emulation, twin-to-twin process automation and cross-party communication. STL Partners defines the stages of the continuum as follows:

• Visualisation : the digital twin is used to create new dynamic visual representations, models and virtual dashboards of assets and processes (e.g. 3D schematics/Virtual Reality);
• Simulation/Emulation : the digital twin is used to perform realistic simulation or emulation modelling of assets; digital twins at this stage can enable the organisation to perform detailed ‘what-if’ scenarios to streamline their processes (e.g. R&D);
• Twin-to-Twin (T2T) process automation : digital twins are used to enable real-time analysis and T2T communication to automate certain processes through the creation of closed loops;
• Cross-party communication : digital twins communicate and interact (in real-time) with twins across multiple 3rd parties within the organisation’s ecosystem or supply chain.

Enterprise should view digital twins as an evolution

Research findings suggests it is most useful to consider digital twins as an evolution. STL Partners expects industries to adopt and progress along the digital twin continuum at different rates, depending on their different business needs and challenges, as well as their existing resources and environments.

Some industries, such as manufacturing, aerospace and defence, are leading adoption, while others such as smart cities, construction and logistics, will be one or two years behind. Slower adoption is expected in verticals such as oil and gas and infrastructure where more industry transformation is required.

Early adoption in manufacturing

Almost 50% of respondents in STL Partners’ research programme identified manufacturing as the leading vertical in digital twins adoption. There is a clear role for digital twins in:

• AR/VR and advanced visualisation : using augmented and virtual reality headsets to guide a worker via augmented display and/or a remote expert when carrying out maintenance and repair tasks
• Precision monitoring and control : conducting real-time, granular monitoring of a factory or plant (and robots/machinery) to reduce number of defects and optimise production process
• Advanced predictive maintenance : collecting data (i.e. through sensors) on the condition of machines to predict maintenance requirements and avoid unplanned downtime and associated costs – we explore this use case in more depth, below

Use of predictive maintenance solutions is already common in manufacturing, and in a plant or factory, it can reduce breakdowns by up to 70% and lower maintenance costs by 25% . However, as plants become more ‘connected’ with more sensors, available data will grow, enabling advanced predictive maintenance use cases for digital twins.

In this type of application, an emulation of equipment in the manufacturing plant is created and sophisticated extraction, management and analysis of sensor data is enabled to test different fault conditions. Data is the fundamental aspect of this application: the right data must be quickly moved to where it is required.

Smart cities adoption in near term

Smart city solutions are forecast to be worth over $250 billion by 2025 . Their intrinsic dependence on data, and being able to collect, process and apply it efficiently, makes it a strong candidate for digital twins use cases. These include:

• Connected devices e.g. autonomous drones, for example being explored by Atrius
• Connected vehicles e.g. V2V/X and smart traffic management
• Connected infrastructure e.g. smart building management

Smart city use cases are inherently wide ranging and involve connecting multiple assets, such as buildings, vehicles and devices, to enable more autonomous functions. One example is smart building management, for which there are two key phases: the construction process and the smart building itself.

In the construction phase of a smart building, traditional building information modelling (BIM) only provides a high-level view of a building, while digital twins enables the organisation to design and shape every aspect of the building in a dynamic way. For example, an emulation of the building can be used to demonstrate how aspects such as connectivity and lighting will function to determine the optimal set-up for purpose. This on-site use case would depend on having the right connectivity in place on the building site.

Post construction, digital twins can be used to manage other aspects of a connected building. For example, a network of sensors can collect a variety of different real data sets (e.g. on people flow, temperature, occupied rooms, lights etc.). This can be extracted, managed and stored via digital twins to enable decision makers to interact with the information more easily. Without digital twins, the noise from real datasets could make it prohibitively difficult to extract and interpret data to inform operational efficiency improvements.

As organisations progress along the digital twins continuum and explore more advanced use cases, digital twins could enable the creation of closed feedback loops, automating the management of buildings, to a certain extent (e.g. when a room is empty turn off the lights, when the temperature reaches x degrees turn down the heating). These relatively simple sequences will drive efficiency in the building and help to cut operational costs.

5G supports digital twins use cases

The use cases explored above will depend on connectivity. Take the example of smart cities: it is crucial that relevant data is extracted and moved from A to B at the right time in order to enable mission critical applications such as autonomous vehicles. It is therefore essential that organisations consider the quality of their network as part of a digital twins strategy.

5G’s capabilities (shown below) can enable a number of digital twins use cases that depend on the efficient and reliable movement of vast quantities of data. 5G could amplify the business case for digital twins.

5G’s capabilities support digital twins use cases

Source: STL Partners

Digital twins can support the case for 5G

Digital twins can enable industries to build the 5G business case by emulating 5G network capabilities:

• In manufacturing, this can help to address integration concerns and demonstrate 5G value. Developers and manufacturers can plan, test, configure and optimise their infrastructure and explore 5G use cases without the upfront risk of adoption.
• In smart cities, this can enable organisations and developers to understand the interaction between 5G and other technologies, for example edge computing, so they can explore how workloads can be moved using different network architectures and topologies (find out more about why 5G needs edge computing here ). They can also test 5G-enabled security solutions to ensure against potential failure or breach before network rollout.

This will help to overcome reluctance to invest in new infrastructure and replace existing equipment. In this way, 5G and digital twins mutually benefit one another and may serve to accelerate adoption and innovation in Industry 4.0.

Telcos and digital twins

There are two ways that telcos could leverage digital twins: to accelerate the rollout of 5G and as a digital twins service to offer customers.

1. Accelerating 5G rollout

Many operators have been hindered by a lack of funding for 5G, resulting in slow and staggered deployments. This is partly due to the lack of defined business cases in many regions. Telcos could use digital twins, within their own network, to address this.

For example, there are doubts surrounding the ability of 5G to provide reliable indoor coverage. By creating an emulation of the network, it may be possible to test and address this scepticism. Furthermore, by leveraging digital twins to provide a virtual copy of the network and right time visibility on network performance and key metrics, operators can take a step towards deploying truly cloud native, dynamic and virtualized networks. This will be essential for deploying standalone (SA) 5G and delivering the promised capabilities laid out above. Telcos could also use a network emulation to evidence what industry standards would be needed to enable specific use cases. The use of digital twins could inform decision making in this regard.

2. Providing digital twins services

Telcos could offer enterprises a digital twins private network as a managed service. Digital twins can provide increased visibility and control for enterprise customers as a private network, enabling them to view and interact with their network twin.

Beyond this, digital twins services could also allow telcos to access a wide range of industry verticals, and play in different parts of the value chain, for example in smart cities, where there are many and various use cases and multiple components depend on the same connectivity.

Recommendations

As a result of this research, STL Partners suggests the following:

• Keep abreast of and contribute to work on digital twins in your industry and in cross-industry bodies to learn more about how digital twins can support you in your position on the digital twins’ continuum;
• Consider digital twins to create a holistic view of your environment, including different connectivity solutions and associated characteristics for different use cases (connectivity is a key part of delivering transformative use cases which can drive efficiencies across Industry 4.0);
• Recognise how 5G and digital twins support each other in a virtuous circle – leverage digital twins to understand the business case for 5G and determine how 5G fits within your connectivity ecosystem and existing connectivity solutions;
• Engage with academia on (applied) research activities to progress digital twins’ adoption and usage across the industry (85% of research respondents agreed that academia would be a key partner for organisations and developers at different stages of their digital transformation).

The full research report can be found here .

Read more of our research on digital twins:

• Digital twins: A catalyst of disruption in the Coordination Age

Darius Singh

Consulting Director & Practice Lead at STL Partners

Darius Singh is the Consulting Director & Practice Lead at STL Partners. He has led a range of client projects within edge computing, 5G, and data analytics

Download this article as a PDF

Read more about growing enterprise revenues.

Growing Enterprise Revenues overview pack

Our overview pack explores how the telecoms industry can leverage new business models to meet enterprise customer needs

MNO IoT security roles across the IoT stack

As part of our study IoT security: The foundation for growth beyond connectivity, we looked at the IoT security services of 12 MNOs to see what they are doing to address enterprise IoT security needs beyond that of secure connectivity.

Vonage: What is Ericsson’s end game?

If Ericsson is able to successfully leverage Vonage’s developer ecosystem to drive the use of 4G and 5G APIs, then it could play an interesting role as a strategic partner and channel for operators as a means of accessing developer communities.

Coordination Age: Digital Twin applications for building construction

Sample reports.

Access some of our recent research for free

First Name (required)

Last Name (required)

Company (required)

Your Email (required)

Please let us know what you're interested in here

Internet of Things

Gsma 5g transformation hub case study: how 5g drones will help deliver digital twins.

Tuesday 10 May 2022 ｜ Advanced Air Mobility | GSMA Foundry |

Autonomous drones could save infrastructure owners tens of billions of dollars a year

Skydio is preparing to use 5G to enable its autonomous inspection drones to coordinate their activities and quickly and reliably relay images and scans of critical infrastructure, and construction and accident sites to decision makers. Regular inspections will enable infrastructure owners to create digital twins of their assets, which can be used to guide proactive maintenance and design improvements.

This case study is part of the GSMA 5G Transformation Hub , demonstrating the world’s most innovative 5G solutions.

Like what you read? Share.

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Sensors (Basel)

When Digital Twin Meets Network Softwarization in the Industrial IoT: Real-Time Requirements Case Study

Associated data.

The data presented in this study are available on request from the corresponding author.

The Industrial Internet of Things (IIoT) is known to be a complex system because of its severe constraints as it controls critical applications. It is difficult to manage such networks and keep control of all the variables impacting their operation during their whole lifecycle. Meanwhile, Digital Twinning technology has been increasingly used to optimize the performances of industrial systems and has been ranked as one of the top ten most promising technological trends in the next decade. Many Digital Twins of industrial systems exist nowadays but only few are destined to networks. In this paper, we propose a holistic digital twinning architecture for the IIoT where the network is integrated along with the other industrial components of the system. To do so, the concept of Network Digital Twin is introduced. The main motivation is to permit a closed-loop network management across the whole network lifecycle, from the design to the service phase. Our architecture leverages the Software Defined Networking (SDN) paradigm as an expression of network softwarization. Mainly, the SDN controller allows for setting up the connection between each Digital Twin of the industrial system and its physical counterpart. We validate the feasibility of the proposed architecture in the process of choosing the most suitable communication mechanism that satisfies the real-time requirements of a Flexible Production System.

1. Introduction

The Internet of Things (IoT) is experiencing strong growth in various fields and is constantly evolving, with a forecast of more than 125 billion connected objects in the world by 2030 [ 1 ]. These are mainly wireless sensors connected to the Internet but also any other physical or virtual object that can communicate via this global network. The IoT is foreseen to play a vital role in the fourth industrial revolution with the advent of Industrial IoT (IIoT) paving the way for a wide range of industrial applications to benefit from full automation and high productivity. Powerful industrial systems can be designed through the deployment of wireless sensors, actuators, controllers and other smart devices.

Facilitated by this dramatic development of the IIoT along with the rise of big data analytics, the last decade witnessed the revival of the concept of Digital Twining (DT) [ 2 ]. In particular, the IoT allows for keeping a digital twin consistent and synchronized with the physical entity it represents thanks to its sensing technology coupled with the communication capabilities it provides. The digital and physical twins in addition to the IoT ensuring the twins connection form a Cyber-Physical System (CPS). Digital Twining is ranked by Gartner [ 3 ] as one of the top ten most promising technological trends for the next decade. Digital Twining is particularly promising in creating a continuously updated model of a physical system to enable rapid adaptation to dynamics mainly unpredicted and undesirable changes. A wide range of industrial fields are concerned [ 4 ] such as manufacturing [ 5 , 6 , 7 ], healthcare [ 8 , 9 ], maritime and Shipping [ 10 , 11 ], city management [ 12 , 13 ], and aerospace [ 14 , 15 ].

In Cyber-Physical Production Systems (CPPS) and industry 4.0, digital twins are made for physical assets that compose the industrial system. The major effort is underway to narrow any gaps that may occur between the twins. Some CPPS architectures are proposed [ 16 , 17 ], however, the communication network connecting the twins was omitted despite its vital importance in the whole CPPS. Recently, some DT architectures for networks have been proposed. The architecture proposed in [ 18 ] only aims to provide adaptive routing in software defined vehicular networks. While the Internet Engineering Task Force (IETF) (IETF is a large open international community of network designers, operators, vendors, and researchers concerned with the evolution of the Internet architecture and the smooth operation of the Internet) draft [ 19 ] presents a reference architecture of a Digital Twin Network, the work is not yet mature and it does not target the Industrial IoT with a holistic approach. Once again, ref. [ 20 ] discusses the opportunities that could provide Digital Twinning to fulfill the potentials of 5G networks without considering an industrial system.

In this paper we propose a holistic Network Digital Twin (NDT) architecture for the IIoT to enable closed-loop network management across the entire network life-cycle. This allows for movement from the current network design methodology to a more dynamic one. In fact, the NDT allows to leverage the output from both twins to suggest improvements on the designed networking protocols and algorithms. An ongoing evolution of the physical network is possible by taking actions during the network service phase with the objective of maximizing the performance. In practice, we chose to leverage the Software Defined Networking (SDN) paradigm as an expression of network softwarization. SDN decouples the data plane from the control plane and allows centralized network orchestration [ 21 ]. Controllers form the control plane and hold the control of the network by sending instructions and commands to the devices in the data plane. They also collect the required information from the physical network to build a centralized global view. This two-way connection can be exploited by the NDT for real-time network monitoring, predictive maintenance mechanisms, and network diagnostics.

To validate our proposed architecture, an industrial project aiming to connect a Flexible Production System (FPS) to the Internet using sensor networks is considered. The concept of NDT is used in the early stage of the project. One design issue consists in choosing the communication mechanism that suits the real-time requirements of the FPS application. An NDT is built to allow the assessment of different networking policies that aim to achieve reliability and timeliness prior to the deployment of the most suitable one. To the best of our knowledge this is the first work that introduces the concept of network digital twin for the IIoT based on a holistic approach. The remainder of the paper is organized as follows. Section 2 presents some important concepts used to design our architecture along with some related work. Section 3 describes our proposed architecture while Section 4 details the use case we considered. Section 5 concludes the paper and discusses some future work.

2. Background and Related Work

In this section, we present some notions that make the fundamental bricks of our architecture. First of all, we provide a brief definition of the SDN paradigm along with its architecture. Then we provide an overview of the Digital Twinning concept covering its origins and its main notions. Finally, we discuss some existing network specific DT architectures as related works and the main characteristics that make our architecture original.

2.1. Software-Defined Networks (SDN)

According to the Software-Defined Networking paradigm, computer networks can be split in three planes: management, control and data planes as depicted in Figure 1 . The data plane is composed of the forwarding devices (switches, routers, etc.). The control plane is responsible for sending commands to the forwarding devices in order for them to apply the required networking policy. The management plane is represented by network management applications such as those related to traffic engineering, mobility management and wireless communications, security, and reliability. These applications are implemented using network programming languages and they interact with the control plane thanks to an open northbound Application Programming Interface (API). The control plane controls the forwarding devices via an open southbound API such as OpenFlow [ 22 ].

Simplified view of SDN architecture.

This three-plane division allows for more flexibility, scalability, and performance compared to traditional networking. In fact, SDN abstracts network intelligence from forwarding devices to a logically centralized controller. This simplifies network management, configuration, and evolution [ 23 ]. It can be defined as a network architecture with four characteristics according to Kreutz et al. [ 21 ]: (i) the control and data plane are decoupled; (ii) flow-based forwarding, meaning that routing decisions will be made based on packet properties instead of its destination, which offers high network flexibility; (iii) control logic is moved to an external entity, called SDN controller, which facilitates the programming of network devices by providing an abstract global view of the network; (iv) the network is programmable by the means of software applications implemented on top of the control logic.

2.2. Digital Twining

A Digital Twin (DT) can be viewed as a machine that is emulating or “twinning” the life of a physical entity [ 24 ]. A DT is more than just a simple simulation or a static model, it is a continuously evolving model that is always aware of the events happening in its physical twin as it follows its lifecycle to supervise and optimize its functions. The synchronization between the DT and its physical counterpart is possible thanks to the real-time data uploading ensured by IoT devices and sensor technology, while big data storage capabilities allow keeping historical data that can be useful for the Digital Twin. AI algorithms, for instance, can be used to predict future states of the physical twin. A DT can also simulate new configurations in order to apply preventive maintenance operations [ 24 ]. Theses properties result in reducing costs and resources in many industries such as manufacturing and healthcare systems.

DT as a concept includes a real space, a virtual space and information/data exchange between the two spaces. It was first introduced in 2002 by Professor Grieves in the Product Lifecycle Management (PLM) course of the University of Michigan [ 25 ]. The concept went through a slow-going development phase from 2003 to 2011 where only few articles were published [ 26 ]. Right after 2011, the availability of low-cost sensors and communication technologies in the era of the IoT, the emergence of big data analytics and simulations technologies, the remarkable advance in Machine Learning (ML) along with powerful computation infrastructure, triggered the rise of DT technology. It gained widespread popularity among researchers in both academia and industry when NASA gave a formalized definition of DT and how it would ameliorate performance in the astronautics and aerospace field [ 27 ]. In 2014, Grieves published the first white paper extending the DT concept from one conceptual idea to various practical applications [ 28 ]. Since then, numerous DT applications have appeared and DT technology has been classified as one of the top ten most promising technology trends in the next decade by Gartner in 2017 and 2018 [ 3 ]. More recently, control-oriented DTs are emerging. In [ 29 ] is proposed a digital twin offering uncertainty management and robust process control.

2.3. Digital Twin Architectures

Some network specific DT architectures have been proposed in the literature. A Digital Twin enhanced Industrial Internet (DT-II) reference framework towards smart manufacturing is proposed in [ 30 ]. The interplay between digital twin and Industrial Internet is discussed in three aspects of product, enterprise, and business. On the one hand, they show how the sensing/transmission network capability of Industrial Internet can provide data acquisition and transmission to digital twins. On the other hand, they show how the physical-virtual real-time synchronization feature of digital twin is consistent with the interconnection of intelligent machines required by Industrial Internet. The Intelligent Digital Twin-based Software-Defined Vehicular network (SDVN) [ 18 ] has been designed to allow intelligent and adaptive routing in vehicular networks. By having a DT that replicates the vehicular network in real-time, a report is sent to it when a routing error occurs. The DT verifies different routing schemes before choosing the one with the best performance to deploy in the physical network.

The IETF group is currently working on standardizing the concept of Digital Twin Network (DTN). The last draft [ 19 ] presents the basic concepts and a reference architecture along with the main challenges for building a DTN. The presented DTN in this work is for computer networks in general and does not consider the specificity of the IIoT. A 5G Digital Twin is presented in [ 20 ] to help developing and deploying 5G complex networks and permit cost-effective access to 5G knowing that deploying 5G networks is too expensive. In [ 31 ], an Application-driven Digital Twin Networking (ADTN) middleware is proposed in order to simplify the interaction with heterogeneous distributed industrial devices and dynamic management of network resources by adopting an application-level point of view. This architecture adopts an SDN-based cross-layer approach to dynamically orchestrate the industrial environment while taking into consideration Quality of Service (QoS) requirements and network configuration adaptation capabilities. The leading communications service provider, Huawei, proposed Intent-Driven Network (IDN) [ 32 ] which builds a digital twin between the physical network and business intent with the goal of helping enterprises accelerate business innovation and boost operation efficiency. The built digital twin could enable quick detection and resolution of network problems, prediction of future network status and enhancement of network reliability by eliminating network risks in advance.

Our architecture is targeted to IIoT networks, uses SDN as the communication interface between the digital and the physical world and considers the interaction between the Network Digital Twin and other Digital Twins in an industrial environment.

3. A Holistic Digital Twinning Architecture for the IIoT

The process of designing and validating network solutions can go through a theoretical analysis as a preliminary step to prove the underlying algorithms convergence and their correctness. On the other hand, simulation (or emulation) tools are widely used by network researchers to develop and evaluate their algorithms and protocols. This is due to the fact that these tools are a good way to quickly test protocols on a large scale at a low cost. To evolve the developed solution, the process is repeated as in Agile methodology using inputs from the previous steps and eventually from experimental validation and deployment steps, until it is fully functional and ready for deployment. These iterations are done off-line implying human intervention which makes this approach prone to errors. We argue that the problem of this methodology is the lack of connection with the real world network throughout the entire life cycle from the first to the final phase. In other words, the coordination between the different steps can be quite challenging. Moreover, with this approach, the designed solution can only be completely validated at the end of the deployment phase. This is further exacerbated in the context of Wireless Sensor Networks (WSN), a basic building block of the IoT.

We estimate that the industrial IoT is a complex system since it is characterized by a large network of components, many-to-many communication channels and sophisticated information processing that makes prediction of system states difficult [ 33 ]. We would contend that complex systems have a major element of surprise, as in “I didn’t see that coming”. That surprise is generally, although not always, an unwelcome one. So the element of surprise is not to be ignored. In fact, in the real world many factors can impact the operation of a WSN: outdour conditions, weather conditions, radio interference from other wireless technologies, etc. The LOFAR-agro project presented in [ 34 ] is the perfect example on how things can go extremely wrong after deployment. The project members intended to deploy a large scale WSN of up to 100 nodes for a pilot in precision agriculture. Everything worked fine in simulation and in short-scale deployment (10 nodes), but when coming to the real world deployment, they faced an endless stream of hardware malfunctions, programming bugs, software incompatibilities, combined with the harsh nature conditions and time pressure. That made them face unsolvable problems due to the layering of the different problems and, in our opinion, the lack of continuous connection between the real world network and the design/validation process.

In the networking field, it is known that simulation is not a tool for fully validating a solution since it cannot take into account all the environmental variables surrounding the network. Although, it helps gaining a better understanding of current performances. On the other hand, deployment testing is costly in terms of time and money. That is why, we are proposing a novel Digital Twinning based architecture for the IIoT that should permit closed-loop network management across the entire network life-cycle from the early design stages to the service and maintenance phases. To do so, we suggest to introduce the concept of the Network Digital Twin (NDT).

By creating a digital twin of the industrial network, a ’living model’ that is kept constantly updated, decisions are, therefore, made based on current conditions rather than those of the original study. As depicted in Figure 2 , modeling and analysis can be tightly coupled with execution, enabling a cycle of continuous improvement and innovation. Enhancing network reliability and dealing with network risks in advance by predicting future network status using, for instance, AI algorithms in the digital twin. Improved performance can also be achieved by adjusting network configuration based on different options, adaptation to evolving traffic and resource demands and by experimenting safely different solutions to determine the optimal configuration of networks without jeopardizing the operation of the physical network. This eliminates the risks related to testing new network policies in a production environment and decreases the corresponding costs since the experiments are done in the network digital twin.

NDT role throughout the whole network development process.

In order to implement the proposed design approach, we propose a holistic architecture composed of three main parts as shown in Figure 3 . A physical world composed of physical plants which could be any industrial object/system such as a 3D printer, an oil platform, a conveyor, a machine, etc. Each physical plant is equipped with wireless sensor nodes that are responsible of collecting data on their operating conditions. The second part consists in the cyber world that integrates a digital twin for each industrial system present in the physical world all interacting with the NDT, the digital twin of the industrial network. The physical and cyber worlds are connected via an SDN controller that acts as a bridge between the two worlds, forwarding information flows from the cyber world to the physical world and vice-versa. The SDN paradigm is adopted since it facilitates the management of networks, enables network centralization, allows network programmability and also network slicing when combined with Network Function Virtualization (NFV). With this approach, all the data describing the physical network is captured by the SDN Controller that constructs the network topology model and provides the necessary intelligence to the NDT.

Digital Twin Networking based architecture for industry 4.0.

When it comes to practical implementation of our architecture, we need to consider the constrained nature of the IIoT in terms of computation and storage means as well as in terms of energy. By including an NDT in the architecture, more data has to be exchanged in a bidirectional way with more packets processing to ensure the synchronization between the physical and digital network. This leads to increasing the traffic load in the network and the nodes energy consumption. So, the challenge when implementing an NDT would be to find a trade-off between increasing energy consumption and ensuring the digital twinning operations. For example, the NDT would provoke more energy consumption in the early stages of the network operations (due to the high amount of information that should be exchanged to synchronize the two sides) but once it becomes stable, it can apply mechanisms that should increase the network’s remaining useful life.

In what follows, we give some of the benefits we can obtain when adopting this architecture in the design and the service phases. The latter includes production and maintenance.

3.1. Design Phase

With the proposed architecture, it is possible for the NDT to leverage collected data from the real world to provide insights on the changes to be made to the network solution being designed in order to ensure the intended operation and get the required performance, thus validating the solution more quickly. Moreover, data visualization and analysis tools can be implemented in the NDT which may help network designers to accurately interpret the behavior of a network protocol and analyze the interactive behaviors among the network components. More interestingly, the NDT would permit testing new network solutions under various conditions, having more accurate results than current simulation tools because the NDT can take into consideration the surrounding environmental conditions and provides more immersive experiments thanks to its permanent connection with the real world. This helps network designers to detect and eradicate the eventual surprising undesirable behaviors that could occur in the deployment environment. Last but not least, the NDT can interact continuously with industrial systems digital twins to get insights on the networking requirements of each one and adjust the network’s resource allocation policy according to that.

3.2. Service Phase

Our architecture allows a continuous evolution of the physical network. In fact, the NDT would provide continuous real-time remote network monitoring based on the information it receives from the physical network. In addition, predictive maintenance mechanisms could be implemented in the NDT using, for instance, AI algorithms and specific network policies could be applied to increase the network remaining useful life. Network diagnostics can also be ensured by the NDT. Based on the reports generated, it adapts continuously the implemented mechanism to improve their efficiency. Also, the NDT can take action to ensure that the requirements of network applications are answered by adapting the network configuration. For instance, when multiple network applications are running on the same network stack, the NDT can provide a network configuration that makes a balance between the current network capacity and the applications requirements.

4. Industrial Case Study

The proposed architecture can be applied in many DT-based architectures to allow an efficient connection between the real and the digital worlds. In the personalized production and distributed manufacturing context for instance, a digital twin for a connected micro smart factory is designed and implemented in [ 35 ]. The DT uses an IIoT network to ensure the synchronization with the physical manufacturing components. This synchronization allows the DT to monitor the present in real-time, to track the past, and make predictions to support decision-making for the future. The NDT concept can be included in this architecture to manage the IIoT network and boost its performance.

In order to validate the proposed Network Digital Twin architecture, we consider its application to the early design stage of an industrial project. The purpose of this project is to satisfy the real-time requirements of a control application that monitors the operation of a Flexible Production System (FPS) [ 36 ]. To do so, a WSN needs to be deployed to allow collecting information on the manufacturing operations carried out by the FPS. One raised question concerns which mechanism allows to meet our real-time requirements. As a preliminary step, a Network Digital Twin is built to assess the performance of the platform equipped with the WSN under three MAC protocols along with an oversampling mechanism. Figure 4 depicts the practical scheme of our proposed NDT architecture adapted to the considered industrial case study.

Practical architecture adapted to our industrial case study.

In the physical world, there is the industrial platform that consists in an FPS installed in approximately a 20 m 2 area within our university. This platform aims to support teaching activities in automation engineering, industrial supervision, industrial communication, control and system integration. The FPS is composed of six assembly stations connected by a conveyor where each station is equipped with a Programmable Logic Controller (PLC). A sensor node is installed at each station in order to report information on its operation process to a central node suspended in the ceiling centrally above the FPS.

In order to carry out the design of our project, we proceed by creating an NDT to get insights on the most appropriate choices with regard to network protocols to satisfy the real-time constraints of our FPS. That is, in the cyber world, cooja simulator [ 37 ] is used in order to replicate the behavior of the WSN deployed in the physical world. To do so, the different distances between the sensor nodes and the central node (the Sink) are measured and the corresponding topology in cooja is reproduced as shown on the left side of Figure 4 . Green node numbered 1 is the Sink and it is located at an altitude of 2.35 m with respect to sensor nodes.

In order to allow communication between the real and the cyber worlds, SDN-WISE [ 38 , 39 ] is used as it is more suitable to WSNs. SDN-WISE focuses on network flexibility and security with the goal of making WSNs modular in terms of communication and processing, reducing the amount of information exchanged between the sensor nodes and the SDN controller, and making the nodes programmable as finite state machines. The SDN-WISE controller keeps track of the network topology using a graph where vertices are the nodes and the edges are the links between these nodes. In SDN-WISE, the sink is the intermediary between the sensor nodes and the controller. It starts by broadcasting a packet called “beacon” that contains the identity of the sink that generated it, a battery level, and the current distance from the sink which is initially set to 0. A neighbor node, upon receiving such a packet, inserts the source node in its neighbor nodes list. If the current distance from the Sink is better than the one it holds then the source node becomes its next hop to the Sink. The current distance is incremented in the beacon packet before it is rebroadcast. After constructing its list of neighbors using the different “beacon” packets received, a node generates a “report” packet containing its current list of neighbors and sends it to the controller. This latter, uses report packets to construct a global view of the network. This protocol is run periodically to ensure that the controller always have an updated view of the network. The frequency of sending “beacon” and “report” packets are application specific and impacts the performance of the network.

The process of building a global view of the network is a key feature in the proposed architecture. Even if in the considered case study, the network topology was manually defined, one can make automatic topology discovery. This may be useful to consider already deployed networks in harsh environments. More interestingly, any change in the physical network, mainly during the service phase, would be detected by the controller which updates the digital replica consequently. These dynamics can be accommodated by the NDT and the running solution can be updated accordingly. In the proposed practical implementation, the network topology can be described within the XML (Extensible Markup Language) file that describes the scenario to run by the cooja simulator.

4.1. Real-Time Requirements

In an industrial environment, the control of automation applications is usually based on cyclic processes that run according to a predefined sampling. The sampling period must allow the control system to be updated while counting for the communication overhead. In order for the system to be updated every period, it must therefore be ensured that all communications arrive within the period. If a network message gets lost then the controller will be deprived of fresh input data and/or a remote control action will not be executed. Reliability and timeliness are of a paramount importance in an industrial application. Since, we aim to endow our FPS with a WSN, it is worth noting that real-time communication in WSNs is more challenging due to their severe constraints in terms of processing and communication means, in addition to the unreliable nature of the wireless medium and its shared access.

To overcome the above mentioned problem, a common solution is to apply an oversampling mechanism where a message is sent more than once in the sampling period to ensure the timely delivery of at least one copy of this message. The drawback of this solution is that sending multiple copies of each message within a given period would lead to network overloading due to congestion which decreases reliability and increases the experienced delays. As a result, a careful setting of the amount of redundancy to apply is crucial to ensure timeliness without affecting the application reliability.

Another solution consists in adopting a contention free access protocol. Recently, the IEEE 802.15.4e amendment [ 40 ] introduced several channel access modes that are contention free. This includes Time Slotted Channel Hopping (TSCH) which received the attention of both academia and industry due to its high performance in real-time constrained applications. TSCH targets application areas such as industrial automation and process control and offers support for multi-hop and multi-channel communications [ 41 ]. It has been designed to satisfy the requirements of IIoT applications as it provides time critical assurances and very high reliability [ 42 ]. It schedules data communication between network nodes by combining time slotted access with the channel hopping mechanism. The first mechanism avoids collisions between competing nodes, so it increases throughput and provides deterministic latency to applications. The channel hopping mechanism in turn allows multiple nodes to communicate simultaneously using different channels. Therefore, it increases the capacity and reliability of the network by mitigating the negative effects of interference.

In TSCH, nodes synchronize to a periodic slotframe composed of a number of timeslots, which is repeated throughout the network lifecycle. Communication between the nodes follows a schedule that is defined to allow the nodes to communicate as efficiently as possible. This schedule can be modeled by a matrix whose rows are the available channels and columns are the timeslots in a slotframe. Each cell of the matrix represents a specific link having as coordinates (ChannelOffset, SlotOffset) and can be reserved for a single link or shared by several links. The CSMA-CA algorithm is executed if a collision occurs in the latter case. The frequency in which two nodes can communicate in a timeslot is calculated as follows:

where ASN is the total number of timeslots that have elapsed since the start of network service, incremented in each timeslot. Function F can be implemented as a look-up table, usually it is defined as the hopping sequence specified in TSCH. For example, if four channels are used, the hopping sequence can be { 15 , 20 , 25 , 26 } . It should be mentioned that Equation ( 1 ) can return different frequencies for the same link in different timeslots, which ensures the channel hopping mechanism.

4.2. Scenarios and Metrics

To answer our question on which solution would satisfy the real-time requirements of our industrial platform, a progressive methodology is followed. Based on the obtained results at each step, we decide whether a new mechanism is to be investigated. CSMA and TSCH minimal schedule (The TSCH minimal schedule is composed of one slotframe with three timeslots and only the first one is used and shared between all nodes) are considered first as they are already provided in cooja. Afterwards, a centralized scheduling algorithm called TASA (Traffic Aware Scheduling Algorithm) [ 43 ] is implemented and evaluated at the SDN controller. Table 1 presents TASA parameters setting.

Parameters used for TSCH with the Traffic Aware Scheduling Algorithm (TASA) scenario.

The platform’s refreshing time is 2 s so the main data rate is one packet every 2 s. To assess the oversampling mechanism, we considered both duplicating and triplicating a data packet within a 2-s observation period. The data transmission rate without oversampling is one packet every 2 s. In the case of duplication the rates becomes 1 pps (packet per second) where one packet is sent at the beginning of the period and the second message with an offset of one second. When triplication is performed, the first message is sent at the beginning of the period, the second with an offset of 2 / 3 s, and the third with a 4 / 3 s offset, resulting in a data rate of 3 / 2 pps. The purpose of these offsets is to avoid creating congestion in the output buffers. Thus in the case of the duplication for instance, the first message meets the requirements if it is received in less than two seconds, while the second message will only have one second to arrive within the time limit. It is sufficient for one of the messages to meet these constraints to conclude that the system has been refreshed. The reported data from the FPS to the controller are of boolean type, that is the payload of each transmitted packet is one byte long. Note that additionally ten bytes are used by the SDN-WISE header. Every scenario has been executed during 40 min and repeated four times with a different random seed in each run. The process of sending data packets starts after an initialization phase of six minutes.

In addition to two network performance indicators, namely packet delay and packet delivery ratio (PDR), an application performance indicator called the freshness indicator (FI) is considered. The packet delay is the difference between the time of its reception by the Sink and its transmission time by a sensor node. The PDR is the ratio of the number of received packets by the Sink to the number of sent packets by the different sensor nodes of the platform. The FI is defined as the ratio of the number of periods in which at least one packet is received by the Sink within the period duration to the total number of periods of the whole simulation. In what follows, the simulations results are presented using box plots in order to visualize the distribution of obtained data points. Mean values are also plotted as empty squares inside the box plots.

4.3. Simulation Results

4.3.1. csma.

Experiments are first conducted using CSMA as MAC protocol without oversampling (i.e., using a data rate of 1p/2s) then an oversampling with 2p/2s and 3p/2s data rates is considered. Figure 5 plots the obtained delays (log-scale) when using CSMA for each sensor node numbered 2 to 7 in the x -axis. The horizontal line at ordinate 2 s recalls the refreshing time period. It is noted that all nodes obtain similar latencies for the different settings. This is due to the fact that they are located almost at similar distances from the Sink reached in one hop resulting in a star topology. In the absence of oversampling (1p/2s), an average delay of 719 ms with an average median delay of 575 ms is obtained. However, it is observed that some packets (about 6 % as shown in Table 2 ) arrive after a delay that exceeds 2 s. This translates into an average freshness indicator of 97 % while the PDR achieves its maximum value ( 100 % ). The distribution of PDR and FI obtained in the different experimented MAC protocols with or without oversampling are presented in Figure 6 and Figure 7 , respectively.

Packet Delay—CSMA.

PDR results.

FI results.

Percentages of packet delays exceeding 2 s for each scenario along with their respective PDR values.

Since no loss is experienced, the duplication of the transmission rate (to 2p/2s) can be afforded by sending each packet twice in the 2-s period window. Not only a PDR of 100 % is kept but also a 100 % of freshness is achieved for all nodes. Delay results show an increase of the mean values, with an average of 931 ms, but the median values decreases to 696 ms. This latter explains the obtained results of the FI even if some packets still arrive out of time as depicted in Figure 5 . Certainly, at least one copy of each message is received and losses only concern duplicates. Duplication increases the probability that a message is received within the 2-s window. Given the obtained results, triplicating seem to be of no interest. This is confirmed by our experiments where an increase is noticed in the delay results with 33 % of packet delays exceed 2 s and a decrease in terms of PDR as an average value of 97.5 % is obtained. Moreover, a significant decrease in the FI ( 77 % in average) is to be noted. The increase of delay values is due to more experienced collisions that result from overloading the network by higher data rates. In fact, CSMA uses a contention window that imposes a waiting time (back_off time) for nodes to avoid collisions. This window doubles in size every time a collision occurs which causes greater delay values.

4.3.2. Minimal TSCH

As opposed to CSMA, TSCH has been introduced to meet the real-time needs of industrial applications as the nodes do not compete to access the medium. We decided to consider a TSCH-based MAC protocol in our study. Its minimal scheduling available in the Cooja simulator is considered first. Figure 8 presents the obtained results where lower delays can be observed when compared to CSMA. For instance, without oversampling (1p/2s), it is obtain in average, a mean and median delay of 133 ms and 56 ms, respectively. Among received packets, only 1 % are out of time as shown in Table 2 . Despite that, the achieved PDR and FI as shown in Figure 6 and Figure 7 are lower with an average value of 86 % and 85 % , respectively. Even worse, the minimum value for FI may drop as low as 60 % because of 1% of the packets that arrives with a delay that exceeds 2 s.

Packet Delay—Minimal TSCH.

At this stage, is it worth applying oversampling? A priori no, but we further increased the transmission rate to confirm our assumptions. When duplication (2p/2s) is considered, the delays are slightly increased when compared to the case of 1p/2s with an average of 189 ms and a median of 57 ms. Duplication allows a slight improvement of the FI from 85 % to 87 % of the periods getting fresh data as shown in Figure 7 but decreases the PDR to 85 % as shown in Figure 6 . It is noted that the FI box plot spreads over a wider range with a minimum value of 66 % and a maximum value of 98 % . The former value results from the 2 % of packets with delay exceeding 2 s and the latter (as well as the median) value is due to the fact that the probability of receiving a data packet within the two-second window is increased. Going further and triplicating messages (3p/2s) is not worth doing either since higher delays are experienced with lower PDR and FI values as we obtain 74 % and 71 % in average, respectively.

The increase in delay when applying oversampling is due to the fact that each sensor node has more packets to transmit within the same duration which adds to the delay, the time each node spends waiting turn to transmit. The experienced losses can be explained by the fact that the scheduling with Minimal TSCH is not optimal.

4.3.3. TSCH TASA

The obtained results when using TSCH with its minimal schedule is inefficient and can not suit our needs. This is why, we considered implementing and testing a more advanced scheduling algorithm called TASA. As done with the previously considered MAC protocols, we began by experimenting TASA without oversampling i.e., using a data rate of 1p/2s. The obtained delays as shown in Figure 9 fully satisfy our real time requirement as all packets arrive with a delay that does not exceed 2 s. Even better, almost 75 % of the packets experience a latency below the 200 ms. An overall average and a median of 178 ms and 148 ms are recorded. Both the PDR and the FI achieve their maximum values ( 100 % ) as shown in Figure 6 and Figure 7 , respectively. As a result, this configuration (i.e., TASA without oversampling) answers perfectly the real-time requirements of the industrial system of interest, the object of this study. Both reliability and timeliness are achieved with the minimum transmission rate. This allows to not consume extra energy that could have been consumed by the oversampling for packets replications that are no longer needed.

TSCH/TASA delay results.

Following the obtained results, the corresponding firmware is uploaded to the FPS sensors using the SDN controller. This latter will have in charge to monitor their operation based on the reports sent periodically. When an anomaly is detected, the designer is alerted to correct and/or adapt the current implementation. The newly obtained firmware is then uploaded and a new “agile” iteration is undertaken. It is worth noting that AI algorithms can be leveraged to automate this process operations as much as possible in order to gain more efficiency.

5. Conclusions

In this paper, we proposed a holistic digital twin architecture for the IIoT where the network component is considered through the adoption of the concept of a Network Digital Twin (NDT). The aim is to enable quick validation of networking solutions in an industrial environment since the NDT is continuously linked to the physical world from the early design stage to the production phase. Moreover, the NDT provides a continuous evolution of the industrial network by exploiting the data collected along with AI algorithms to enhance the networking performance, prevent network failure and increase the network remaining useful life. The proposed NDT can be included in any DT-based architecture where a communication network is required in the physical world.

We validated the proposed architecture through its application to the early design stage of an industrial project with real-time requirements. Precisely, we used the concept of NDT to assess different scenarios in order to choose the best suited communication mechanism satisfying the real time requirements of the FPS application. We mainly found that oversampling is no longer required when using a TDMA-like MAC protocol such as TASA. The next step consists in undertaking subsequent iterations to consider the effectiveness of the proposed solution in the real network. More elaborated network topologies will be considered to assess the scalability of the proposed solution to more complex industrial systems. Moreover, the adaptive character of the NDT in response to failures in the network in addition to its dynamics will be investigated possibly using AI-based solutions for network diagnosis and prediction. A new communication scheduling algorithm will be designed to further improve the performance of the industrial application. This would include an adaptive TSCH scheduler that exploits the data collected as an input to the NDT to provide the most efficient communication scheduling to the sensor nodes. Finally, it is envisaged to enhance the proposed architecture where a tight collaboration between the different DTs including the NDT with the aim of achieving a better orchestration of the whole industrial system.

Acknowledgments

We would like to thank The AIP-PRIMECA Lorraine for their support regarding the industrial case study.

Author Contributions

Conceptualization, M.K., M.M. and E.R.; methodology, M.K., M.M. and E.R.; software, M.K.; validation, E.R. and M.M.; investigation, M.K.; data curation, M.K.; writing—original draft preparation, M.K.; writing—review and editing, M.M.; visualization, M.K.; supervision, E.R. and M.M.; project administration, E.R. and M.M. All authors have read and agreed to the published version of the manuscript.

This research received no external funding.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not Applicable.

Data Availability Statement

Conflicts of interest.

The authors declare no conflict of interest.

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Published: 26 March 2024

The increasing potential and challenges of digital twins

Nature Computational Science volume  4 ,  pages 145–146 ( 2024 ) Cite this article

2472 Accesses

4 Altmetric

Metrics details

• Climate sciences
• Computational biology and bioinformatics
• Computational science
• Engineering
• Environmental sciences
• Social sciences

This issue of Nature Computational Science includes a Focus that highlights recent advancements, challenges, and opportunities in the development and use of digital twins across different domains.

More than 50 years ago, as part of the Apollo 13 mission, NASA used high-fidelity simulators controlled by a network of digital computers to train astronauts and mission controllers. These simulators were particularly important for testing scenarios of failure and refining instructions that determined success in critical mission situations. During the mission, an in-flight explosion critically damaged the spacecraft’s main engine, and mission controllers used data from the spacecraft to modify their simulators in order to reflect the condition of the corresponding physical counterpart 1 . Ultimately, the simulations were used as a tool to help inform the decisions that safely brought the astronauts back home.

The term ‘digital twin’ would only be coined more than 40 years later 2 , but the Apollo simulators were already notable examples of the concept behind this term (although this has been recently disputed 3 ). After all, these simulators represented a set of virtual assets that were supposed to mimic the structure and behavior of their corresponding physical asset (meaning, the spacecraft), and most importantly, that received feedback of data from the physical asset to inform critical decisions that realized value. It is worth noting that these simulators were more than ‘just traditional model simulations’: they were designed to acquire and process real-time data to help mission controllers with responding to high-stakes situations and making real-time decisions that, in turn, would directly influence the physical asset. The aforementioned events of Apollo 13, for instance, happened over just a few days.

Since then, there have been major developments in the technologies that surround digital twins. We have experienced an ever-growing amount of data being generated and made available, including in real time. We have implemented sophisticated modeling capabilities and observed the sharp rise in data-driven methodologies, including machine learning, which has allowed us to take advantage of the data deluge. While NASA made use of state-of-the-art telecommunications technology at the time of the Apollo 13 mission, we now have access to advanced Internet of Things networks that can substantially accelerate data movement. The list goes on and on.

Another important development is in regards to applications. The engineering and industrial domains have arguably leveraged digital twins for longer, such as for developing, testing, and maintaining aircraft and spacecraft in aerospace engineering, and for optimizing product life-cycle management in manufacturing systems (the concept of a digital twin in this context was first introduced by Michael Grieves 4 , before the term was coined). More recently, however, many other distinct areas of science have realized the potential of digital twins, from biomedical sciences to climate sciences and social sciences. For instance, digital twins could enable improved precision medicine, more accurate weather and climate predictions, and more informed urban planning.

Undoubtedly, all of these developments bring exciting new capabilities and opportunities for digital twins — but not without myriad challenges. This issue presents a Focus that highlights these recent developments within this burgeoning field, bringing together experts’ opinions on the requirements, gaps, and opportunities when implementing digital twins across different domains.

Advancements and current challenges for industrial applications of digital twins are discussed in a Perspective by Fei Tao and colleagues. Digital twins have become very popular in industry and manufacturing, with different conceptual models proposed in the past and large investments from many well-known companies being made within this space. Nevertheless, according to the authors, we still have a long way to go to improve the maturity of digital twins and to facilitate large-scale industrial applications. Among the many challenges and opportunities that still need to be addressed, the authors argue that the trade-offs between overly simplistic models (which are less expensive, but less accurate) and overly complex models (which are more accurate, but can be prohibitively expensive) need to be well-understood and evaluated on a case-by-case basis; that the opportunities and risks brought by artificial intelligence need to be better assessed; and that validation benchmarks and international standards are urgently needed to make the field more mature.

Across aerospace and mechanical engineering, the use cases for digital twins are vast, and the potential benefits, gaps, and future directions in these domains are highlighted in a Perspective by Karen Willcox and Alberto Ferrari. Notably, the authors advocate for the value of considering the digital twin as an asset in its own right: similar to how cost–benefit–risk tradeoffs are employed in the design and development of physical assets, the development and the life cycle of digital twins cannot be an afterthought and must involve investments paralleling those that are made over the life cycle of a physical asset. The computational cost and the complexity of digital twins are also discussed by Willcox and Ferrari: they argue that, to satisfy stringent computational constraints, the use of surrogate models — meaning, approximation models that behave similarly to the simulation model but that are computationally more accessible — may play an essential role, and that for complex systems it is not beneficial to see a digital twin as an identical twin of a physical asset — instead, a digital twin must be envisioned to be fit for purpose, depending on cost–benefit tradeoffs and required capabilities.

The applications of digital twins in the biomedical sciences are explored in the Focus by Reinhard Laubenbacher and colleagues in a Perspective . The authors argue that, different from industry and engineering, there is no broad consensus as to what constitutes a digital twin in medicine, mainly due to some of the unique challenges faced in the field, including the fact that the relevant underlying biology is partially or completely unknown and that the required data are often not available or difficult to collect, with the latter challenge impacting the exchange of data between physical and digital twin. That said, the authors discuss many potential promising applications for medical digital twins in curative and preventive medicine, as well as in developing novel therapeutics and helping with health disparities and inequalities.

The definition of a digital twin is also examined by Michael Batty in a Perspective , this time in the context of urban planning. Batty argues that, while the coupling between real and digital tends to be strong and formalized for physical assets, the same is not true for social, economic, and organizational systems (such as cities), since the transfer of data is often non-automated. Batty also discusses the need of the human in the loop in the design and use of digital twins, and the fact that cities may be intrinsically unpredictable, which brings challenges to applying the standard definition of digital twins to the field. On the other hand, Luís M. A. Bettencourt argues in a Comment that cities do present many levels of predictability that can be leveraged and represented in digital twins, in particular related to processes that are being increasingly understood via statistical models and theory. Both authors talk about the fact that cities are very complex and are associated with long-term dynamics (in contrast to the typical short-term dynamics of many applications), as well as about the different computational challenges that come with building digital twins of cities, such as the high computational complexity and the required multiscale modeling support.

In the area of Earth systems science, Peter Bauer, Torsten Hoefler, and colleagues discuss in a Comment — similar to the arguments put forth by Batty for city planning — that flexible human-in-the-loop interaction is essential for understanding and making efficient use of the data provided by digital twins of Earth. The authors argue that, in order to achieve such needed interaction, large pre-trained data-driven models are expected to be key to facilitate access to information hidden in complex data and to implement the human interface portion of digital twins (for instance, by using intelligent conversational agents). The authors also indicate the need for agreed upon standards for data quality and model quality, echoing some of the issues discussed by Tao et al. for industrial applications.

Evidently, there are many commonalities across these domains when it comes to current obstacles and opportunities for digital twins — but at the same time, there is also variability in how digital twins are perceived and used depending upon the specific challenges faced by each research community. Accordingly, the National Academies of Sciences, Engineering, and Medicine (NASEM) recently published a report 5 that identified gaps in the research that underlies the digital twin technology across multiple areas of science. The report — recapitulated by Karen Willcox and Brittany Segundo in a Comment — proposes a cross-domain definition for digital twins based on a previously published definition 6 and highlights many issues and gaps also echoed by some of the manuscripts in the Focus, such as the critical role of verification, validation, and uncertainty quantification; the notion that a digital twin should be ‘fit for purpose’, and not necessarily an exact replica of the corresponding physical twin; the need for protecting individual privacy when relying on identifiable, proprietary, and sensitive data; the importance of the human in the loop; and the need of sophisticated and scalable methods for enabling an efficient bidirectional flow of data between the virtual and physical assets.

It goes without saying that the Focus covers only a fraction of the potential applications for digital twins: many other areas have the potential to benefit from this technology, including, but not limited to, civil engineering 7 , 8 , chemical and materials synthesis 9 , sustainability 10 , 11 , and agriculture 12 . Without a doubt, there are many challenges to overcome and many research gaps to be addressed before we can bring the promise of digital twins to fruition. As stated in the NASEM report 5 , realizing the potential of digital twins will require an integrated research agenda, as well as an interdisciplinary workforce and collaborations between domains. We hope that this Focus will facilitate such collaborations and further discussion within the broad computational science community.

Ferguson, S. Apollo 13: The First Digital Twin (Siemens, 2020); https://blogs.sw.siemens.com/simcenter/apollo-13-the-first-digital-twin/

National Research Council. NASA Space Technology Roadmaps and Priorities: Restoring NASA’s Technological Edge and Paving the Way for a New Era in Space (The National Academies Press, 2012); https://doi.org/10.17226/13354

Grieves, M. Physical twins, digital twins, and the Apollo myth. LinkedIn (8 September 2022); https://go.nature.com/3ILB4Im

Grieves, M. Product Lifecycle Management: Driving the Next Generation of Lean Thinking (McGraw Hill Professional, 2005).

National Academies of Sciences, Engineering, and Medicine. Foundational Research Gaps and Future Directions for Digital Twins (The National Academies Press, 2023); https://doi.org/10.17226/26894

Digital Twin: Definition & Value — An AIAA and AIA Position Paper (AIAA, 2021); https://www.aia-aerospace.org/publications/digital-twin-definition-value-an-aiaa-and-aia-position-paper/

Feng, J., Ma, L., Broyd, T. & Chen, K. Autom. Constr. 130 , 103838 (2021).

Article   Google Scholar  

Torzoni, M., Tezzele, M., Mariani, S., Manzoni, A. & Willcox, K. E. Comput. Methods Appl. Mech. Eng. 418 , 116584 (2024).

David, N., Sun, W. & Coley, C. W. Nat. Comput. Sci. 3 , 362–364 (2023).

Tzachor, A. et al. Nat. Sustain. 5 , 822–829 (2022).

Mehrabi, Z. Nature 615 , 189 (2023).

Shrivastava, C. et al. Nat. Food 3 , 413–427 (2022).

Download references

Rights and permissions

Reprints and permissions

About this article

Cite this article.

The increasing potential and challenges of digital twins. Nat Comput Sci 4 , 145–146 (2024). https://doi.org/10.1038/s43588-024-00617-4

Download citation

Published : 26 March 2024

Issue Date : March 2024

DOI : https://doi.org/10.1038/s43588-024-00617-4

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Synergizing Digital Twins and Metaverse for Consumer Health: A Case Study Approach

Ieee account.

• Change Username/Password
• Update Address

Purchase Details

• Payment Options
• Order History
• View Purchased Documents

Profile Information

• Communications Preferences
• Profession and Education
• Technical Interests
• US & Canada: +1 800 678 4333
• Worldwide: +1 732 981 0060
• Contact & Support
• About IEEE Xplore
• Accessibility
• Terms of Use
• Nondiscrimination Policy
• Privacy & Opting Out of Cookies

A not-for-profit organization, IEEE is the world's largest technical professional organization dedicated to advancing technology for the benefit of humanity. © Copyright 2024 IEEE - All rights reserved. Use of this web site signifies your agreement to the terms and conditions.