While there has been a recent growth of interest in the Digital Twin, a variety of definitions employed across industry and academia remain. There is a need to consolidate research such to maintain a common understanding of the topic and ensure future research efforts are to be based on solid foundations. Through a systematic literature review and a thematic analysis of 92 Digital Twin publications from the last ten years, this paper provides a characterisation of the Digital Twin, identification of gaps in knowledge, and required areas of future research. In characterising the Digital Twin, the state of the concept, key terminology, and associated processes are identified, discussed, and consolidated to produce 13 characteristics (Physical Entity/Twin; Virtual Entity/Twin; Physical Environment; Virtual Environment; State; Realisation; Metrology; Twinning; Twinning Rate; Physical-to-Virtual Connection/Twinning; Virtual-to-Physical Connection/Twinning; Physical Processes; and Virtual Processes) and a complete framework of the Digital Twin and its process of operation. Following this characterisation, seven knowledge gaps and topics for future research focus are identified: Perceived Benefits; Digital Twin across the Product Life-Cycle; Use-Cases; Technical Implementations; Levels of Fidelity; Data Ownership; and Integration between Virtual Entities; each of which are required to realise the Digital Twin.

https://research-information.bris.ac.uk/files/229965388/Characterising_The_Digital_Twin.pdf

Characterising the Digital Twin: A systematic literature review

The digital twin (DT) is a virtual representation of a physical object, which has been proposed as one of the key concepts for Industry 4.0. The DT provides a virtual representation of products along their lifecycle that enables the prediction and optimization of the behavior of a production system and its components. A methodology design using model-driven engineering (MDE) is proposed that strives toward being both flexible and generic. This approach is presented at two levels: first, a DT is modeled as a composition of basic components that provide basic functionalities, such as identification, storage, communication, security, data management, human–machine interface (HMI), and simulation; second, an aggregated DT is defined as a hierarchical composition of other DTs. A generic reference architecture based on these concepts and a concrete implementation methodology are proposed using AutomationML. This methodology follows an MDE approach that supports most of the DT features currently proposed in the literature. A case study has been developed, the proposed ideas are being evaluated with industrial case studies, and some of the preliminary results are described in this article. With the case study, it is possible to verify that the proposed methodology supports the creation and the deployment process of a DT.

A Methodology for Digital Twin Modeling and Deployment for Industry 4.0

Digital Twin is a new concept that consists of creating an up-to-date virtual asset in cyberspace which mimics the original physical asset in most of its aspects, ultimately to monitor, analyze, test, and optimize the physical asset. In this article, we investigate and discuss the use of the digital twin concept of the roads as a step toward realizing the dream of smart cities. To this end, we propose the deployment of a Digital Twin Box to the roads that is composed of a 360° camera and a set of IoT devices connected to a Single Onboard Computer. The Digital Twin Box creates a digital twin of the physical road asset by constantly sending real-time data to the edge/cloud, including the 360° live stream, GPS location, and measurements of the temperature and humidity. This data will be used for realtime monitoring and other purposes by displaying the live stream via head-mounted devices or using a 360° web-based player. Additionally, we perform an object detection process to extract all possible objects from the captured stream. For some specific objects (person and vehicle), an identification module and a tracking module are employed to identify the corresponding objects and keep track of all video frames where these objects appeared. The outcome of the latter step would be of utmost importance to many other services and domains such as national security. To show the viability of the proposed solution, we have implemented and conducted real-world experiments where we focus more on the detection and recognition processes. The achieved results show the effectiveness of the proposed solution in creating a digital twin of the roads, a step forward to enable self-driving vehicles as a crucial component of smart mobility, using the Digital Twin Box.

/pdf/roads-infrastructure-digital-twin-a-step-toward-smarter-2pmevn5ftj.pdf

Roads Infrastructure Digital Twin: A Step Toward Smarter Cities Realization

Industry 4.0 aims to support smarter and autonomous processes while improving agility, cost efficiency, and user experience. To fulfill its promises, properly processing the data of the industrial processes and infrastructures is required. Artificial intelligence (AI) appears as a strong candidate to handle all generated data, and to help in the automation and smartification process. This article overviews the digital twin as a true embodiment of a cyber-physical system (CPS) in Industry 4.0, showing the mission of AI in this concept. It presents the key enabling technologies of the digital twin such as edge, fog, and 5G, where the physical processes are integrated with the computing and network domains. The role of AI in each technology domain is identified by analyzing a set of AI agents at the application and infrastructure levels. Finally, movement prediction is selected and experimentally validated using real data generated by a digital twin for robotic arms with results showcasing its potential.

Toward Intelligent Cyber-Physical Systems: Digital Twin Meets Artificial Intelligence

The growing demand of industrial, automotive and service robots presents a challenge to the centralized Cloud Robotics model in terms of privacy, security, latency, bandwidth, and reliability. In this paper, we present a ‘Fog Robotics’ approach to deep robot learning that distributes compute, storage and networking resources between the Cloud and the Edge in a federated manner. Deep models are trained on non-private (public) synthetic images in the Cloud; the models are adapted to the private real images of the environment at the Edge within a trusted network and subsequently, deployed as a service for low-latency and secure inference/prediction for other robots in the network. We apply this approach to surface decluttering, where a mobile robot picks and sorts objects from a cluttered floor by learning a deep object recognition and a grasp planning model. Experiments suggest that Fog Robotics can improve performance by sim-to-real domain adaptation in comparison to exclusively using Cloud or Edge resources, while reducing the inference cycle time by $4\times$ to successfully declutter 86% of objects over 213 attempts.

A Fog Robotics Approach to Deep Robot Learning: Application to Object Recognition and Grasp Planning in Surface Decluttering

The concept of "Industry 4.0" considers smart factories as data-driven and knowledge enabled enterprise intelligence. In such kind of factory, manufacturing processes and final products are accompanied by virtual models -- Digital Twins. To support Digital Twins concept, a simulation model for each process or system should be implemented as independent computational service. The only way to implement an orchestration of a set of independent services and provide scalability for simulation is to use a cloud computing platform as a provider of the computing infrastructure. In this paper, we describe a Digital Twin-as-a-Service (DTaaS) model for simulation and prediction of industrial processes using Digital Twins.

Towards Digital Twins Cloud Platform: Microservices and Computational Workflows to Rule a Smart Factory

Comprehensive analysis of the angular and linear accelerations of moving elements (shafts, gears) allows an increase in the quality of the condition monitoring of mechanisms. However, existing tools and methods measure either linear or angular acceleration with postprocessing. This paper suggests a new construction design of an angular acceleration sensor for moving elements. The sensor is mounted on a moving element and, among other things, the data transfer and electric power supply are carried out wirelessly. In addition, the authors introduce a method for processing the received information which makes it possible to divide the measured acceleration into the angular and linear components. The design has been validated by the results of laboratory tests of an experimental model of the sensor. The study has shown that this method provides a definite separation of the measured acceleration into linear and angular components, even in noise. This research contributes an advance in the range of methods and tools for condition monitoring of mechanisms.

Wireless acceleration sensor of moving elements for condition monitoring of mechanisms

The neural network inverse model of a sensor with filtration of the sequentially recovered signal is considered. This model effectively reduces the dynamic measurement errors due to deep mathematical processing of measurement data. The result of the experimental data processing of a dynamic temperature measurement validates the efficiency of the proposed neural network approach to reduce dynamic measurement errors.

/pdf/neural-network-approach-to-reduce-dynamic-measurement-errors-4jtx32npzs.pdf

Neural network approach to reduce dynamic measurement errors

This paper provides a review of modern information technologies used in management of industrial enterprises and corporations, as well as of government authorities. Despite the success achieved in the field of the speed of operation of computers and supercomputers, and of various analyzing systems, the implementation of modern information technologies does not ensure the improvement of the quality of managing different organizational and corporate structures. The many-years’ experience in engineering and operation of various automated, information-computing, analytical, expert, and other systems has shown that the implementation of these systems was not always beneficial in terms of improving the organizational management quality, or making the activity of the structures, where these systems had been introduced, more efficient. In this context, today, it is of utmost importance to understand that the reasons for such slightly disappointing results of implementing various information systems do not lie in the technical capacities of the computing machines, but can rather be explained by the fact that the methods, models and technologies of management at all the stages of elaboration and making of decisions have remained essentially the same as in the last quarter of the past century. The lagging behind in the development of the methods and algorithms has lead to a situation when the capacities of the computing machines are not used to their maximum. Thus, an important task of the management theory and practice is to close the gap between the methods of decision making and the modern level of the computing equipment. The paper sets forth a task on making sure that the development of the computing machines is harmonized with the processes of improving and implementing the managerial algorithms. The causes of the abovementioned lagging behind are described, and the recommendations are given on how to eliminate the existing gap.

Aleksandr L. Shestakov

Papers

Towards Digital Twins Cloud Platform: Microservices and Computational Workflows to Rule a Smart Factory

Wireless acceleration sensor of moving elements for condition monitoring of mechanisms

Neural network approach to reduce dynamic measurement errors

Modern Information Technologies and the Need for Improving the Quality of Organizational and Corporate Structures’ Management