Using Workspace to implement Digital Twins in the Mixed Reality Lab
2021
The growing need for decision support, situational awareness and predictive analytics systems in real world iterative scenarios has given rise to the research field of "Digital Twins"; the combining of cyberphysical systems, artificial intelligence, computational processing platforms and mixed-reality visualisation.Digital Twins are living, digital replicas of a physical object, process or system and are comprised of three elements; a physical system; a digital model of that system; and the interactions that flow between these two systems.In this paper we describe our approach to the ongoing task of building our Mixed Reality Lab (MRL) using our Scientific Workflow System (SWS) called Workspace and describe its characteristics that make it suitable for use in producing Digital Twins for both research and commercial applications.Digital Twins need to address many of the same issues that SWSs address, for example platform independence, handling large data sets, local and remote processing, algorithm integration, hardware integration, advanced visualisation, reproducibility and scalability.In general, we believe that the elements that make our SWS applicable to the creation of Digital Twins could also apply to other SWSs, subject to their capabilities in areas discussed in this paper.Two prototypical use cases supported in our MRL, Defect Detection in Manufacturing scenarios and Human Movement (a.k.a.Digital Human) are outlined.Although these two use cases work at different scales, we see how both can be modelled in the MRL using the infrastructure constructed on top of Workspace.Capability strengths of our SWS that are of particular importance to our Digital Twin use cases are discussed.These include: The use of both continuous execution (interactive) workflows with rich user interfaces as well as distributed, cross-platform, remote execution processing workflows; Advanced interactive visualisation capabilities ranging from 2D widgets and charts, OpenGL-based 3D rendering through to a real-time Virtual and Augmented Reality pipelines using third party rendering engines; a modular, extensible plugin based architecture that exposes capabilities from many permissive open source libraries combined with in-house developed capabilities; and a streamlined deployment and productisation process to support different distribution or commercialisation needs.The MRL is very much a work in progress, optimisations to the current systems are ongoing and work to add extra facilities and infrastructure to support more scenarios is progressing.