John Fitzgerald’s research while affiliated with Newcastle University and other places

Pioneering work in model-oriented specification showed that properties of software could be confirmed or refuted from the early stages of development and be verifiably preserved through design to implementation. As verification technology has evolved, research and innovation have had to balance the insight generated by hand-crafted proof with the benefits of automation. In this paper, we describe QuickCheck, an ‘as-you-specify’ checker for proof obligations in the Vienna Development Method (VDM) Specification Language. Using a classical example of model-oriented specification in VDM, we demonstrate the value of such automated checking of basic proof obligations as a precursor to full machine-assisted proof using the Isabelle theorem prover.

We outline the technical processes defined in the systems engineering standard ISO/IEC/IEEE 15288, considering how each of them plays out when engineering a system enabled by a Digital Twin (DT).We consider how these processes, and the teams working on them, might be tailored to the challenges of engineering DTs, including the roles and competencies needed. A DT-enabled tempeh incubator system is used as a running example.

We report a virtual “roundtable” discussion on Digital Twins (DTs) involving experts from four engineering sectors covering consumer electronics, the built environment, aerospace and infrastructure. Topics covered include the vision, benefits and risks for DT technology in each sector, approaches to DT engineering and the role of dependability. Themes identified by our contributors include the need for DT engineering life cycles that deal with the creation of DTs for both new and existing products, competencies in ensuring that engineering disciplines work together, and the future need for DT integration.

This chapter introduces concepts that form the foundations for systematically considering the engineering of digital twins for cyber-physical systems. We introduce an example based on an incubator for growing tempeh, which serves as a running study for the remainder of this book. Based on this example,we discuss basic system concepts and core features of the models that are at the heart of any digital twin.We then consider the structure of a digital twin as an assembly of services, built upwards from models and data to the delivery of advanced analytics, visualisation and “what-if” analysis. This involves the management of these services, their communication protocols, data storage, model management, and other crucial aspects inherent to distributed software systems implemented within a service-oriented architecture. The rest of the book gives details about how to realise the different services and their management.

Although digital twins have been around for decades, there are many areas of research that need to be pursued in order to enable trustworthy predictions. The existing challenges and the future directions to take can be organised in many different ways. In this chapter, the foundational issues that we still think require clarification comes first. This is followed by research challenges in the different platforms needed to be able to produce digital twins in a cost-efficient manner. Afterwards the research changes for enabling autonomy of digital twins (uncertainty in many areas may hinder an accurate estimation of the state). Finally, research in composition of digital twins is considered and this really goes into system of systems considerations where an eco-system of stakeholders needs to agree about the means of composition.

In this chapter we introduce the concept of Digital Twin (DTs) for cyberphysical systems. We describe aspects of the technical engineering of DT-enabled systems, including their model-centric character, the need to maintain fidelity to the physical world, and the need to add value to the system being twinned. We examine the conceptual structure of a DT based on the assets it maintains and the services it offers, and consider the ways in which DTs may emerge through development and federation.

This paper describes challenges in the composition of cyber-physical systems enabled by digital twins. Digital twins link virtual models with the real systems that they represent, adding value to these models throughout the system’s life. As we come to rely on federations of systems, there is good reason to be interested in the composition of both models and twins. A practical example demonstrates the composition of twins of two robots, building on a “Digital Twin as a Service” platform. In considering how this field may develop, we highlight the influence of business and commercial factors, such as the need to deal with the diverse interests of their stakeholders, that are expected to have a significant technical impact.

The model-based engineering of dependable robotic systems brings challenges that include the need to work effectively with a range of models, owners and operators, and addressing the extent to which the systems evolve over time. This paper explores the linking of co-simulation, which allows for the exploration of models composed of multiple diverse simulation units, with the concept of a digital twin, which provides a framework for coupling virtual models with the real, evolving systems that they are intended to represent. We describe a systematic and generic approach to fault injection in multi-formalism models that use the Functional Mockup Interface standard, alongside the embedding of such multi-models in digital twins. We illustrate the potential and challenges of such an approach using a small “desktop” version of a commercial semi-autonomous agricultural robot. Inspired by the work of Jan Peleska on test automation, we identify directions for future research and innovation in this setting.