A Framework for Model-Driven Digital Twin Engineering

Digital twins (DTs) are an emerging concept to handle the complexity of modern cyber-physical systems. They are virtual representations of physical objects (referred to as physical twins or PTs) that allow interaction with these objects. Experimentable DTs extend this notion through simulations of the physical object that allow experimentation and exploiting what-if scenarios. Based on these DTs, engineers can build innovative applications, from predictive maintenance to self-adaptation.

Problem: Connecting DT services, together with DTs and simulations, into so-called DT systems, and managing the evolution of these DT systems, currently requires a lot of effort from engineers. More precisely, this high effort stems from (i) redundant effort when specifying DT models in addition to engineering models, (ii) redundant effort when connecting DT services to both PT and simulations, and (iii) lack of dedicated support for managing variability when evolving DT systems in time and space.