Mosaic Digital Twins: A Variability-Centric Approach to Engineering Adaptive Cyber-Physical Systems

Digital twin (DT) engineering often relies on ad-hoc practices, hindering the scalability, consistency, and adaptability of systems developed.Thus, modeling dynamic systems with rich variability options, such as the DT of the road system, where appropriate thermomechanical simulations must be applied and dynamically updated for every vehicle individually, remains challenging.This research introduces mosaic digital twins (MDTs) -a novel architecture that leverages variability-centric design principles for engineering digital twin systems.We present the Waffle workbench -a framework for constructing MDTs that features a specialized domain-specific language (DSL) for advanced variability management.The latter enables the consistent composition of unique DT instances from a rich 'mosaic' of features.A key innovation is the MDT's proactive self-adaptation mechanism, where behavioral model simulation results trigger structural reconfigurations managed by Waffle.This doctoral work details the MDT concept, the Waffle workbench architecture, and its novel features.It outlines an empirical evaluation plan based on a complex highway digital twin case study, aiming to demonstrate improvements in DT consistency, scalability, and adaptive capabilities.