ºÚÁϸ£ÀûÍø

Modern engineering systems are becoming increasingly complex, from autonomous robots to smart factories. To design, monitor, and optimize them, engineers often create digital twins, interactive virtual replicas that reflect both the structure and behavior of the real systems. These digital models allow engineers to test scenarios, simulate changes, and predict maintenance needs used in robotics, manufacturing, and other industries without interrupting physical operations.

Building digital twins, however, presents major challenges. They typically involve multiple digital models, each developed by experts from different fields using specialized tools. Ensuring these models remain consistent and accurately represent the real system is essential for reliable operation. Even small mismatches can lead to errors, unreliable predictions, or interruptions in system performance.

Investigating the challenge

To understand these challenges, Muctadir conducted interviews with 19 digital twin experts. He found that inconsistencies are common, while existing tools and methods are limited in capability and scope. Although consistency management has long been studied in traditional software engineering, it is much less explored for digital twins, which are more heterogeneous and dynamic.

This insight motivated the development of a new approach specifically tailored to digital twin models, which are often built with domain-specific tools and evolve continuously as systems are updated.

 

Beyond model-level consistency

Digital twins are dynamic and data-driven, so new challenges arise beyond static model consistency. For example, the virtual system may behave differently than its physical counterpart.

To explore this, Muctadir applied his framework to a digital twin of an autonomous TURTLE soccer robot, similar to those employed by Tech United in RoboCup and other tournaments. Using advanced computer techniques, he detected mismatches between expected and actual behavior and monitored alignment with the real system. This shows how digital twins can actively support their own reliability.

Towards reliable and trustworthy digital twins

Muctadir’s work shows that as digital twins become more central to industry, their reliability depends on careful consistency management. His framework provides a foundation for future tools that can handle large, complex systems, improve performance, and simplify implementation.

By detecting and addressing inconsistencies early, engineers can create digital twins that are more robust, adaptable, and trustworthy. As these systems spread across fields such as manufacturing, robotics, energy, and healthcare, Muctadir’s research offers practical methods to ensure they continue to evolve safely and effectively.