Research Topics

Our research in the Oakes lab focuses on enabling domain experts to efficiently capture and utilise their knowledge through an intelligent model-driven approach, to support knowledge engineering for complex cyber-physical systems. The goal is to minimise the cognitive and time effort for constructing, verifying, and validating these systems, while still maximising the insights gained during the systems engineering process.

Interested in joining the Oakes Lab?

Digital twins — faithful, executable models of physical systems — need rigorous engineering approaches to be trustworthy and actionable. Our work spans the full DT lifecycle: foundational modelling concepts and description frameworks, service-driven construction methods that organise DT components around well-defined interfaces, and knowledge graph architectures for accessing historical and streaming data. A recurring concern is making DT behaviour transparent: our reporting frameworks and real-time monitoring tools help operators understand what a DT is doing and why. This work is validated through case studies including cooperative robotics and industrial fermentation processes.

Effective systems engineering of complex cyber-physical systems requires modelling techniques that scale to the diversity of the engineers who use them. Our work develops MBSE and MDE frameworks that bridge the gap between domain expertise and formal models — including the openCAESAR framework for balancing agility and rigor, tools for intelligently generating domain-specific modelling languages, and frameworks for supporting functional safety processes. A survey of 76 MBSE practitioners grounds this work in real organizational challenges, informing where tool and methodology improvements have the greatest impact.

Machine learning offers targeted capabilities for specific engineering challenges — but applying it effectively requires understanding the structure of the problem. Our work investigates ML and reinforcement learning for fault injection and hazard analysis, where efficient exploration of fault parameter space is critical for safety-critical systems. We also apply ML to mining developer rationale from commit messages, and to automatically repairing model transformations with semantic errors. Alongside these applications, we study how domain experts build ML workflows, identifying the key challenges they face in translating domain knowledge into executable pipelines.

Engineering complex systems depends on capturing expert knowledge in a form that tools can reason over. Our work uses ontologies, knowledge graphs, and semantic modelling to structure domain knowledge across multiple contexts: ontological frameworks for MBSE, knowledge graph architectures for querying DT data, and semantic approaches to mining developer rationale. A key goal is interoperability — enabling knowledge captured in one tool or project to be reused and reasoned about across the broader engineering ecosystem.