Application area: Industrial Automation
Researcher in charge: Marina Rantanen Modéer
Supervisor: Prof. Dr.-Ing. Sebastian Engell
Host: Technische Universität Dortmund, Germany
Secondments: INCHRON GmbH, Germany, Eindhoven University of Technology, the Netherlands and KTH Royal Institute of Technology, Sweden
Working title of thesis “Reliable design of cyber-physical systems based on partial and heterogeneous models”
Currently there is a lack of rigorous and comprehensive design methodologies for complex CPS. Major efforts are spent on testing and on the often costly correction of design flaws which could be reduced if an integrated model-based design would be used. On the other hand, the required effort for building all-encompassing models of complex systems often is prohibitive, so the modelling efforts should be focused on those elements and aspects of a system which are crucial for the correct function and the performance of the overall system.
For each design task, a conscious decision has to be made whether it should be performed in a model-based fashion and how accurate a model must be to answer the questions that arise in the design process. The effort for modeling must be focused on the critical aspects for the overall system. As a result, on the one hand, the models that are used in the design process are of different scope and level of detail and are formulated in different formalisms. On the other hand, the models and the design decisions are connected and must be managed in a comprehensive environment to avoid inconsistencies and to be able to track the effect of local decisions on the global system performance.
This research is expected to output a framework that connects different models, e.g. continuous time models and discrete time event models, for the design and analysis of comprehensive CPS. The researcher will investigate the relationships between the models, documenting model assumptions, ensuring consistency of models and model parameters, propagation of specifications and performance parameters from the global to the local level and back, and handling of system-wide effects of design decisions. The project will be carried out for controlled and automated systems with focus on performance related criteria. Some relevant research questions include:
- How can heterogeneous models that represent different properties of different elements of complex systems be used in the overall process of design and validation?
- How can the dependencies of the models and the dependencies between the results of a partial model-based analysis, e.g. a simulation or a verification of certain properties of a subsystem be managed so that consistency is maintained? Can this be formulated in a contract-based or assume-guarantee fashion?
- How can model abstraction be formulated in the case of continuous-time dynamic systems?
- How can the integration of different hierarchy levels in automated systems (both from the process and manufacturing industries) be performed?
- How can uncertainties of the models be represented and handled in the design process?
- How can the overall performance of a complex system be predicted on the basis of partial models?