System-level feature-based modeling of cyber-physical systems

System-level feature-based modeling of cyber-physical systems: A theoretical framework and methodological fundamentals

Pourtalebi Hendehkhaleh, S. (TU Delft Cyber-Physical Systems)

Horvath, I. (promotor)

Delft University of Technology

2017-04-26

Cyber-physical systems are complex trans-disciplinary systems. Designing this kind of systems requires cooperation of several groups of experts with various backgrounds such as mechatronics and robotics, software engineering, data management, knowledge engineering, system on a chip, embedded systems, humans and systems interaction, and social and cognitive engineering. Design of a new CPS is typically initiated by system designers, who focus on system-level architecting and functional design (pre-embodiment design), and completed by embodiment/detail designers (who focus on components and their interoperation). Although there are numerous software tools to assist designers in detail design, it is hard to find dedicated tools that can sufficiently support CPS designers in pre-embodiment design phase. The tools available for this purpose are mostly based on logical, analytical, and mathematical modeling. They usually apply various levels of abstraction and simplification in modeling. As a result, many important chunks of information about attributes and parameterization of components are ignored or lost in system-level conceptualization. To overcome this problem, we introduced the framework and computational methodology of a system-level feature-based modeling, which can be used as a basis for developing next-generation CPSs modeling tools.
As an general underpinning, we have developed the mereo-operandi theory (MOT), which combines several classic theories in a uniform body and captures information about architectural and operational relations in a system in an integral manner. As a complementary to MOT, the theory of system manifestation features (SMFs) has been proposed that introduced the methodological fundamentals. SMFs have been defined as multi-stage, knowledge-intensive, compound modeling entities. The developed knowledge frames have been implemented as relational tables of the modeling warehouses. The computational procedures required for modeling of CPSs have also been elaborated. For validation, the developed modeling framework was benchmarked against multiple available modeling tools. It has been concluded that SMFs-based modeling offers many new opportunities for compositional modeling of CPSs. Some of the novelties are: (i) imposing strictly physical view, (ii) concurrent modeling of architecture and operation aspects, (iii) uniform information structure for all kinds of components, and (iv) benefiting from active ontologies.

Cyber-Physical Systems
System manifestation features (SMFs)
system-level features
system-level modeling
Pre-embodiment design
Mass customization
Feature technology

https://doi.org/10.4233/uuid:a75df963-aede-44a7-9216-274cdc7c4278

978-94-028-0621-2

Institutional Repository

doctoral thesis

© 2017 S. Pourtalebi Hendehkhaleh