Model-based Documentation of Context Uncertainty for Cyber-Physical Systems
Torsten Bandyszak, Marian Daun, Bastian Tenbergen, Thorsten Weyer · 2018
Cyber-physical systems are increasingly developed for highly-automated purposes, such as industrial production plants. Not only the level of autonomy of these cyber-physical production systems, but also their potential lifespan, and the expectation that the system adapts throughout its life introduces challenges for the development process. To develop such systems in a structured manner, consideration of their context is essential. Suitable abstractions and perspectives must be employed to highlight different aspects of the context, such as structural or functional concerns. To ensure longevity and adaptability, the development process must account for changes in the context at runtime. This introduces a great deal of uncertainty. Current uncertainty modeling approaches cannot be applied right away, as they do not explicitly consider the multiple aspects of context analysis, which need to be documented in different artifacts. To aid the engineering of collaborative cyber-physical production systems, this paper presents a model-based approach to document context uncertainty. We propose “Orthogonal Uncertainty Models”, which closely integrate with context models from different perspectives. The approach has been evaluated using a case example from the industry automation domain, which is also presented herein.