A Process for Identifying and Modeling Relevant System Context for the Reconfiguration of Automated Systems

TL;DR: In this article , the authors present a new method for identifying and modeling the context information that is relevant to trigger a system reconfiguration as well as automated context model creation in the form of ontologies and feature models.

Abstract: Context-aware systems are gaining increasing importance in industry as managing and using context information opens new opportunities for reconfiguration, digital twin concepts, performance monitoring, and so forth. Today, there is an immense variety of context-aware systems, from mobile applications to complex systems which require run time reconfiguration. The diversity of methods for acquiring the relevant data, as well as the specific use case requirements, complicate the management and usage of context information. Thus, the usage of context information for the reconfiguration management of systems demands adequate processes to identify and model system-relevant context. Consequently, in this research, we present a new method for identifying and modeling the context information that is relevant to trigger a system reconfiguration as well as automated context model creation in the form of ontologies and feature models. We suggest the ontology usage to disambiguate the data from different sources. The implementation of context-aware behavior is beyond the scope of this article and will be included in future research. We validate our ideas using cases study in the wind energy domain and unmanned automated vehicle domain. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —Dynamically changing environments, require systems that continuously monitor their environment and adapt when necessary to ensure safe and optimal operation. For the system to adapt, it is necessary to define which of the environmental variables are relevant and when a change should lead to a reconfiguration of the system. The method presented in this article helps practitioners to identify these environmental variables. A rapidly changing environment also means having to interact with different systems from different vendors. The relevant environmental variables could be provided by different systems or collected by different sensors that rely on different data models. To cope with this heterogeneous environment, it is useful to create a formalized and reusable context model, describing the relevant environmental variables. A formalized context model serves as a semantic layer to link heterogeneous data and enables operation in a changing environment. In this way, the result of the presented method is transformed automatically into an ontology.