Towards an Approach for Modeling and Architecting of Self-Adaptive Cyber-Physical Systems

Cyber-physical systems (CPS) emerge as a paradigm in which smart things actively collaborate among them, providing value-added services. The interconnection of physical things and their integration into software systems caused the CPS to be more and more complex. CPS is required to carry out a mission with minimum supervision. In fact, CPSs plan to be flexible systems that should autonomously, or with minimal manual intervention, decide how they adapt their internal behavior in response to dynamic environmental events. In this paper we discuss the CPS paradigm from the perspective of self-adaptive systems. Indeed, this will allow CPSs to modify their behavior, at runtime, in response to environment changes. We propose a multi-layer architecture for CPSs. The latter contains the main software modules that convert CPS to be self-aware, self-learning and self-reconfiguring. A case study about fault prediction in industrial CPS helps to evaluate the proposed architecture.