Towards a Model-Integrated Runtime Monitoring Infrastructure for Cyber-Physical Systems

Abstract

Runtime monitoring is essential for ensuring the safe operation and enabling self-adaptive behavior of Cyber-Physical Systems (CPS). It requires the creation of system monitors, instrumentation for data collection, and the definition of constraints. All of these aspects need to evolve to accommodate changes in the system. However, most existing approaches lack support for the automated generation and setup of monitors and constraints for diverse technologies and do not provide adequate support for evolving the monitoring infrastructure. Without this support, constraints and monitors can become stale and become less effective in long-running, rapidly changing CPS. In this "new and emerging results" paper we propose a novel framework for model-integrated runtime monitoring. We combine model-driven techniques and runtime monitoring to automatically generate large parts of the monitoring framework and to reduce the maintenance effort necessary when parts of the monitored system change. We build a prototype and evaluate our approach against a system for controlling the flights of unmanned aerial vehicles.

Adaptive hybrid reasoning for agent-based digital twins of distributed multi-robot systems

“…For this reason, we plan to incorporate model-driven engineering concepts and techniques to add more services such as monitoring [59], simplifying, and expediting the process of building and updating the design models [38] of DTs. This would provide a layer of abstraction that facilitates the creation of multipurpose models that can be modified and used for other DT implementations that target different systems and domains.…”

Towards developing a cyber-physical warehouse system for automated order-picking for online shopping

“…"Cyber-physical systems" are defined as "the integration of physical processes and computation." This gives the production process a whole new level of control, monitoring, and efficiency [25].…”