When Things Go Wrong: The SENSE Ontology for Explaining CPS Anomalies

Cyber-physical systems (CPS) integrate computational and physical components, using sensor data for advanced monitoring and analysis. However, as these systems grow in complexity, their internal processes and decision-making become less transparent, making it difficult for stakeholders to interpret, trust, or intervene in system behavior. Integrating heterogeneous data sources to support explainability, remains a challenge. Ontologies have proven effective for data integration and representation of diverse knowledge in a structured and machine-readable format. They can encode both general concepts and instance-specific information as well as provenance. However, existing ontologies lack support for dynamic temporal and causal relations present in CPS. To address this gap, we introduce Semantics-based Explanation of Cyber-physical Systems (SENSE), a domain-independent ontology designed to support user-centered explanations of anomalies in CPS. SENSE integrates five key knowledge areas: topology, observation, causality, user context, and explanation knowledge. Each of these areas is partially covered by previous ontologies. With SENSE, all of these aspects are integrated into a single, unified ontology, enabling causal reasoning over system anomalies at runtime. Following the Linked Open Terms (LOT) methodology, SENSE was developed and evaluated across three use cases in the smart grid and smart building domain, showcasing one proof of concept of a smart charging garage in this paper. The SENSE ontology and related development artifacts are publicly available under the CC-BY 4.0 License, providing an open resource for advancing semantic explainability in CPS. Future work will focus on expanding causal representations, scalability and domain adaptability.