Intention Abstraction Layer Prevents Industrial System Conflicts

Modern industrial environments increasingly deploy multiple autonomous subsystems, such as schedulers and energy managers, which operate independently but share physical resources. A critical issue arises because high-level human intentions are often translated into low-level control logic and then discarded. This means no running component can verify if its actions still align with the original intent, leading to goal conflicts that only become apparent after causing operational failures. To address this, the Intention Abstraction Layer (IAL) is proposed as a domain-agnostic middleware. The IAL treats intentions as first-class, persistent, and explainable runtime objects. It utilizes a large language model, grounded in a formal OWL ontology, to parse natural language goals into structured intentions. A key feature of the IAL is its consistency monitor, which detects potential conflicts at the registration stage, *before* any execution begins. A transparency module then explains these conflicts in natural language. A proof-of-concept demonstrated the IAL successfully flagging and explaining a conflict between production and energy intentions registered by two autonomous agents, effectively shifting behavioral assurance from reactive failure analysis to proactive, intention-level checking.