Empowering Adaptive Manufacturing with Interactive Diagnostics: A Multi-Agent Approach

This paper presents a novel approach towards proactive manufacturing control that integrates automated diagnostics with human interaction, resulting in a flexible adaptation of machine capabilities which helps to avoid damage in case of abnormalities. The model-based interpretation process supports predictive diagnostics using abductive reasoning, relying on plausibility thresholds and human intervention to resolve the resulting ambiguity between competing solutions. This enables the system to detect and avoid potential failure states before they actually occur. The proposed architecture additionally integrates intelligent products as mobile sensors, improving robustness and dependability of the production system.