Special Session 4: Trustworthy Agentic Intelligence for PHM-Oriented Industrial Reliability and Safety
Please submit your manuscript via https://www.easychair.org/conferences/?conf=icsrs2026 and select Special Session 4
Chairs
Brief Introduction
Agentic intelligence is reshaping Prognostics and Health Management (PHM) and industrial reliability by enabling systems that can interpret heterogeneous evidence, coordinate analytical tools, reason over domain knowledge, and support maintenance decisions. This special session invites original research, reviews, and applications on trustworthy AI agents for PHM-oriented reliability and safety in complex engineering and manufacturing systems. Topics include multi-agent architectures, large/foundation models, multimodal evidence fusion, digital twins, RAG and knowledge-grounded reasoning, uncertainty-aware prognosis, human-in-the-loop validation, auditable workflows, and safety assurance. The session emphasizes reliable, explainable, traceable, and deployable agentic intelligence for health assessment, fault diagnosis, RUL prediction, and maintenance optimization.
Sub-topics
• Trustworthy agentic AI for PHM and industrial reliability
• Multi-agent architectures and coordination mechanisms for diagnosis, prognosis, and maintenance
• LLM- and foundation-model-based PHM assistants
• Multimodal evidence fusion using sensor, image, text, log, and operational data
• Knowledge-grounded reasoning, RAG, and GraphRAG for PHM decision support
• Digital twin-enabled agentic PHM for cyber-physical and manufacturing systems
• Uncertainty-aware, risk-informed, and explainable RUL prediction
• Component orchestration for fault diagnosis, prediction, and maintenance optimization
• Human-in-the-loop validation, deterministic replay, and auditable PHM workflows
• Safety assurance, benchmarking, governance, and reliability evaluation of PHM agents