From Fifth Generation Dreams to Intelligent Industrial SCADA: Integrating LLMs, Expert Systems and Knowledge Graphs

This paper revisits the Japanese Fifth Generation Computer Systems project, reframing its historical lessons to inspire modern industrial SCADA architectures. It integrates large language models (LLMs), expert systems, and knowledge graph techniques to enable natural language interfaces, real-time analytics, and adaptive control in production lines. Retrieval-augmented generation and fine-tuning strategies are examined to overcome LLM limitations — hallucinations, knowledge cutoffs and lack of personalization — within high-reliability expert systems. The proposed framework envisions scalable, domain-adaptive AI that merges human-centered interaction with advanced reasoning, offering a pathway toward more intelligent, responsive and context-aware industrial automation systems.

Authors:

Fabian Barreiro | ANT Automation LLC
 

Javier Barreiro | ANT Automation LLC
 

Session Chairs:

Brad Morgan | Nucor Business Technology
 

Chalapathy Devarakonda | Big River Steel