Computational Models Of Human–Machine Integration In Automated Systems

Abstract

Modern automated manufacturing systems require high flexibility to accommodate rapid product customization and dynamic production demands. However, conventional Programmable Logic Controller (PLC)-based automation systems rely on static control logic and manual programming, resulting in long reconfiguration times, high operational complexity, increased human intervention, and susceptibility to errors. The primary objective of this research is to develop an intelligent, adaptive, and interpretable control framework that enables efficient real-time reconfiguration of manufacturing systems while minimizing manual effort and improving operational safety. To achieve this, the proposed Intelligent Adaptive Control Framework (IACF) integrates Knowledge Graph Modeling (KGM) for structured representation of machines, tasks, and dependencies, and a Double Deep Q-Network (DDQN) for optimal decision-making under dynamic conditions. A Human-in-the-Loop Explainable Artificial Intelligence (XAI) module provides transparent and interpretable recommendations, ensuring operator trust and validation. Additionally, a Model-Based Design (MBD) Programmable Logic Controller (PLC) auto-code generation engine automatically generates executable PLC code, while a digital twin simulation layer validates system performance and safety before deployment. Z-score normalization is applied to input features to improve learning stability. Experimental evaluation implemented in Python demonstrates that the proposed IACF achieves strong predictive performance with an Mean Absolute Error (MAE) attained lower value of 0.097. The framework provides precise forecasts with few errors, enhancing flexibility, effectiveness, and dependability in the reconfiguration of intelligent manufacturing systems, making it appropriate for use in practical applications.