A REVIEW AND COMPARATIVE ANALYSIS OF AUTOMATED PROCESS CONTROL METHODS IN MODERN MANUFACTURING WITHIN THE CONTEXT OF INDUSTRY 5.0

Abstract

Relevance, topic and main objective. The development of automated process control systems (APCS) is a critical factor for ensuring stability, efficiency, resilience, and human-centric operation in modern manufacturing, particularly within the emerging framework of Industry 5.0. This article provides a structured review and comparative analysis of classical and intelligent APCS methods based on extended criteria such as control accuracy, adaptivity, computational efficiency, sustainability impact, resilience, integrability, explainability, and human-in-the-loop compatibility. The main objective is to evaluate and classify the principal control approaches and highlight their evolution toward hybrid architectures integrating machine learning, digital twins, and advanced operator-interaction mechanisms. Methods. The study applies a multi-criteria analytical framework informed by theoretical research, industrial reports, and documented implementations in manufacturing systems. The analysis incorporates modernized controller models, including an intelligent PID loop, a multi-objective MPC scheme supported by digital-twin-based prediction, and a neural-network-based architecture enhanced with explainable AI and resilience management. Results. The findings show that none of the examined APCS methods is universally optimal. PID remains effective for stable, well-characterized processes; MPC excels in multivariable, constraint-dominated environments; fuzzy and adaptive systems offer flexibility for uncertain conditions; neural networks demonstrate strong nonlinear modeling and fault tolerance but require substantial computational and data resources. Conclusions. Each method exhibits context-dependent strengths, and the most promising direction for APCS development lies in hybrid solutions that integrate classical techniques with intelligent, interpretable, and digitally interconnected components. Future research should focus on unifying enhanced PID, MPC, and neural-network-based controllers into a single hybrid architecture suitable for adaptive, transparent, and resource-efficient control in Industry 5.0 environments.