Back to search

University of Western Macedonia

Ιndustrial internet of things optimization techniques using machine learning

Abstract

dc:description

The rapid evolution of Industry 4.0 has redefined industrial environments by integrating IndustrialInternet of Things (IIoT) systems, enabling unprecedented levels of real-time monitoring, process optimization, and automation. However, this transformation also brings significant challenges, including defect detection, efficient resource allocation, cybersecurity risks, and usability concerns. This thesis addresses these critical issues by developing and implementing advanced Artificial Intelligence (AI)techniques, optimization strategies, and explainability frameworks to enhance performance, security, and reliability within IIoT ecosystems. The thesis begins in Chapter 1 with a comprehensive introduction that outlines the motivation and objectives of the research. This chapter highlights the pressing need for AI-driven solutions to address IIoT challenges while introducing the methodological approach adopted throughout the work. The research contributions, structure, and scope of the thesis are also clearly defined. In Chapter 2, the focus shifts to the foundational role of AI in IIoT ecosystems. The chapter examinesthe application of Machine Learning (ML) and Deep Learning (DL) techniques in industrial processes such as Smart Manufacturing (SM) and Smart Grid (SG) management. It also discusses the unique characteristics of IIoT data, such as its high volume, heterogeneity, and real-time processing requirements, emphasizing the need for tailored AI approaches to extract actionable insights. Chapter 3 presents an extensive analysis of AI models and optimization strategies for industrial environments. This chapter explores techniques like Convolutional Neural Networks (CNNs) for defect detection, hyperparameter tuning, and transfer learning for model refinement, and optimization frameworks such as Mixed-Integer Linear Programming (MILP) to streamline resource allocation. The combination of these methods ensures scalable, efficient, and accurate solutions that align with the computational constraints of IIoT systems. To bridge the gap between AI systems and human operators, Chapter 4 introduces a novel cognitive IIoT platform that integrates Explainable AI (XAI) and Human-in-the-Loop (HITL) methodologies. By enhancing the transparency and interpretability of AI-driven decisions, this platform empowers users to collaborate with AI systems effectively. Real-world applications in defect detection demonstrate how HITL mechanisms improve accuracy while fostering trust and usability for system operators. Recognizing the growing cybersecurity threats within interconnected IIoT systems, Chapter 5 addresses the need for robust security mechanisms. This chapter introduces AI-based Intrusion Detection Systems (IDS) and Trust Management Frameworks that leverage advanced models for detecting cyberattacks, mitigating vulnerabilities, and managing reputational risks. Special emphasis is placedon securing industrial protocols such as Modbus/TCP and ensuring resilience in critical infrastructure like Smart Grids.In Chapter 6, the proposed AI-driven models and frameworks are rigorously evaluated across multiple industrial use cases, including defect detection, anomaly identification, and cybersecurity applications. This chapter outlines the evaluation environments, datasets, and performance metrics used to validatethe effectiveness, scalability, and robustness of the solutions. Comparative analyses highlight the superior performance of the developed techniques over traditional approaches. Finally, Chapter 7 concludes the thesis by summarizing the key findings and contributions of the research. This chapter also outlines promising directions for future work, such as Reinforcement Learning for adaptive resource optimization, and Graph Neural Networks for network recovery mechanisms.

Degree

thesis:*
Grantor dc:publisher
University of Western Macedonia
Year dc:date
2025

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Kotsiopoulos, Athanasios
  • Κωτσιόπουλος, Αθανάσιος

Subjects

dc:subject × 22

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/60432

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Kotsiopoulos, Athanasios; Κωτσιόπουλος, Αθανάσιος. Ιndustrial internet of things optimization techniques using machine learning. University of Western Macedonia, 2025. http://hdl.handle.net/10442/hedi/60432