Preface
Acknowledgments
Contents
Editor and Contributors
About the Editor
Contributors
Abbreviations
Part I Architectures and Knowledge Modelling for AI in Manufacturing
Reference Architecture for AI-Based Industry 5.0 Applications
1 Introduction
2 Relevant Work
3 Architecture for AI-Based Industry 5.0 Systems (STAR-RA)
3.1 High-Level Reference Model for AI-Based Industry 5.0 Systems
3.1.1 Overview
3.1.2 Cybersecurity Domain Functionalities
3.1.3 HRC Domain Functionalities
3.1.4 Safety Domain Functionalities
3.2 Logical Architecture for AI-Based Industry 5.0 Systems
3.2.1 Driving Principles
3.2.2 Logical Modules
4 Solution Blueprints for Industry 5.0 Applications
4.1 The Industry 5.0 Blueprints Concept
4.2 Technological Solutions Blueprints
4.3 Regulatory Compliance Blueprints
5 Conclusions
References
Designing a Marketplace to Exchange AI Models for Industry 5.0
1 Introduction
2 Functionalities and Proposed System Architecture of knowlEdge Marketplace
3 Implementation of knowlEdge Marketplace for Exchanging AI Models in Industry 5.0
3.1 knowlEdge AI Model Repository
3.1.1 Overview of Key Components
3.1.2 Ontology
3.2 NFT-Based Monetization Framework for AI Models
3.3 User Interfaces and Functionalities
4 Conclusions
References
Human-AI Interaction for Semantic Knowledge Enrichment of AI Model Output
1 Introduction
2 Related Works
3 Human Feedback into AI Model
3.1 Interface Abstraction
3.2 Model and Data Selection
3.3 Parameter Optimization
3.4 Configuration Adaptation
3.5 Domain Knowledge Enrichment
3.6 Domain Knowledge Repository
4 Interaction for Model Selection and Parameter Optimization
5 Conclusion and Future Works
References
Examining the Adoption of Knowledge Graphs in the Manufacturing Industry: A Comprehensive Review
1 Introduction
2 Antecedents and Motivation
3 Research Questions and Search Strategy
3.1 Research Questions
3.2 Dataset
3.3 Subject Area
3.4 Manufacturing Domain
3.5 Kinds of KGs
3.6 Different Approaches for KG Creation
4 Insights
4.1 Answers to the Research Questions
4.2 Additional Lessons Learned
4.3 Open Problems
5 Conclusion
References
Leveraging Semantic Representations via Knowledge Graph Embeddings
1 Introduction
2 Knowledge Graph Embeddings
3 Representation Learning
3.1 Representation Learning for Knowledge Graphs
3.1.1 Tensor Decomposition Models
3.1.2 Geometric Models
3.1.3 Deep Learning Models
4 Industrial Applications of Knowledge Graph Embeddings
5 The Navi Approach: Dynamic Knowledge Graph Embeddings via Local Embedding Reconstructions
6 Conclusions
References
Architecture of a Software Platform for Affordable Artificial Intelligence in Manufacturing
1 Introduction
2 Platforms in the AI Ecosystem
3 KITT4SME: A Platform Delivering AI to SMEs
3.1 High-Level Concept and Architecture
3.2 Functionalities and Component Description
4 KITT4SME to Bring AI to an Injection Molding Use Case
5 Conclusion
References
Multisided Business Model for Platform Offering AI Services
1 Introduction
2 Methodologies for MSPs Business Modeling
3 Application of PDT for the Design of AI Platform as a Service Business Model โ KITT4SME Case Study
3.1 Introduction to the KITT4SME Project
3.2 Needs Elicitation
3.3 KITT4SME Business Model
3.4 Business Model Design Canvas
3.5 Revenue Model for the KITT4SME Platform
4 Conclusions and Next Steps
References
Self-Reconfiguration for Smart Manufacturing Based on Artificial Intelligence: A Review and Case Study
1 Introduction
2 Reconfiguration in Manufacturing
2.1 Precursors of Reconfigurable Systems: Flexible Manufacturing Systems
2.2 Reconfigurable Manufacturing Systems
2.3 Evolution Towards Self-Reconfiguration
3 Current Approaches
3.1 Computer Simulation
3.2 Fuzzy Systems
3.3 Data-Driven Methods
3.4 Reinforcement Learning
4 Lighthouse Demonstrator: GAMHE 5.0 Pilot Line
4.1 Deep Learning-Based Visual Inspection
4.2 Automating the Machine Learning Workflow
4.2.1 Typical Machine Learning Workflow
4.2.2 Process Optimization
4.2.3 Application of AutoML to the Pilot Line
4.3 Fuzzy Logic-Based Reconfigurator
5 Conclusions
References
Part II Multi-agent Systems and AI-Based Digital Twins for Manufacturing Applications
Digital-Twin-Enabled Framework for Training and Deploying AI Agents for Production Scheduling
1 Introduction
2 Related Works
3 Multi-Agent System Framework
3.1 System Architecture
3.2 Paint Shop Scheduling Agents
3.2.1 Mathematical Optimization
3.2.2 Data-Driven Optimization
3.3 Deep Reinforcement Learning Scheduling Agent
3.4 Heuristic Optimization
4 Case Study
5 Conclusion
References
A Manufacturing Digital Twin Framework
1 Introduction
1.1 Definition, Usages, and Types of Digital Twins
1.2 Digital Twin in Manufacturing
2 knowlEdge Manufacturing Digital Twin Framework
2.1 Digital Twin Standardization Initiatives
2.2 knowlEdge Digital Twin Framework
2.3 knowlEdge Digital Twin Framework Alignment with Current Initiatives
3 knowlEdge Digital Twin for Process Improvement
4 Conclusions
References
Reinforcement Learning-Based Approaches in Manufacturing Environments
1 Introduction
2 Reinforcement Learning
2.1 Toward Reinforcement Learning in Manufacturing
3 Deep Reinforcement Learning in Virtual Manufacturing Environments
3.1 CNC Cutting Machine
3.2 Robotic Manipulation of Complex Materials
4 Conclusions
References
A Participatory Modelling Approach to Agents in Industry Using AAS
1 Introduction
2 Background
3 AAS Model for an Agent
3.1 General Model Structure
3.2 Generic Submodels
3.3 Specific Submodels
3.4 Usage
4 Methodology for Developing an AAS
4.1 Phases
4.2 Agent Modelling
5 AAS Model Repository
5.1 Functionality
5.2 Working Principle
6 Discussion
6.1 Maturity of the AAS
7 Conclusion
References
I4.0 Holonic Multi-agent Testbed Enabling Shared Production
1 Introduction
2 State of the Art
2.1 Control Architectures in the Manufacturing Domain
2.2 Cyber-Physical Production Systems
2.3 Agents and Holons
2.4 Agent Framework
3 Towards an Architecture for Modern Industrial Manufacturing Systems
3.1 Multi-agent System Manufacturing Architecture
3.2 Service Holon
3.3 Product Holon
4 Execution System: Resource Holon
4.1 Behaviors and Skills of a Holon
4.2 Demonstrator Use Case
5 Conclusion
References
A Multi-intelligent Agent Solution in the Automotive ComponentโManufacturing Industry
1 Introduction
1.1 Fersa's Pilot Plant
2 Experimental Development
2.1 Data Aggregation
2.2 Tooling Agent
2.3 Machine Agent
2.4 Prescheduling Agent
2.5 Holon
3 Conclusion
References
Integrating Knowledge into Conversational Agents for Worker Upskilling
1 Introduction
2 Related Work
3 Existing Conversational Agents
4 Skills, Competences, and Occupations
5 Proposed Solution
6 Expected Challenges, Benefits, and Impact
7 Conclusions
References
Advancing Networked Production Through Decentralised Technical Intelligence
1 Introduction
2 Decentralised Technical Intelligence
3 Implications for Networked Production Management
4 Building Blocks and Implementation Roadmap
5 DTI Deployment from Business and Organisational Perspective
6 Discussion
7 Conclusion
References
Part III Trusted, Explainable and Human-Centered AI Systems
Wearable Sensor-Based Human Activity Recognition for Worker Safety in Manufacturing Line
1 Introduction
2 Background
3 Wearable Sensor-Based Worker Activity Recognition in Manufacturing Line
3.1 Use Case at the SmartFactory Testbed
3.2 Data Acquisition
3.3 Worker Activity Recognition Results
4 Deep Learning Techniques for Human Activity Recognition Improvement
4.1 Adversarial Learning
4.2 Contrastive Learning
5 Conclusion
References
Object Detection for HumanโRobot Interaction and Worker Assistance Systems
1 Introduction: Why Object Detection in the Industrial Environment is Helpful?
2 Background
2.1 Dataset
2.2 Architectures
2.3 Application in Industrial Environment
2.4 Challenges
3 Scenarios
3.1 Object Detection for HumanโRobot Interaction in the STAR Project
3.2 Object Detection for Manual Assembly Assistance System in InCoRAP
3.3 Methodology: Context-Based Two-Step Object Detection
4 Ongoing Research
4.1 Hybrid Dataset
4.2 Continual Learning (CL)
5 Conclusion
References
Boosting AutoML and XAI in Manufacturing: AI Model Generation Framework
1 Introduction
2 AI Model Generation Framework
3 System Architecture
4 Use Cases
5 Core Components
5.1 Data Retrieval Module
5.2 Automatic Pre-processing Module
5.3 Cost Computation Module
5.4 Automatic Hyperparameter Tuning Module
5.5 Automatic Training, Inference, and Standardization
5.6 Explainability Generation Module
5.7 Pipeline Execution Module
5.8 Edge Embedded AI Kit
6 Conclusions and Future Work
References
Anomaly Detection in Manufacturing
1 Introduction
2 Anomaly Detection in Industry
3 Feature Selection and Engineering
4 Autoencoder Case Study
4.1 Autoencoders
4.2 Anomaly Detection for Blow Molding
4.3 Human-Enhanced Interaction
5 Conclusions
References
Towards Industry 5.0 by Incorporation of Trustworthy and Human-Centric Approaches
1 Introduction
1.1 Understanding the Transition to Industry 5.0
1.2 AI, Trustworthy AI, and its Link to Industry 5.0
1.3 PRM and Considerations for AI Implementation
2 Failure Mode and Effects Analysis
3 TAIโPRM Protocol
3.1 e-Risk Management Process
4 Risk Analysis and Evaluation Activity
4.1 Use FMEA Activity
4.2 Heat Map Construction
4.3 Perform Analysis
5 Risk Treatment Transfer Terminate or Tolerate Activity
6 Validation and Real Case Scenario
7 TAI-PRM Tool
8 Conclusions
References
Human in the AI Loop via xAI and Active Learning for Visual Inspection
1 Introduction
2 Background
2.1 Overview on Human-Machine Collaboration
2.2 Industry 5.0 and Human-Centric Manufacturing
2.2.1 New Technological Opportunities to Reshape the Human Workforce
2.2.2 Trustworthiness and Implications for AI-Driven Industrial Systems
2.3 Automated Quality Inspection
2.3.1 The Role of Robotics
2.3.2 Artificial IntelligenceโEnabled Visual Inspection
2.4 Realizing Human-Machine Collaboration in Visual Inspection
3 Industrial Applications
3.1 Machine Learning and Visual Inspection
3.2 Human-Digital Twins in Quality Control
3.3 Making AI Visual Inspection Robust Against Adversarial Attacks
4 Conclusion
References
Multi-Stakeholder Perspective on Human-AI Collaboration in Industry 5.0
1 Introduction
2 Related Work
3 Manufacturing Context
3.1 UC1: Quality Inspection
3.2 UC2: Parameter Optimization
3.3 UC3: Ergonomic Risk Assessment
4 Stakeholder Roles
5 Identified Pains
5.1 UC1: Quality Inspection
5.2 UC2: Parameter Optimization
5.3 UC3: Ergonomic Risk Assessment
5.4 Total Results: Pains
6 Expectations Toward the Technical Realization
7 Team Effectiveness
8 Conclusions and Future Work
References
Holistic Production Overview: Using XAI for Production Optimization
1 Use Case Context
2 XAI Approach
2.1 Identification of XAI Needs
2.2 Hybrid Models
2.2.1 Interpretation of XAI Outputs
2.3 Graph Machine Learning Models
2.3.1 Graph Models
2.3.2 Explainability Techniques
3 XMANAI Platform Usage
4 Achievements, Conclusions, and Open work lines
References
XAI for Product Demand Planning: Models, Experiences, and Lessons Learnt
1 Introduction
2 Whirlpool as XMANAI Demonstrator
3 White Appliances Use Case Description
4 Explainable AI Approach
4.1 In-Depth Analysis
4.2 Data Acquisition and Exploration
4.3 Development and Validation of Hybrid XAI Models
4.4 Delivery of the XMANAI Demand Forecasting Manufacturing App
5 Explainable AI Implications and Added Value
5.1 Explanations at Data Level
5.2 Explanations at Instance and Model Levels
6 Application of the XMANAI Platform and Manufacturing App
6.1 XMANAI Platform
6.2 XMANAI Manufacturing App
7 Evaluation of the XMANAI Solution
8 Conclusions and Lessons Learnt
References
Process and Product Quality Optimization with Explainable Artificial Intelligence
1 Introduction
2 The CNH Industrial XMANAI Demonstrator
2.1 Use Case Description
2.2 XAI Technical Implementation
2.3 Explainability Value
2.4 XMANAI Manufacturing App Experience
2.4.1 XMANAI Platform and Components Usage
2.4.2 XAI Powered Manufacturing Web App
2.4.3 AR App Design
2.4.4 Evaluation of XAI Platform
3 Conclusion and Lessons Learnt
References
Toward Explainable Metrology 4.0: Utilizing Explainable AI to Predict the Pointwise Accuracy of Laser Scanning Devices in Industrial Manufacturing
1 Introduction
2 Background and Related Work
2.1 Optical Metrology
2.2 Explainable Artificial Intelligence
3 Methodology
3.1 Setting Up a Supervised Learning Task
3.2 Data Sources
3.3 Data Pre-processing
3.4 Feature Extraction
3.5 Model Training, Optimization, and Validation
3.6 Comparative Analysis
3.7 Generation of Explanations
4 Results
4.1 Model Performance โ Quantitative Analysis
4.2 Qualitative Analysis โ Error Maps
4.3 Explanations
5 Discussion
5.1 Discussion on Experimental Results
5.2 Limitations of the Analysis
5.3 Practical Implications and Future Perspectives
6 Conclusions
References
Index