The transition towards Industry 5.0 is marked by a human-centric approach to technology. It prioritizes the well-being and skill development of operators within industrial processes. In this context, eXtended Reality (XR) has emerged as a powerful tool for industrial training. However, the adoption of XR is currently hindered by the complexity and cost of creating and maintaining training content. This Thesis proposes a paradigm shift towards Training 5.0, introducing an innovative, AI-assisted architecture designed to democratize the creation of XR training programs. The research begins by analysing the efficacy of immersive learning technologies and identifying the gap between technological potential and practical implementation. To bridge this gap, the Thesis presents a novel authoring framework that uses AI to streamline the creation of immersive training material. This approach allows domain experts without coding expertise to create immersive experiences directly by performing technical procedures. The proposed methodology is validated through multiple case studies, such as high-complexity maintenance procedures for the Extremely Large Telescope (ELT), conducted in collaboration with the Italian National Institute for Astrophysics (INAF), as well as industrial decision-making processes within Volvo Construction Equipment. The results show that integrating AI into the XR creation process can significantly reduce authoring effort while maintaining the effectiveness of training, making it easier to prepare industrial training materials.
La transizione verso l'Industria 5.0 è caratterizzata da un approccio alla tecnologia human-centric. Essa privilegia il benessere e lo sviluppo delle competenze degli operatori all'interno dei processi industriali. In questo contesto, la Realtà Estesa (XR) si è affermata come un potente strumento per la formazione industriale. Tuttavia, l'adozione dell'XR è attualmente rallentata dalla complessità e dai costi legati alla creazione e alla manutenzione dei contenuti formativi. Questa Tesi propone un cambiamento di prospettiva verso il Training 5.0, introducendo un'architettura innovativa basata sull'intelligenza artificiale, progettata per rendere accessibile a tutti la creazione di programmi di formazione XR. La ricerca inizia analizzando l'efficacia delle tecnologie di apprendimento immersivo e identificando il divario tra il potenziale tecnologico e l'implementazione pratica. Per colmare questo divario, la Tesi presenta un nuovo framework di authoring che utilizza l'IA per semplificare la creazione di materiale formativo immersivo. Questo approccio consente agli esperti del settore, anche senza competenze di programmazione, di creare esperienze immersive direttamente eseguendo procedure tecniche. La metodologia proposta è convalidata attraverso diversi casi di studio, tra cui procedure di manutenzione ad alta criticità per l'Extremely Large Telescope (ELT), condotte in collaborazione con l'Istituto Nazionale di Astrofisica (INAF), nonché i processi decisionali industriali all'interno di Volvo Construction Equipment. I risultati mostrano che l'integrazione dell'IA nel processo di creazione in XR può ridurre significativamente lo sforzo di authoring, mantenendo l'efficacia della formazione, rendendo più semplice la preparazione di materiali formativi industriali.
Towards training 5.0: an AI-assisted extended reality approach to industrial training
Giussani, Riccardo
2025/2026
Abstract
The transition towards Industry 5.0 is marked by a human-centric approach to technology. It prioritizes the well-being and skill development of operators within industrial processes. In this context, eXtended Reality (XR) has emerged as a powerful tool for industrial training. However, the adoption of XR is currently hindered by the complexity and cost of creating and maintaining training content. This Thesis proposes a paradigm shift towards Training 5.0, introducing an innovative, AI-assisted architecture designed to democratize the creation of XR training programs. The research begins by analysing the efficacy of immersive learning technologies and identifying the gap between technological potential and practical implementation. To bridge this gap, the Thesis presents a novel authoring framework that uses AI to streamline the creation of immersive training material. This approach allows domain experts without coding expertise to create immersive experiences directly by performing technical procedures. The proposed methodology is validated through multiple case studies, such as high-complexity maintenance procedures for the Extremely Large Telescope (ELT), conducted in collaboration with the Italian National Institute for Astrophysics (INAF), as well as industrial decision-making processes within Volvo Construction Equipment. The results show that integrating AI into the XR creation process can significantly reduce authoring effort while maintaining the effectiveness of training, making it easier to prepare industrial training materials.| File | Dimensione | Formato | |
|---|---|---|---|
|
Thesis_Giussani.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: The transition towards Industry 5.0 is marked by a human-centric approach to technology. It prioritizes the well-being and skill development of operators within industrial processes. In this context, eXtended Reality (XR) has emerged as a powerful tool for industrial training. However, the adoption of XR is currently hindered by the complexity and cost of creating and maintaining training content. This Thesis proposes a paradigm shift towards Training 5.0, introducing an innovative, AI-assisted architecture designed to democratize the creation of XR training programs. The research begins by analysing the efficacy of immersive learning technologies and identifying the gap between technological potential and practical implementation. To bridge this gap, the Thesis presents a novel authoring framework that uses AI to streamline the creation of immersive training material. This approach allows domain experts without coding expertise to create immersive experiences directly by performing technical procedures. The proposed methodology is validated through multiple case studies, such as high-complexity maintenance procedures for the Extremely Large Telescope (ELT), conducted in collaboration with the Italian National Institute for Astrophysics (INAF), as well as industrial decision-making processes within Volvo Construction Equipment. The results show that integrating AI into the XR creation process can significantly reduce authoring effort while maintaining the effectiveness of training, making it easier to prepare industrial training materials.
Dimensione
4.13 MB
Formato
Adobe PDF
|
4.13 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/255677