XR Training in Industry 5.0: Advancing Human-Machine Collaboration with the XR5.0 Training Platform

XR Training in Industry 5.0: Advancing Human-Machine Collaboration with the XR5.0 Training Platform Jorge Oliveira1,2(B), Tomaz Saraiva1, Harsh Manoj Shah3, Emmanouil Mavrogiorgis4, Argyro Mavrogiorgou5, and Athanasios Kiourtis5 1 Immersive Lives, Lispolis, n44, 1600-546 Lisbon, Portugal {jorge.oliveira,tomaz.saraiva}@immersivelives.pt 2 University Lusófona/HEI-Lab, Campo Grande, 376, 1749-024 Lisbon, Portugal 3 HOLO-Industrie 4.0 Software GmbH, Carl-Zeiss-Ring 19, 85737 Ismaning, Germany [email protected] 4 Synelixis Solutions S.A, 157 Perissou and Chalkidos, 14343 Athens, Greece [email protected] 5 University of Piraeus, 80 M. Karaoli and A. Dimitriou, 18534 Piraeus, Greece {margy,kiourtis}@unipi.gr Abstract. Industrial Training, especially training geared towards Industry 5.0 – referring to robot and smart machines working alongside people, is an evolving field, and recent technological advancements in Extended Reality (XR) and Artificial Intelligence (AI) have propelled interest toward this goal. The combination of these technologies allows the implementation of immersive, adaptive, and personalized learning experiences, which can be utilized by the workforce in onand off-the-job contexts to address training in increasingly complex industrial systems. However, the adoption of XR-based training faces several challenges, including computational demands, latency, usability constraints, and personalization. To address these limitations, the XR5.0 Training Platform provides a state-of-the-art cloud infrastructure and AI-enhanced training solution designed to create, manage, and display XR content to users with optimized performance and accessibility. The platform is structured around three (3) core components, namely: (i) the Hololight Hub, for managing and orchestrating XR applications enabling low-latency streaming via a cloud-based infrastructure; (ii) the XR Training Asset Repository to ensure secure storage of training materials; and (iii) the XR Training Management System for the creation, management, and visualization of XR-native training programs. This platform addresses the limitations of existing training platforms while reducing hardware dependency by adopting a device-agnostic approach. T his ensures a more efficient and scalable training ecosystem, enhancing workforce alignment with Industry 5.0 environments. This paper presents the platform’s architecture, key functionalities, and integration strategies while discussing its potential to transform industrial training through intelligent, immersive, and data-driven learning solutions. Keywords: Extended Reality · XR Streaming · Artificial Intelligence · Training · Industry 5.0 © The Author(s), under exclusive license to Springer Nature Switzerland AG 2026 L. T. De Paolis et al. (Eds.): XR Salento 2025, LNCS 15738, pp. 93–100, 2026. https://doi.org/10.1007/978-3-031-97769-5_7 94 J. Oliveira et al. 1 Introduction Industry 5.0 marks a significant shift from the automation paradigm of Industry 4.0 towards a human-centric paradigm, emphasizing on human-machine cooperation [1]. This approach integrates advanced technologies with human expertise to create more productive and worker-friendly environments [2]. One of the key areas where Industry 5.0 is driving innovation is industrial training, mainly through the utilization of advanced technologies such as Extended Reality (XR) and Artificial Intelligence (AI) [3]. These technologies enable more immersive, adaptive, and personalized learning experiences, providing customized environments for training industrial operators in complex industrial procedures [4]. Following a report by MarketsandMarkets [5], the XR market is predicted to grow from $15.3 billion in 2020 to $77.0 billion by 2025, driven by the increasing utilization of AR in domains including industrial training, showcasing its overall importance and value. Despite these opportunities, several challenges contribute to the slow adoption of XR-based solutions in the industrial domain [6]. The current limitations include latency needed for industrial settings that are constrained by technical specifications related to rendering frame rate, the refresh rate of XR display, and the input lag from the interacting content due to higher computational demands [7]. Additionally, XR’s difficulty for beginners and usability challenges (especially pronounced when XR systems do not adequately address concepts like presence and immersion) are limiting XR-based training solution adoption in real-world industrial settings [8]. Reviews of existing training platforms for industrial scenarios suggest a predominant focus on technical and technological subjects while often neglecting the human-centric aspects of training [9, 10]. To this context, in this manuscript the XR5.0 Training Platform is being introduced, as a component of the XR5.0 project. The latter is an EU-funded initiative aiming to overcome these challenges by establishing specific guiding principles and strategies for implementing XR technologies within Industry 5.0. XR 5.0 aims to ensure seamless human-machine collaboration, enabling operators to learn in highly interactive environments that adjust in real-time according to user performance and needs [11]. As part of this project, the XR5.0 Training Platform will be developed that integrates advanced XR technologies with AI tools to enhance training efficiency, adaptability, and realworld applicability [8]. This project utilizes XR streaming solutions that stream XR content directly from the cloud to XR end-devices to mitigate latency and computational demand-related issues, ensuring a fluid and more natural interaction in industrial settings. It also contains a module designed to create XR-native training programs, able to create a more dynamic, immersive, and personalized training experience. The rest of this paper is organized as follows. In Sect. 2 it is being described the XR5.0 Training Platform’s architecture and vision, including its core components and key features, while Sect. 3 concludes this work, discussing the future work of the platform. While this paper focuses primarily on the architectural and functional design of the XR5.0 Training Platform, a future iteration will detail the experimental design, data collection protocols, and analysis techniques planned for evaluating training outcomes, usability, and system performance in real-world industrial scenarios. XR Training in Industry 5.0 95 2 XR5.0 Training Platform The XR5.0 Training Platform is a cloud-based XR training and streaming platform that will enable Industry 5.0 operators to on-demand render XR training applications specific to selected industrial scenarios within the Industry 5.0 ecosystem. The end-users or operators will be able to leverage the training platform to undergo immersive and AIpowered training in Augmented Reality/Virtual Reality (AR/VR) mode with real-time interactions. The platform supports administrative operations through web browsers and delivers XR experiences to operators via XR devices, such as AR/VR headsets (e.g., Meta Quest and Microsoft Hololens). The platform’s architecture (Fig. 1) ensures seamless management of applications and users, robust orchestration of cloud resources, and efficient XR content delivery. Fig. 1. XR5.0 Training Platform Architecture 2.1 Training Platform Architecture As shown in Fig. 1, the core of the XR5.0 Training Platform is composed of the below three (3) primary components: •Hololight Hub -It is the core streaming platform for managing, hosting, and streaming XR applications. •XR Asset Repository - It refers to a cloud-based file storage system for storing training materials and programs. •XR Training Management System - It allows administrators to create training materials and immersive step-by-step training modules, by linking training materials stored in the repository. In addition to the core components, the platform utilizes two (2) supporting layers, namely the Graphical User Interface (GUI) layer, which serves as the interaction layer, as well as the Virtualization Provider, which serves as the infrastructure layer. More information is provided below: 96 J. Oliveira et al. •Graphical User Interface (GUI) - It consists of the Web Browser and the XR Device Interface, where: •The Web Browser can be used by administrators/managers t hrough a central dashboard to manage XR applications, training programs, and user activities. •The XR Device Interface can be used by the operators to access immersive training through XR devices (e.g., HoloLens 2, Meta Quest 3), enabling real-time interaction with XR content. •Virtualization Provider - It consists of High-performance Virtual Machines (VMs) and the XR applications, where: – The High-performance VMs with Graphics Processing Units (GPUs) run XR applications hosted on the Hololight Hub, ensuring scalability and reliability. – The XR applications built in Unity communicate with external services developed within XR5.0 project like the Human-Centric Digital Twin and AI models (AL, NSAI, GenAI) via Application Programming Interfaces (APIs), incorporating personalization and General Data Protection Regulation (GDPR)-compliant data security. It should be mentioned that the XR applications are built with Unity based plugins, referring to: (i) the Hololight Stream that is used to enable XR application streaming from the cloud to the XR end devices; (ii) the XR Repository plugin that is used to download files such as 3D models from the asset repository storage into the XR environment; and (iii) the XR Training Plugin that is used to render the training programs developed using the web browser into the XR environment. Within the XR5.0 Tr aining Platform, the training interaction workflow begins with the administrator adding users, XR applications, related assets on the platform, and training materials, which can be further combined in the training programs via the web interface. Authorized trainees run the XR application assigned to them on the dedicated XR device, whereas within the XR application, the trainees can access the training programs and materials via the XR Training Plugin and run the programs in an immersive environment. 2.2 Core Components The platform is developed around a virtualization environment using Amazon Web Services (AWS)-based cloud services for managing multiple XR applications, each equipped with plugins that allow the communication between Hololight Hub, the Asset Repository Plugin, and the XR Training Plugin providing real-time rendering and content delivery to XR devices. The platform will lean on its core components, further explained below. Hololight Hub. The Hololight Hub (Fig. 2) is a cloud-based system designed to host, manage, and stream XR applications in real-time to ensure a low-latency and smooth XR experience across different devices, such as the Meta Quest and Microsoft Hololens. This component centralizes XR application management, allowing organizations to manage users and XR applications and settings effectively. The Hololight Hub uses AWS cloud XR Training in Industry 5.0 97 infrastructure to orchestrate and run XR applications and leverages Hololight Stream to ensure immersive, lag-free, and secure XR streaming. Fig. 2. Hololight Hub Architecture The s ystem follows a microservices-based framework, with an API Gateway sending requests to different services related to user and application management, analytics, and resource orchestration. It also integrates external services such as Single Sign-On (SSO) for secure login, a database for managing user and application data, and AWS virtual machines for high-performance XR rendering. XR Asset Repository. The XR Training Asset Repository (Fig. 3) is designed to provide flexible and secure storage of industrial training materials. The core element of the XR Training Asset Repository is a Web App designed to manage the storage of files and metadata linked to training materials. Its primary purpose is to ensure flexibility, allowing files to either be stored close to the Virtualization Provider for enhanced performance or towards the edge to address security concerns or technical constraints (For example, specific 3D models would need to stay on-premises due to security concerns). The app incorporates Owncloud, an open-source file-sharing solution, to store training-related assets and serves as a connector between Owncloud and Training Platform components like Hololight Hub and the XR Training Management System. It enables administrators to create training materials, assign permissions, share content with other users, and allocate storage space for specific users based on their requirements. Additionally, the Web App allows XR Devices to access (view/download) training materials via the XR training plugin or enables users to interact with the content through a standard Web browser. XR Training Management System. The XR Training Management System (Fig. 4) enables the creation, management, and visualization of the XR Training Materials and Programs. It consists of two main modules: (i) the Training Programs Authoring Tool (TPAT), a web application for creating and managing training content, and (ii) the XR 98 J. Oliveira et al. Fig. 3. XR Asset Repository Architecture Training Plugin (XRTP), a Unity plugin that allows the visualization of XR training programs and materials on XR devices. Fig. 4. XR Training Management System Architecture The TPAT provides a structured environment for developing XR training content. It includes options for uploading and managing training resources, adapting traditional materials for XR, and generating new training content. The supported formats of materials to be uploaded include PDF files, video files (i.e., MPEG4), images (i.e., PNG) and 3D Models currently supporting different formats (i.e., FBX) that can be uploaded and used as XR training materials. The system allows managers to organize training programs into structured workflows, using features such as interactive checklists, workflow-based instructions, and multimedia content (e.g., images, videos). The TPAT also integrates AI-powered tools to automate the creation of content, f or instance, by extracting and creating text from PDF materials. It is also planned to include AI assistants to support operators within the XR environment. The XRTP is responsible for visualizing XR training content in immersive environments. It provides an intuitive interface optimized for XR interactions, supporting hand tracking, and allowing a natural and immersive experience. Training materials such as checklists, images, videos, and workflows are presented through dedicated viewers, allowing operators to interact dynamically with content. A Training Program Manager organizes available programs, guiding users through structured learning experiences. Planned features include a virtual dashboard for managing views within the XR space and the adaptation feature that will consist of an adaptive interface that personalizes the user experience based on interaction data. XR Training in Industry 5.0 99 2.3 Integration of Components As it has been already explained, the XR5.0 Training Platform integrates the three (3) core services (i.e., Hololight Hub, XR Training Asset Repository, XR Training Management System) for ensuring seamless delivery of XR training content, user management, and system scalability. These components are designed to work in an integrated manner, allowing enterprises to deliver scalable and efficient training experiences. The platform employs a microservices-based architecture with API-driven communication, cloudcentric data exchange, and event-driven orchestration, whereas SSO provides secure, unified user authentication. The Hololight Hub manages authentication and session control, while the XR Asset Repository stores training materials accessed via the XR Training Plugin. Real-time streaming and dynamic resource allocation optimize performance, whereas the orchestration of these components enables efficient management of applications, training modules, resources, assets, user permissions, and secure file sharing. This integration ensures secure access control, efficient content management, and dynamic resource allocation, enabling a robust and scalable XR training experience. To this context, future refinements aim to enhance orchestration, authentication flows, and streaming performance. 3 Conclusion In this paper, the XR5.0 Training Platform has been presented, as a robust solution to address the challenges of industrial training in the Industry 5.0 paradigm. This platform integrates XR technologies with AI and cloud-based systems, providing immersive, adaptive, and human-centric training environments that go beyond traditional methods. By relying on components such as the Hololight Hub, XR Asset Repository, and XR Training Management System, the platform is flexible enough to allow creating, managing and delivering XR content for training in different industrial scenarios aiming to enhance experiential learning. Its architecture ensures scalability, low-latency streaming, and secure access, contributing to effective deployment in real-world applications. This approach advances beyond the state-of-art by offering an integrated s olution that combines XR, AI, and a cloud-based infrastructure to deliver immersive, adaptive and personalized learning experiences better aligned with human factors. Being an XR-based solution, the platform can support both on-the-job and off-the-job training, providing a hybrid learning model where foundational knowledge can be acquired in controlled environments and later reinforced through real-world application being particularly relevant for sectors s uch as manufacturing and maintenance. However, one limitation of the platform is its reliance on high-bandwidth internet connections for real-time streaming, which can be a limitation in remote or low-connectivity industrial settings. As the XR5.0 project evolves, future work will focus on enhancing orchestration capabilities, exploring the metrics for the adaptation and personalization, and expanding interoperability across devices and industrial contexts. Ultimately, the XR5.0 Training Platform lays the foundation for scalable, intuitive, and efficient XR training solutions that take advantage of human-machine collaboration in industrial environments. As part of future work, experimental validation will include pilot s tudies in industrial settings to measure improvements in training efficiency, user engagement, and learning 100 J. Oliveira et al. retention, using both quantitative metrics (e.g., task completion time, error rates) and qualitative user feedback. This will ensure a comprehensive assessment of the platform’s effectiveness. Acknowledgments. The research presented in this paper has received funding from the EU funded Project XR5.0 (GA no 101135209). Disclosure of Interests.. The authors have no competing interests to declare that are relevant to the content of this article. References 1. Zalozhnev, A.Y., Ginz, V.N.: Industry 4.0: underlying technologies. Industry 5.0: humancomputer interaction as a Tech Bridge from Industry 4.0 to Industry 5.0. In: 9th International Conference on Web Research (ICWR) (2023) 2. Moser, T., Wolfartsberger, J., Sorko, S., Abu Naim, B.: Extended reality based education and training for human-centric Industry 5.0 skill enhancement. In: NOMS 2024–2024 IEEE Network Operations and Management S ymposium (2024) 3. Rauschnabel, P.A., Felix, R., Hinsch, C., Shahab, H., Alt, F.: What is XR? Towards a framework for augmented and virtual reality. Comput. Hum. Behav. 133, 107289 (2023) 4. Moser, T., Schlager, A.: Training and assistance in industrial environments using extended reality. In: IEEE Conference on Virtual Reality and 3D User Interfaces Abstracts and Workshops (VRW), pp. 97–103 (2024) 5. Markets and Markets: Augmented and virtual reality market size, share and growth (2024). https://www.marketsandmarkets.com/Market-Reports/augmented-realitymarket-82758548.html 6. Wang, B., Zheng, L., Wang, Y., Fang, W., Wang, L.: Towards the industry 5.0 frontier: review and prospect of XR in product assembly. J. Manuf. Syst., 777–811 (2024) 7. Simões, B., Creus, C., Carretero, M.P., Ochaíta, A.G.: Streamlining XR technology into industrial training and maintenance processes. In: Web3D ‘20: Proceedings of the 25th International Conference on 3D Web Technology, vol. 5, pp. 1–7 (2020) 8. Daling, L.M., Schlittmeier, S.J.: Effects of augmented reality, virtual reality-, and mixed reality–based training on objective performance measures and subjective evaluations in manual assembly tasks: a scoping review. Hum. Factors 66(2), 589–626 (2024) 9. Cazeri, G.T., de Santa-Eulalia, L.A., Serafim, M.P., Anholon, R.: Training for Industry 4.0: a systematic literature review and directions for future research. Brazilian J. Oper. Prod. Manag. 19, 1–21 (2022) 10. Crnobrnja, J., Stefanovic, D., Romero, D., Softic, S., Marjanovic, U.: Digital transformation towards Industry 5.0: a systematic literature review. In: Alfnes, E., Romsdal, A., Strandhagen, J.O., von Cieminski, G., Romero, D. (eds) Advances in Production Management Systems. Production Management Systems for Responsible Manufacturing, Service, and Logistics Futures. APMS 2023. IFIP Advances in Information and Communication Technology, vol 689. Springer, Cham (2023) 11. Kiourtis, A. et al.: XR5.0: Human-centric AI-enabled extended reality applications for Industry 5.0. In: 36th Conference of Open Innovations Association (FRUCT), Lappeenranta, Finland, 2024, pp. 314–323 (2024)