The integration of Extended Reality (XR) and the Internet of Things (IoT) enables immersive and interactive control of physical devices within virtual environments, offering significant potential across domains such as smart environments, education, and industrial automation. However, achieving real-time responsiveness remains a critical challenge, particularly when cloud-based solutions introduce latency and reliability concerns. This paper explores two approaches to integrating IoT devices into a Unity-based XR environment deployed on the Meta Quest 2: a Tapo smart bulb controlled via IFTTT webhooks and an ESP32 microcontroller operating as a local HTTP server. The XR environment, designed to replicate a university building, incorporates physics-based interactions and hand tracking to provide an intuitive user interface. Comparative evaluation of the two approaches highlights the advantages of localized IoT processing using the ESP32, which achieves low-latency performance and greater reliability compared to the cloud-dependent Tapo bulb integration. The findings demonstrate the feasibility of cost-effective, real-time IoT-XR integration and provide insights into future development of scalable and responsive systems.

Hyperspectral image (HSI) classification is central to environmental monitoring, yet real-time deployment of deep learning models on resource-constrained edge platforms remains challenging due to high spectral dimensionality and computational overhead. In this paper, we propose TinyCapsViT, an ultra-lightweight hybrid architecture that integrates convolutional feature extraction, transformer-based attention, and capsule-inspired representation learning for efficient HSI classification under strict TinyML constraints. The model employs a minimal convolutional stem using pointwise and depthwise separable convolutions to capture local spectral-spatial features, followed by a compact tokenization strategy and learnable positional embeddings. A lightweight self-attention module enables global context modeling with reduced computational complexity, while a capsule-inspired refinement block with squash nonlinearity and residual scaling enhances feature discrimination. The proposed architecture contains only 2781 trainable parameters, representing up to a 30× reduction compared with larger architectures such as ResNet and Vision Transformer, and approximately 11× and 9.5× fewer parameters than the CNN and 3D-CNN baselines, respectively. Extensive experiments across benchmark hyperspectral datasets demonstrate that TinyCapsViT achieves up to 99.61% overall accuracy and maintains competitive classification performance despite its substantially reduced model complexity. Although TinyCapsViT does not consistently achieve the highest classification accuracy compared with larger baseline models, it provides a favourable trade-off between classification performance and computational efficiency. These results demonstrate the potential of TinyCapsViT as a practical solution for real-time hyperspectral analysis on resource-constrained edge platforms and UAV-based environmental monitoring applications.

This paper proposes a new Unified Namespace (UNS)-based architecture to improve predictive and prescriptive maintenance of industrial equipment and overcome challenges such as incomplete data, poor interoperability, and disconnected IT/OT environments. The framework combines interoperable data formats in real-time sensor data, predictive modeling, prescriptive analytics, and simulations of digital twins, using UNS as a centralized, protocol-agnostic data layer that is scalable and complies with Industry 4.0 and Pharma 4.0 standards. The suggested methodology increases data accessibility, reduces integration complexity, and allows low-latency analytics and automated decision-making. Machine learning predictive models achieved more than 94% accuracy in predicting equipment failures. Prescriptive analytics provides maintenance recommendations to reduce downtime and risks. The feedback loops of digital twins can enhance the accuracy of predictions and allow decision optimization through what-if analysis. A test-bench deployment showed a higher performance compared to traditional point-to-point integration, with lower latency (approximately 18 ms vs. approximately 31 ms), decreasing packet loss (0.40% vs. 3.11%), and higher model accuracy (94.20% vs. 87.51%). The structure avoided more than 4000 simulated breakdowns in the test-bench environment, indicating dependability. The study connects the theoretical applications of the UNS with the actual maintenance processes and provides a sound approach to the industrial analytics and optimization of the equipment.

The primary focus of enhancing the efficiency of operations in the Industry 4.0 setting is Predictive and Preventive Maintenance (PPM). The paper introduces a predictive-maintenance system based on the Unified Namespace (UNS), which involves real-time sensor measurements, photogrammetry, and modelling of a digital twin to improve fault prediction and responsiveness to maintenance. This experiment was conducted over six months in a medium-sized discrete electromechanical production plant equipped with motors, Variable Speed Drives (VSDs), robot/cobots, precision grip systems, pipework systems, Magnemotion/linear motor drives, and a CNC machine. The continuous data, such as high-frequency vibration, temperature, current, and pressure, were monitored and analysed with machine-learning models, including support-vector machines, Gradient Boosting, long-short-term memory, and Random Forest, through which temporal degradation can be predicted. UNS architecture integrated all sensor and imaging data into a vendor-neutral data model through OPC UA to help ensure that all experiments could be integrated consistently and be updated in real time to real digital twins. The suggested system correctly identified mechanical and electrical failures and predicted failures before they really took place. Consequently, machine downtime was reduced by 42.25%, and Mean Time to Repair (MTTR) by 36%, compared to the prior six-month baseline period. These improvements were associated with earlier anomaly detection and digital-twin-supported pre-inspection. Overall, the findings indicate that the integration of UNS with multi-modal sensing and digital-twin technologies may enhance predictive maintenance performance in comparable industrial settings. The framework provides a data-driven, scalable solution to organisations that aim to modernise their maintenance processes, attain greater reliability and better equipment utilisation, as well as enhanced Industry 4.0 preparedness.

Hyperspectral image (HSI) classification is vital for environmental monitoring, land cover mapping, and precision agriculture, but its effectiveness is often constrained by the scarcity of labeled samples and high spectral similarity among classes. To address these challenges, we propose CrossCapsViT, a hybrid classification framework that integrates capsule networks (CapsNets) and vision transformers (ViTs) through a cross-attention fusion mechanism and a cross-layer adaptive fusion module, enabling richer and more discriminative spectral-spatial feature learning. To further improve efficiency in data scarce scenarios, we embed an actor-critic reinforcement learning-based active learning (RAL) strategy that jointly leverages accuracy, uncertainty, and diversity in the reward structure, guiding the selection of the most informative samples while reducing labeling effort. Experiments conducted on four benchmark datasets (Kennedy Space Center, Pavia University, Houston University 2013, and Salinas) and a custom UAV-based saltmarsh dataset (Derrymore, collected with a Pika-L sensor) demonstrate that CrossCapsViT with RAL consistently outperforms CapsViT and other baseline models in terms of classification accuracy, robustness, and generalizability. The proposed framework achieves up to 25% improvement in class-level accuracy on challenging vegetation classes, while reducing dependence on large annotated datasets, highlighting its potential for practical deployment in real-world ecological monitoring and remote sensing applications.

The transition toward fully wireless industrial automation under Industry 4.0 requires communication systems capable of deterministic and ultra-reliable performance. This study experimentally evaluates the potential of multi-connectivity, specifically the simultaneous use of private 5G, public 5G/LTE, and Wi-Fi in order to meet the 99.9% reliability target defined in 3GPP TS 22.104 for mission-critical applications. Using a mobile robotic platform within the Confirm Smart Manufacturing Centre, four experimental scenarios were conducted under stationary and mobile conditions with 64 byte and 25000 byte payloads. Results show that no individual interface achieved the 99.9% reliability threshold independently, however composite multi-path configurations employing packet duplication across interfaces consistently met or exceeded this benchmark, achieving up to 100% reliability. Dual private 5G + Wi-Fi and triple-interface configurations eliminated single-path deficiencies, while private 5G demonstrated superior stability under mobility due to its single cell architecture. Although packet duplication increases overall network load, it ensures mission-critical communication with a minimum reliability of 99.98%, fulfilling deterministic performance requirements for industrial control and robotic systems. These findings confirm that multi-connectivity offers a practical and robust approach to achieving ultra-reliable wireless communication in industrial environments.

Cyber-physical systems (CPS) represent a significant advancement in the integration of computational and physical processes, with impactful applications across industries such as healthcare, manufacturing, transportation, and smart grids. This paper examines the current state of CPS, focusing on key applications, security and privacy challenges, and the complexities of systems integration and interoperability. An extensive literature search synthesizes findings from recent research to outline fundamental CPS concepts and their practical implementations. The paper provides a critical analysis of existing solutions to security vulnerabilities and discusses ongoing efforts to enhance privacy protections within CPS environments. Identified gaps in the current research landscape are presented, alongside potential directions for future investigations. The analysis underscores the necessity for continued innovation and robust security measures to harness the full potential of CPS, ensuring reliability and efficiency in an increasingly interconnected world.

Saltmarshes are critical coastal ecosystems that are increasingly threatened by the accumulation of marine and terrestrial debris, including plastics, metals, wood, and other anthropogenic litter. This search focuses on a UAV based remote sensing framework for detecting and isolating such debris within saltmarsh environments by masking out natural background elements such as vegetation, soil, and water. High resolution multispectral was collected using a DJI Matrice 300 UAV equipped with an AGROWING Alpha 7R Sextuple camera and ground truthed via vegetation quadrats at Derrymore Island, Ireland. Spectral indices including NDVI, GNDVI, and MNDWI were employed to generate binary masks for key land cover types, enabling precise identification of anomalous debris. The proposed method demonstrates a scalable, noninvasive and semiautomated approach to debris detection in complex saltmarsh terrains. Results demonstrate high accuracy in debris detection and classification, highlighting UAV based multispectral imaging as an efficient, accurate, and scalable approach for environmental monitoring and marine debris assessment in sensitive coastal ecosystems.

The integration of private 5G networks in Industry 4.0 presents a significant opportunity to enhance manufacturing automation, real-time machine-to-machine (M2M) communication, and industrial IoT applications. This study evaluates the throughput and latency performance of private 5G networks in smart manufacturing environments, focusing on two distinct Radio Access Network (RAN) platforms: Stock RAN and Open RAN. The experimental analysis assesses subscriber-to-subscriber and subscriber-to-data centre communication, measuring performance under varying network loads. The results highlight key trade-offs between Open RAN and Stock RAN, emphasizing the need for application-specific deployment strategies in private 5G networks. These insights contribute to ongoing research in wireless industrial networking, guiding network engineers and system integrators in optimizing private 5G deployments for smart manufacturing applications.

The transition from the basic, reactive approach to maintenance toward the more sophisticated predictive and prescriptive maintenance options respond to key issues in industrial processes, including unforeseen halts, costly repairs, and poor resource management. Reactive maintenance focuses its efforts on post-failure action leading to interruptions as well as the imposition of significant economic costs while on the other end preventive maintenance, which often prompts additional unneeded activities alongside high costs. Predictive maintenance is useful in monitoring failures and its major disadvantage is that it may provide recommendations that are difficult to implement while prescriptive maintenance faces challenges such as integration of data, standardization of the process and lack of skilled professionals. To overcome these challenges, the proposed solution incorporates a Unified Namespace (UNS) architecture. UNS has become a real-time data management platform collecting data in various formats and providing intuitive workflow between devices. It is pivotal in the case of prognostic models that anticipate equipment breakdown and decision models that suggest choices that will result in better utilization of available resources, minimum time wastage, and high equipment reliability. Other applications for predictive maintenance include digital twins that improve the monitoring and testing processes required for sound decision making. Thus, the proposed UNS allows for deciphering integration and data standardization challenges and, therefore, contributes to scaling up the industrial maintenance practices.

The convergence of the Internet of Things (IoT) and the construction industry represents a potent opportunity to fundamentally redefine the processes of constructing and maintaining buildings. The integration of IoT technology has the potential to revolutionize building safety, efficiency, and sustainability, offering a promising avenue for innovation and advancement within the industry. In this work, we propose an innovative design and implementation of an interoperable physical access control IoT-based system that can be easily integrated with existing access control security systems and used in various environments such as offices, homes, and public buildings. Based on our developed and implemented physical access control IoT-based system, we identified several challenges in the current state of access control systems. These challenges include difficulties in integrating different security systems, the absence of interoperable standards, and the need for further research to enhance the scalability and performance of such systems.

This paper presents a detailed performance evaluation of two private 5G standalone (SA) packet core implementations Nokia and Raemis within a smart manufacturing context. Six representative test scenarios were designed to assess core network behaviour under varying traffic loads and operational conditions. These include subscriber-to-subscriber and subscriber-to-data centre communications, with both uplink and downlink throughput and latency as key metrics. Experimental results reveal that while both packet cores deliver sufficient baseline performance, they exhibit contrasting strengths, the Raemis core consistently achieved higher throughput, particularly in subscriber-to-subscriber scenarios, whereas the Nokia core demonstrated more stable performance under high device density and concurrent traffic. Latency analysis shows that both cores satisfy 3GPP Service performance requirements for mobile robots specifically use case C. Importantly, this study demonstrates the feasibility of deploying both packet cores concurrently using dual Public Land Mobile Network (PLMN) configurations. This architectural flexibility allows network operators to dynamically assign traffic to the core that best meets specific performance requirements, for instance, selecting Raemis for high-throughput applications and Nokia for latency-sensitive workloads. The findings contribute actionable insights into the deployment of private 5G networks for Industry 4.0, emphasizing the need to tailor core selection to application-specific QoS demands in industrial environments.

In the context of Industry 4.0, this study presents a defect-aware, closed-loop manufacturing system that integrates low-cost Internet of Things (IoT) devices with robotic automation to enhance quality control in 3D printing. A custom ESP32-based IoT device continuously monitors environmental and vibrational data to enable early defect detection during the printing process. If the predicted print quality falls below a configurable threshold (60%), the system halts the operation and instructs a UR3 robotic arm to remove the defective part— demonstrating full interoperability within a digital twin framework. The integrated machine learning model provides real-time feedback and autonomous control. More than 50 test scenarios confirmed the system’s reliability, with latency remaining well within real-time operational limits. By reducing material waste and energy consumption, the proposed solution offers a scalable and sustainable approach to intelligent additive manufacturing.

Digital twins (DTs) are transforming the manufacturing sector by enabling real-time monitoring, predictive maintenance, and process optimization. This paper systematically reviews the ten most cited research papers on digital twins in manufacturing, identifying key themes, successes, challenges, and future research directions. While rooted in the manufacturing domain, the findings reveal insights highly transferable to renewable energy and power systems—particularly in areas such as data-driven decision-making, AIdriven predictive modelling, and cyber-physical system integration. Notably, the technologies and methodologies explored—such as real-time data synchronization, simulation-based planning, and AI-enhanced analytics—can support smart grid optimization, renewable energy forecasting, and infrastructure resilience. The paper advocates for greater crosssector adoption of digital twin frameworks, particularly in energy system planning, energy efficiency strategies, and sustainability-driven innovation. This positions digital twins as not only pivotal for Industry 4.0 but also as enablers of the global transition toward secure, efficient, and intelligent energy ecosystems.

Saltmarshes are critical coastal ecosystems that offer biodiversity, shoreline protection, and high carbon sequestration. Monitoring these habitats is essential but challenging because of their spatial complexity and sensitivity. This paper presents a Unmanned Ariel Vehicle (UAV) based hyperspectral survey conducted over Derrymore Saltmarsh, County Kerry, Ireland, using a DJI Matrice 300 drone equipped with a Resonon Pika-L hyperspectral imager. We demonstrate the effectiveness of hyperspectral data and Spectral Angle Mapper (SAM) classification for identifying key saltmarsh vegetation species. Field data from vegetation quadrats were used for validation. Our results highlight the potential of UAV and HSI integration in supporting ecological research, improving habitat mapping accuracy, and enabling scalable, nondestructive monitoring.

The optimisation of 3D printing parameters for manufacturing biomedical devices is an emerging interdisciplinary field that incorporates artificial intelligence techniques such as machine learning and deep learning. In this particular study, the focus is on the fabrication of biocompatible finger splints using digital light processing 3D printing technology, followed by UV curing, to evaluate their quality. By leveraging vibration data from printers, which cannot be captured through visual inspection of layer defects, this study aims to develop a predictive model for assessing the failures of printed parts. Here, a closed-loop detection system is proposed to identify failure phenomena in 3D resin printing, combining both cloud and edge computing technologies to effectively detect and address potential failures in the printing process.

Virtual Reality (VR) technology is a powerful tool in the ongoing quest to address global challenges. VR's immersive and interactive nature enables more engaging and effective solutions for addressing real-world technical challenges and improving the overall human experience. It offers an innovative approach to simulate real-world scenarios within a virtual realm. The work being presented bridges the gap between the physical and the virtual by crafting a virtual environment that empowers users to interact in real time with critical real-world issues through the tool of virtual reality. The exhibited work offers more than just technological innovation; it immerses users in an intricately replicated virtual environment that seamlessly mirrors an actual building allowing the user to interact within the virtual domain and control the physical domain. Through this convergence of technology and reality, a glimpse of the future where advancing VR technologies can open doors (physically and virtually) delivering novel methods of interacting with and managing our surroundings.

Deep learning based visual cognition has greatly improved the accuracy of defect detection, reducing processing times and increasing product throughput across a variety of manufacturing use cases. There is however a continuing need for rigorous procedures to dynamically update model-based detection methods that use sequential streaming during the training phase. This paper reviews how new process, training or validation information is rigorously incorporated in real time when detection exceptions arise during inspection. In particular, consideration is given to how new tasks, classes or decision pathways are added to existing models or datasets in a controlled fashion. An analysis of studies from the incremental learning literature is presented, where the emphasis is on the mitigation of process complexity challenges such as, catastrophic forgetting. Further, practical implementation issues that are known to affect the complexity of deep learning model architecture, including memory allocation for incoming sequential data or incremental learning accuracy, is considered. The paper highlights case study results and methods that have been used to successfully mitigate such real-time manufacturing challenges.

Virtual and Augmented Reality (VR/AR) experiences are being developed by a large global community using diverse game engines and frameworks. Each research laboratory or private company is developing its own solutions, and there is a common tendency to re-implementing basic Extended Reality (XR) tools that have already been developed by other people. The overarching issue is the lack of an open-source framework that would combine a sufficient set of modules to provide common XR functionality without additional development effort. In addition, the modules must have clear semantic distinctions and be easily discoverable for both experienced and novice users. This work aims to propose the architecture of such framework based on the Unity game engine and distributed as open-source software. The envisioned solution can facilitate better collaboration and knowledge exchange between XR researchers and industry professionals by providing a common standard for sharing experiences and toolkits developed in this domain. The description is supplemented by potential use cases of the XR framework in educational and industrial scenarios.

Optimised human machine interfaces for multi-robot systems are essential for human in the loop in cyber-physical production lines and collaborative systems. The speed of changeover and customized production processes together with the need for straightforward easy and user-friendly human-machine interaction methods demands natural and adaptable interfaces. They should be based on flexible software cores and packages that ease and speed up the development processes. This also applies to laboratory testing and assessment in the academic field, in particular when it comes to the deployment of virtual or augmented reality (AR/VR) interfaces. The standardization of human-robot interaction including control methods in the extended reality domain is a work in progress. It needs broader assessment and clearer metrics to realise efficient and reliable tools. This work presents an Extended Reality (XR) user interface for the control and teleoperation of industrial robots. The systems allows the fast integration of the digital twins of robotic arms and path planning interface in AR and VR using Robot Operating System and Unity. Furthermore, a design-of-experiment involving two different robots (ABB IRB 1200 and ABB IRB 1600) in the two geographically distributed locations is proposed along with some preliminary experimental results.