TY - JOUR A1 - Gschoßmann, Lukas A1 - Schedel, Valentin A1 - Süß, Franz A1 - Weber, Markus A1 - Pfingsten, Andrea A1 - Dendorfer, Sebastian T1 - Comparing kinematic and kinetic demands on the knee joint during selected physiotherapy exercises and activities of daily living JF - Technology and health care ; Special issue: BIOMDLORE N2 - Background: The knee is one of the most common areas to suffer injuries or be affected by surgery. Physiotherapy rehabilitation was shown to support recovery, but evidence guiding optimal rehabilitation practices is limited. To recommend appropriate exercises, it is essential to understand the musculoskeletal requirements involved in both physiotherapy and activities of daily living (ADLs).ObjectiveThis study aimed to evaluate and compare the knee joint kinematics, joint forces and muscle activity in knee flexors and extensors during selected rehabilitation exercises and ADLs.MethodsKinematic and kinetic data from 30 healthy participants were collected during 20 different tasks. Full-body musculoskeletal simulations were performed to estimate peak knee joint angles, angular velocities, joint reaction forces, and muscle activity of the knee flexors and extensors.ResultsComparatively high requirements were observed for lunges, squats, stair walking and gait. Medium requirements were observed for sitting down and rising from a chair. Low requirements were observed for balance shifts and variations of the single leg stand.ConclusionOverall, ADLs like gait and stair walking show surprisingly high requirements compared to many exercises employed in physiotherapy. These findings are a step towards biomechanically informed exercise selection and the development of personalized rehabilitation programs. Y1 - 2026 U6 - https://doi.org/10.1177/09287329251413413 N1 - Corresponding author der OTH Regensburg: Lukas Gschoßmann PB - Sage ER - TY - JOUR A1 - Geldof, Arne A1 - Kopačin, Jan A1 - Straus, Izidor A1 - Kriegl, Raphael A1 - Kravanja, Gaia A1 - Hribar, Luka A1 - Jezeršek, Matija A1 - Shamonin, Mikhail A1 - Kokot, Gašper A1 - Drevenšek‐Olenik, Irena T1 - Transfer of Energy and Momentum Between Magnetoactive Surface Microstructure and a Solid Object JF - Advanced Engineering Materials N2 - Transport systems utilizing magnetic materials are very promising for applications that require contactless operation and compatibility with biological processes. A critical parameter in these systems is the efficiency of energy and momentum transfer between the transporting platform and the objects being transported. We investigate the physical mechanisms driving directional transport of solid objects by microlamellar structures laser-inscribed on the surface of a magnetoactive elastomer (MAE). When subjected to a rotating magnetic field with a magnitude of 175 mT and a time period of 0.4 s, the lamellas reorient within a few milliseconds, reaching angular velocities up to 1100 rad s−1. This rapid motion is crucial for efficient momentum and energy transfer to objects in contact with the lamellas. The analysis of collisions of a single lamella with a lead ball with a 2.2 mm diameter shows that the lamella can transfer around 50 nJ of energy, propelling the ball to a speed of around 35 mm s−1. We show how this value sets the upper limit for the ball's transport speed on microlamellar structures. We also explain the background of three distinct transport regimes (kicking, pushing, and bouncing modes) observed on these magnetically driven “conveyor belts”. Y1 - 2026 U6 - https://doi.org/10.1002/adem.202502369 VL - 28 IS - 6 PB - Wiley ER - TY - CHAP A1 - Riemann, Janine A1 - Klug, Katharina A1 - Christ, Luisa A1 - Stadler, Sebastian T1 - Impact of Social Features on User Experience in Sport Tracking Applications T2 - 2025 11th International HCI and UX Conference in Indonesia (CHIuXiD), 05-06. December 2025, Bali, Indonesia N2 - Sport-tracking applications can enhance physical activity but may also induce stress through social comparison. This study examines how social comparison features influence user experience (UX), focusing on motivation, stress, and mental well-being. A survey of 101 German sport-tracking app users assessed intrinsic and extrinsic motivation, perceived stress, self-efficacy, enjoyment, social comparison orientation, and UX. Results reveal that higher social comparison consistently lowers intrinsic motivation, raises extrinsic motivation, and increases stress levels, while positive UX boosts intrinsic motivation and enjoyment and indirectly buffers stress. In particular, users who frequently evaluate their performance against others reported greater pressure to perform, heightened frustration, and a stronger focus on external validation. integrating adaptive feedback, and offering customizable social elements. Overall, the study contributes to understanding how UX and motivational dynamics interact in digital health tools and provides guidance for developing psychologically informed, health-oriented app environments. Conversely, app designs prioritizing mastery and personal progress over competitive ran kings foster autonomy, competence, and relatedness, supporting mental health and long-term adherence. These findings highlight the ambivalent effects of social features and underscore the need for sport-tracking apps to minimize harmful comparisons and promote self-determined motivation through supportive, user-friendly features. Practical implications include designing for sustainable engagement and well-being rather than exclusive competition. Y1 - 2025 SN - 979-8-3315-5572-6 U6 - https://doi.org/10.1109/CHIuXiD68326.2025.11323615 SP - 25 EP - 30 PB - IEEE ER - TY - JOUR A1 - Haug, Sonja A1 - Currle, Edda A1 - Weber, Karsten T1 - Acceptance of Medical History-Taking Supported by Artificial Intelligence and Chatbots: A Population-Based Survey in Germany JF - Healthcare N2 - Background/Objectives: Digital anamnesis tools, including chatbots, are increasingly being developed and evaluated, yet their implementation in German medical practices remains limited. This study examines the acceptance of medical history-taking assisted by artificial intelligence (AI) among the German population. The objective is to derive implications for integrating such systems into digitalization strategies of medical practices. Methods: This study is based on an online survey of the German population, aged between 18 and 74 years, conducted in two independent cross-sectional waves (trend design) in 2024 and 2025 with n = 1000 respondents in each year. Based on the Unified Theory of Acceptance and Use of Technology (UTAUT), three hypotheses regarding the use of AI in medical history-taking were developed and tested using linear regression models. Results: Both waves reveal a high acceptance level of AI-supported anamnesis systems for people aged between 18 and 74, regardless of whether a chatbot is used in medical practice (Scenario 1) or at home (Scenario 2). The latter received slightly less approval for the intention to use (mean intention scores: 3.50 and 3.45, range from 1.0 to 5.0) than Scenario 1 (3.59, 3.56). The indices of Performance Expectancy (PE), Effort Expectancy (EE), and perceived Social Influence (SI) determine the intention to use a chatbot with the strongest correlation of the PE index (Scenario 1: ß =0.466, Scenario 2: ß = 0.475). Most respondents (73% and 75%) expressed a favorable opinion for digitally storing medical history data within their electronic health record (EHR). Conclusions: The findings suggest that gender- and age- specific differentiation—aside from considering the needs of older adults—may be less relevant for designing digitalization strategies than previously assumed. Instead, the focus of medical practices should lie on the practicability of the tool used. Despite currently low EHR utilization rates in Germany, medical practices may expect broad patient approval regarding the digital storage of medical history data. Y1 - 2026 U6 - https://doi.org/10.3390/healthcare14070905 N1 - Corresponding author der OTH Regensburg: Sonja Haug VL - 14 PB - MDPI CY - Basel ER - TY - BOOK A1 - Stadler, Anselm A1 - Mohr, Christa A1 - Teves, Stephanie T1 - Educational Concept - Using VR applications in nursing education T3 - VReduMED Handbook. for VR integration in care education, presenting the project toolkit, an educational concept and recommendations for its sustainable transfer N2 - The following chapter provides teachers with clear guidance on key aspects to consider when introducing XR. It is structured into three main sections. First, the technical prerequisites and conditions are outlined, focusing on aspects that should be taken into account prior to procurement as well as during integration into everyday teaching practice. Building on this, organizational factors that are essential for successful implementation are examined. Finally, the didactic integration of XR is addressed, highlighting the aspects that are critical for effective educational use. The findings from the VReduMED project are complemented by insights from international studies on this topic. Y1 - 2026 UR - https://www.vredumed.eu/wp-content/uploads/VReduMED_Handbook_version_18032026_FINAL.pdf SP - 68 EP - 77 PB - Interreg CENTRAL EUROPE ER - TY - JOUR A1 - Schmiedt, Anja B. A1 - Balakrishnan, Narayanaswamy A1 - Cramer, Erhard T1 - Generalized chi-squared based goodness-of-fit tests under progressive Type-II censoring for exponential and Weibull distributions JF - Communications in Statistics - Simulation and Computation N2 - We propose new goodness-of-fit tests for exponentiality based on progressively Type-II censored data. These tests utilize scale-invariant statistics obtained from the Mahalanobis norm of normalized order statistics, leading to three test statistics, corresponding to 𝐿2-, 𝐿1-, and 𝐿∞-norms of centered uniform spacings. Exact and asymptotic distributions of these statistics are presented. A power study evaluates the proposed tests against existing benchmarks across various alternative distributions and censoring plans, demonstrating superior performance in cases with small and moderate sample sizes. Furthermore, we extend the methodology to approximate goodness-of-fit tests for Weibull distributions via power transformation, ensuring robustness w.r.t. the approximated significance level under unknown shape parameters. An illustrative data example confirms the practical applicability of our tests. Our findings highlight the potential for further extending goodness-of-fit tests under progressive Type-II censoring to other null distributions. Y1 - 2026 U6 - https://doi.org/10.1080/03610918.2026.2625230 PB - Taylor & Francis ER - TY - JOUR A1 - Apel, Thomas A1 - Fellner, Klemens A1 - Kempf, Volker A1 - Salcedo-Lagunero, Reymart A1 - Zilk, Philipp T1 - Lipolysis on Lipid Droplets: Mathematical Modelling and Numerical Discretisations JF - Results in Mathematics N2 - Lipolysis is a life-essential metabolic process, which supplies fatty acids stored in lipid droplets to the body in order to match the demands of building new cells and providing cellular energy. In this paper, we present a first mathematical modelling approach for lipolysis, which takes into account that the involved enzymes act on the surface of lipid droplets. We postulate an active region near the surface where the substrates are within reach of the surface-bound enzymes and formulate a system of reaction-diffusion PDEs, which connect the active region to the inner core of lipid droplets via interface conditions. We establish two numerical discretisations based on finite element method and isogeometric analysis, and validate them to perform reliably. Since numerical tests are best performed on non-zero explicit stationary state solutions, we introduce and analyse a model, which describes besides lipolysis also a reverse process (yet in a physiologically much oversimplified way). The system is not coercive such that establishing well-posedness is a non-standard task. We prove the unique existence of global and equilibrium solutions. We establish exponential convergence to the equilibrium solutions using the entropy method. We then study the stationary state model and compute explicitly for radially symmetric solutions. Concerning the finite element methods, we show numerically the linear and quadratic convergence of the errors with respect to the - and -norms, respectively. Finally, we present numerical simulations of a prototypical PDE model of lipolysis and illustrate that enzyme clustering on lipid droplets can significantly slow down lipolysis. Y1 - 2026 U6 - https://doi.org/10.1007/s00025-026-02639-y VL - 81 IS - 3 PB - Springer Nature ER - TY - CHAP A1 - Hoffmann, Tim A1 - Dietrich, Florian A1 - Melzer, Matthias A1 - Dünnweber, Jan ED - Yurish, Sergey Y. T1 - A Code Generation Framework for Indoor Robot Applications Based on Building Information Modeling (BIM) T2 - Automation, Robotics & Communications for Industry 4.0/5.0: Proceedings of the 6th Winter IFSA Conference on Automation, Robotics & Communications for Industry 4.0/5.0/6.0 (ARCI' 2026) 25-27 February 2026 Salzburg, Austria N2 - We present BIM2Robot, a framework connecting the digital planning world (Building Information Modeling) with real-world robotics. The aim of the tools we developed is to convert building data from IFC building plans (Industry Foundation Classes) such that autonomous robots can use them for navigation and task planning. Conventional robotics systems rely on hard-coded maps or proprietary data. In the BIM2Robot framework, we generate a robot-compatible building model automatically from existing BIM data. This model describes rooms, walls, doors, and connections allowing robots to understand their environment, move around in it, and carry out useful missions. Besides the data transformation tools, our software framework comprises a graphical user interface which allows for interactive planning, path visualization and remote process control. The generated building model can be used directly by ROS-compliant robots or in robot simulation environments, forming a basis for navigation, perception, and interaction. The manual post-processing of the generated robot motion plans is supported at GUI- and also at code-level. Y1 - 2026 UR - https://www.researchgate.net/profile/Sergey-Yurish/publication/401427390_Automation_Robotics_Communications_for_Industry_405060/links/69a567ae16faea00ba9b9b02/Automation-Robotics-Communications-for-Industry-40-50-60.pdf#page=41 SN - 978-84-09-82030-6 SP - 40 EP - 44 PB - IFSA ER - TY - CHAP A1 - Bachir, Hussein A1 - Weikl, Simone T1 - Identifying Cyclist Riding Styles Using Drone-Based Trajectory Data and Volatility Clustering in Free-Flow Traffic Conditions T2 - 2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC), 18-21. November 2025, Gold Coast, Australia N2 - Cycling is increasingly promoted as a sustainable and healthy urban transportation mode. However, understanding cyclists' behavior remains underdeveloped compared to car drivers' behavior research. This paper proposes a two-level unsupervised clustering framework to identify cyclist riding styles in free flow conditions from drone-recorded trajectory data. The first level classifies local riding behavior at the timestamp level using riding volatility measures. The second level identifies overall cyclist behavior through clustering entropy measures, mean rotation fluctuation and surrounding traffic density at the trajectory level. The method was applied to a dataset of 284 trajectories and 100,000 data points from a 500-meter urban road segment in Munich, Germany and revealed two riding patterns on both levels: stable and volatile. The findings highlight the influence of road design clarity and behavioral variability and provide a foundation for cyclist-focused behavior and infrastructure analysis. Y1 - 2025 SN - 979-8-3315-2418-0 U6 - https://doi.org/10.1109/ITSC60802.2025.11423140 SP - 233 EP - 240 PB - IEEE ER - TY - CHAP A1 - Norbisrath, Ulrich A1 - Rossi, Bruno A1 - Jubeh, Ruben A1 - Heydarov, Araz T1 - Multi-Stage Testing for Open Source IoT Frameworks T2 - 2025 IEEE International Conference on Systems, Man, and Cybernetics (SMC), 5-8. October 2025, Vienna N2 - The Internet of Things (IoT) has rapidly evolved, integrating networked intelligence into a web of things, servers, and cloudlets. Although there are various tools and approaches for software testing, the broader field of IoT testing presents unique challenges due to the heterogeneity of devices, large-scale deployments, dynamic environments, and real-time needs. This paper presents our approach to developing a multi-stage testing framework for the IoTempower framework, addressing the challenges of testing a versatile and evolving open-source Internet-of-Things framework used extensively in educational settings and beyond. The framework incorporates compilation testing, integration testing, and system testing with a focus on regression testing. This multi-stage testing approach allows us to validate the framework’s functionality at various granularities, from the correct compilation of individual drivers to the seamless interaction of deployed hardware. This approach aims to proactively identify and prevent regressions, facilitating the integration of new features and enhancements without losing our scope of providing a real hands-on IoT experience in the classroom. Y1 - 2025 SN - 979-8-3315-3358-8 U6 - https://doi.org/10.1109/SMC58881.2025.11343478 SP - 4685 EP - 4690 PB - IEEE ER - TY - CHAP A1 - Barik, Ranjan Kumar A1 - Rawat, Vikram Singh A1 - Manna, Subhrajit A1 - Bandyopadhyay, Ayan Kumar A1 - Hausladen, Matthias A1 - Asgharzade, Ali A1 - Schreiner, Rupert T1 - On Chip Electron Gun Design Using Silicon Tip Field Emitter Array T2 - 2025 IEEE Microwaves, Antennas, and Propagation Conference (MAPCON), 14-18. Dezember 2025, Kochi, India N2 - Silicon based silicon tip array emitter has developed at OTH Regensburg, Germany. This array emitter is the basis for development of a novel electron gun. This field emitter electron sources were fabricated using laser-micromachining technique followed by MEMS technology. In this work, the design and development of electron gun using complete silicon structure is reported. The goal of this work is to develop electron gun using less effort, easy and hassle free technique. Y1 - 2025 SN - 979-8-3315-3722-7 U6 - https://doi.org/10.1109/MAPCON65020.2025.11426578 PB - IEEE ER - TY - CHAP A1 - Michel, Johanna A1 - Krenkel, Lars ED - Dillmann, Andreas ED - Heller, Gerd ED - Krämer, Ewald ED - Breitsamter, Christian ED - Wagner, Claus ED - Krenkel, Lars T1 - Towards Experimental Validation of Models of Shear-Induced Aerosol Generation in the Human Respiratory System T2 - New Results in Numerical and Experimental Fluid Mechanics XV : Contributions to the 24th STAB/DGLR Symposium, Regensburg, Germany, 2024 N2 - Numerical modeling is a valuable tool to research shear-induced aerosol generation inside the human respiratory system. While the volume of fluid method and Eulerian wall film models have been used to predict the stripping of particles from the mucus film, sufficient validation data is lacking. Here, we present an experimental method to create such validation data. A film of mucus mimetic hydrogel with an initial thickness of 1 mm covering the floor of a rectangular channel (75.5 mm 25.5 mm 3 mm) was exposed to an airflow with a flow rate of 9.5 and 21.6 L/min. The number of created particles and the emergence of waves on the mucus surface were measured. Shear-induced aerosol generation was triggered successfully and caused an increase of mean particle flow. Different wave profiles were observed at varying film depths. Y1 - 2026 SN - 978-3-032-11114-2 U6 - https://doi.org/10.1007/978-3-032-11115-9_13 SP - 135 EP - 144 PB - Springer Nature CY - Cham ER - TY - JOUR A1 - Pointner, Daniel A1 - Kranz, Michael A1 - Wagner, Maria Stella A1 - Haus, Moritz A1 - Lehle, Karla A1 - Krenkel, Lars T1 - Automated deep learning based detection of cellular deposits on clinically used ECMO membrane lungs JF - Frontiers in Bioinformatics N2 - Introduction: Despite the promising application of extracorporeal membrane oxygenation (ECMO) in the treatment of critically ill patients, coagulation-associated technical complications, primarily clot formation and critical bleeding, remain a major challenge during ECMO therapy. The deposition of nucleated cells on the surface has been shown, yet the role of these cells towards complication development is still matter of ongoing research. In particular, the membrane lung (MemL) is prone to clot formation. Therefore, the investigation of nuclear deposits on its hollow-fibers may provide insights for a better understanding of the cellular mechanisms involved in the development of ECMO complications. Methods: To support current research, this study aimed to develop a deep learning–based tool for the automated detection and quantitative analysis of nuclear depositions on MemL hollow-fiber mats. A customized fluorescence microscopy workflow, combined with a semi-automated iterative labeling strategy, was used to generate a high-quality dataset for model training. Results: Six configurations of instance segmentation models were evaluated, with a Mask R-CNN with ResNet 101 backbone using dilated convolution providing the most balanced performance in both nuclei count and area accuracy. Compared with U-Net–based approaches such as Cellpose or StarDist, the proposed model demonstrated superior segmentation of overlapping and low-intensity nuclei, maintaining accuracy even in densely packed cellular regions. Discussion: We present an automated image analysis tool for clinically used MemLs, which exhibit complex three-dimensional hollow-fiber architectures and irregular cellular deposits that challenge conventional tools. A dedicated graphical user interface enables streamlined detection, morphometric analysis, and spatial clustering of nuclei, establishing a reproducible workflow for high-throughput analysis of fluorescence microscopy images. This approach eliminates labor-intensive manual counting and facilitates large-scale studies on cell-fiber interactions and disease-related correlations. Y1 - 2026 U6 - https://doi.org/10.3389/fbinf.2026.1771574 N1 - Corresponding author der OTH Regensburg: Daniel Pointner, Lars Krenkel VL - 6 PB - Frontiers ER - TY - JOUR A1 - Stenglein, Anna A1 - Appelt, Andreas T1 - Einsatz der Thermographie als Steuerungsinstrument im Asphaltbau – Erkenntnisse aus Erprobungsstrecken mit temperaturabgesenktem Asphalt JF - Straße und Autobahn N2 - Die Sicherstellung homogener Einbautemperaturen ist eine zentrale Voraussetzung für eine dauerhafte Asphaltbefestigung. Dieser Beitrag stellt Ergebnisse aus der wissenschaftlichen Begleitung zweier Erprobungsstrecken zum Einsatz von temperaturabgesenktem Walzasphalt vor. Mithilfe von Thermoscannern wurden die Oberflächentemperaturen des eingebauten Mischguts flächendeckend erfasst und georeferenziert ausgewertet. Die Analyse zeigt, dass insbesondere Lkw-Wechselstellen, Fertigerstillstände und komplexe Einbausituationen zu ausgeprägten Temperaturgradienten führen. Bei temperaturabgesenkten Asphalten wirkt sich dies aufgrund des eingeschränkten Verdichtungszeitfensters besonders kritisch aus. Die Untersuchungen verdeutlichen, dass Thermoscans über ihre dokumentierende Funktion hinaus als wirksames Instrument der bauprozessintegrierten Qualitätskontrolle eingesetzt werden können. Sie unterstützen die Ausführung, indem thermisch bedingte Schwachstellen frühzeitig erkannt und gezielte Anpassungen im Bauprozess ermöglicht werden. Darüber hinaus werden die Abschiebetechnik und ihre Eignung als Alternative zur Kipptechnik mit Beschicker hinsichtlich Temperaturhomogenität und logistischer Randbedingungen bewertet. T2 - Use of thermography as a control tool in asphalt construction – findings from test sections with temperature-reduced asphalt Y1 - 2026 UR - https://www.kirschbaum.de/fachzeitschriften/strasse-und-autobahn/strasse-und-autobahn/aktuelles-heft-47-1.html#c12609 U6 - https://doi.org/10.53184/STA3-2026-3 VL - 77 IS - 3 SP - 182 EP - 191 PB - Kirschbaum Verlag GmbH ER - TY - INPR A1 - Gaube, Susanne A1 - Jussupow, Ekaterina A1 - Kokje, Eesha A1 - Khan, Jowaria A1 - Bondi-Kelly, Elizabeth A1 - Schicho, Andreas A1 - Kitamura, Felipe Campos A1 - Koch, Timo Kevin A1 - Ezer, Timur A1 - Mottok, Jürgen A1 - Lermer, Eva A1 - Ghassemi, Marzyeh A1 - Colak, Errol T1 - Examining Reliance Patterns on AI Advice in Medical Imaging: a Mixed-Methods Randomized Crossover Experiment N2 - Background: Artificial intelligence (AI) holds significant potential to support diagnostic decision-making; however, evidence regarding its clinical utility remains mixed. Often, the collaboration between clinicians and AI systems does not surpass the individual performance of unaided humans or standalone AI. Yet, currently, the mechanisms that limit human-AI synergy are poorly understood. This study examined the impact of AI advice on diagnostic decision-making among experts and novices, focusing on reliance patterns. Methods: We used a mixed-methods crossover experimental design with a think-aloud and an eye-tracking study arm. Participants were 50 task experts (radiologists) and 75 novices (non-radiologist physicians and medical trainees) from 10 countries. They reviewed 50 head CT scans and every case was examined in three time-separate sessions in randomized order. In each session, participants were exposed to different experimental conditions: (a) control, no AI prediction; (b) basic advice, AI prediction without annotations; and (c) XAI advice, AI prediction with scan annotations. For each case, participants had to determine if the patients had an intracranial hemorrhage (ICH). The main outcomes were diagnostic performance, confidence in the diagnosis, case reading time, and AI advice usefulness ratings. Findings: Both overreliance on incorrect advice and underreliance on correct advice occurred. Underreliance was associated with high uncertainty and, in absolute terms, had a more detrimental impact on diagnostic performance than overreliance. Correct XAI advice reduced underreliance, improved performance (OR=1·84, p<0·0001), and confidence (b=0·15, p<0·0001), particularly when reviewing more difficult cases with ICH. Surprisingly, correct XAI did not reduce reading time (b=1·81, p=0·0713). XAI was perceived as more useful than basic AI advice (b=0·12, p=0·0029), especially among novices. Interpretation: The occurrence of both under- and overreliance highlights the need for efficient counterstrategies beyond classic XAI methods to foster appropriate reliance and synergy. Y1 - 2026 U6 - https://doi.org/10.31219/osf.io/4wv8j_v3 ER - TY - INPR A1 - Achhammer, Anton A1 - Fioriti, Davide A1 - Patonia, Aliaksei A1 - Sterner, Michael T1 - The impact of hydrogen underground storage on fair partnerships: a GIS-based integration of salt caverns into PyPSA-Earth N2 - The increasing demand for hydrogen in Europe and the development of cross-border infrastructure, such as the SoutH2 Corridorconnecting Tunisia, Italy, Austria, and Germany, underscore the importance for hydrogen storage solutions to ensure supplysecurity and competitive pricing. Without storage, producers face increased market dependency, as electrolyzers require con-tinuous operation to remain economically viable. At the same time, storage offers opportunities to strengthen domestic valuechains by securing hydrogen supply for local industries. To assess the system-level impact of underground hydrogen storageand its implications for hydrogen partnerships, we integrate GIS-based salt cavern potentials into PyPSA-Earth and apply theframework to Tunisia. Salt caverns are currently largely considered the most economical option for large-scale hydrogen storage,offering operational flexibility. Underground storage is represented as an endogenously optimised, regionally constrained option,enabling a direct comparison between scenarios with and without geological storage under identical demand, technology, andpolicy assumptions.Our results show that underground hydrogen storage enables seasonal balancing at multi-terawatt-hour scale, reshaping hydro-gen system design. Storage availability substitutes most aboveground hydrogen tank capacity, improves electrolyser utilisation,and reduces levelised hydrogen production costs by approximately 0.10 € kg−1. Moreover, it decouples hydrogen production fromshort-term electricity variability and export demand, enhancing supply stability and export competitiveness.Beyond the Tunisian case, the findings underscore the strategic role of geological storage in international hydrogen trade. Byincreasing resilience and reducing cost volatility, underground hydrogen storage strengthens the position of exporting regionsand supports more balanced and sustainable hydrogen partnerships. KW - Energy Transition KW - Hydrogen export KW - hydrogen underground storage KW - Hydrogen prices KW - salt caverns KW - Power-to-X KW - PyPSA-Earth Y1 - 2026 U6 - https://doi.org/10.2139/ssrn.6307406 PB - SSRN ER - TY - INPR A1 - Falter, Thomas T1 - Werkstattbericht: Erfahrungen aus der Zusammenarbeit mit KI-Agenten für individualisiertes Lernen N2 - DDer Artikel reflektiert praxisbasiert und philosophisch die Zusammenarbeit von Menschen mit KI-Agenten – insbesondere im Kontext des Lernens von Skills auf Experten-Level im Hochschulbereich. Am Beispiel der Entwicklung des KI-unterstützten Multiagentensystem LASSI, das Lernen unterstützt, wird eine Lernreise in drei Phasen erzählt – vom produktiven Staunen über die Verschiebung von Autorenschaft bis hin zur Übernahme von Verantwortung. Die Perspektiven von Entwickler, Professor und Studierenden zeigen, wie sich Mensch-Agenten-Systeme durch Zusammenarbeit verändern. Ergänzend werden drei philosophische Dimensionen – Embodied Knowledge, Enhanced Technologies und Embedded Ethics – herangezogen, um Potenziale und Risiken dieser Systeme zu beleuchten: Was passiert, wenn Intelligenz entkörperlicht, menschliche Fähigkeiten an Technologien ausgelagert und Werte kodiert werden? Der Text plädiert dafür, Agenten nicht als neutrale Werkzeuge, sondern als Mitgestalter im Soziotechnischen System Bildung zu begreifen. In der Schlussbetrachtung wird die Hochschule als Denkwerkstatt skizziert, die durch KI ihre Rolle neu definieren muss: weg vom Prüfungsbetrieb hin zu einem Raum für Urteilskraft und Selbstreflexion. Lernen mit und von Agenten heißt: Lernen neu denken – gemeinsam gestalten. Y1 - 2026 U6 - https://doi.org/10.13140/RG.2.2.20204.63360 ER - TY - JOUR A1 - Brunner, Philipp A1 - Vogl, Stefanie T1 - Extracting Product Improvement Insights from Social Media Comments Using Machine Learning: a Case Study in the Automotive Industry JF - Machine Learning and Knowledge Extraction N2 - This paper presents a scalable machine learning pipeline for extracting actionable, product-related insights from user-generated social media comments. Leveraging sentence embeddings from SBERT and unsupervised clustering (k-Means and agglomerative), the approach structures informal and noisy comments from Instagram and YouTube into topic groups intended to support thematic analysis. A case study on feedback regarding BMW vehicles, comprising more than 26,000 comments, illustrates how the pipeline can reveal recurring user concerns, such as design critiques, usability issues, and technology-related expectations, even in short and unstructured social media comments. The proposed pipeline operates without labeled data or manual annotation, enabling scalable application and transferability across product categories and industries. By transforming large-scale, unstructured consumer feedback into interpretable themes, the pipeline provides product teams with an efficient and structured basis for data-driven product development and improvement. KW - social media mining; sentence embeddings; unsupervised clustering; product feedback analysis; SBERT; natural language processing Y1 - 2026 U6 - https://doi.org/10.3390/make8020042 VL - 8 IS - 2 PB - MDPI ER - TY - JOUR A1 - Wallner, M. A1 - Gutbrod, Max A1 - Rauber, David A1 - Ebigbo, Alanna A1 - Probst, Andreas A1 - Palm, Christoph A1 - Messmann, Helmut A1 - Roser, David T1 - KI-gestützte Detektion und Segmentierung von Magenkarzinomen in westlichen endoskopischen Bilddaten anhand eines fine-tuned Deep-Learning Ansatzes JF - Zeitschrift für Gastroenterologie N2 - Diese vorläufige monozentrische Studie zeigt, dass ein aus einem Barrett-Ösophagus-KI-System feinjustiertes Deep-Learning-Modell Magenkarzinome in westlichen multimodalen endoskopischen Bilddaten zuverlässig detektieren und präzise segmentieren kann. Die hohe Segmentierungsgenauigkeit und Detektionssensitivität über verschiedene Bildmodalitäten hinweg unterstreichen die Machbarkeit eines pathologiegestützten KI-Ansatzes auch in einer westlichen Niedriginzidenzpopulation. Aufgrund der ausschließlichen Verwendung von Bildern mit sichtbaren Tumoren lassen sich keine Aussagen zur Spezifität treffen; eine Übertragbarkeit auf Screening- oder Mischkollektive ist daher limitiert. Weitere Studien mit a) größerem Datensatz inklusive Videodaten, b) externer Validierung an einer multizentrischen westlichen Kohorte, sowie c) Anwendung und Prüfung an nicht-neoplastischen Vergleichsbildern oder anderen Pathologien sind erforderlich. Nach unserem Kenntnisstand zählt dieses System zu den ersten in einer westlichen Population entwickelten endoskopischen KI-Ansätzen zur Magenkarzinomdetektion, und zu wenigen, die vollständige ESD-präparatbasierte Referenzdaten für Training und Validierung nutzen. Y1 - 2026 U6 - https://doi.org/10.1055/s-0046-1817751 VL - 64 IS - 03 SP - e64 EP - e65 PB - Thieme ER - TY - JOUR A1 - Tan, Jing Jie A1 - Schreiner, Rupert A1 - Hausladen, Matthias A1 - Asgharzade, Ali A1 - Edler, Simon A1 - Bartsch, Julian A1 - Bachmann, Michael A1 - Schels, Andreas A1 - Kwan, Ban-Hoe A1 - Ng, Danny Wee-Kiat A1 - Hum, Yan-Chai T1 - SiMiC: Context-aware silicon microstructure characterization using attention-based convolutional neural networks for field-emission tip analysis JF - Journal of Vacuum Science & Technology B N2 - Accurate characterization of silicon microstructures is essential for advancing microscale fabrication, quality control, and device performance. Traditional analysis using scanning electron microscopy (SEM) often requires labor-intensive, manual evaluation of feature geometry, limiting throughput and reproducibility. In this study, we propose SiMiC: Context-aware Silicon Microstructure Characterization Using Attention-based Convolutional Neural Networks for Field-Emission Tip Analysis. By leveraging deep learning, our approach efficiently extracts morphological features—such as size, shape, and apex curvature—from SEM images, significantly reducing human intervention while improving measurement consistency. A specialized dataset of silicon-based field-emitter tips was developed, and a customized convolutional neural network architecture incorporating attention mechanisms was trained for multiclass microstructure classification and dimensional prediction. Comparative analysis with classical image processing techniques demonstrates that SiMiC achieves high accuracy while maintaining interpretability. The proposed framework establishes a foundation for data-driven microstructure analysis directly linked to field-emission performance, opening avenues for correlating emitter geometry with emission behavior and guiding the design of optimized cold-cathode and SEM electron sources. The related dataset and algorithm repository that could serve as a baseline in this area can be found at https://research.jingjietan.com/?q=SIMIC. KW - Field emitter arrays KW - Quality assurance KW - Convolutional neural network KW - Deep learning KW - Image processing KW - Machine learning KW - Cold cathodes KW - Scanning electron microscopy KW - Electron sources KW - Chemical elements Y1 - 2025 U6 - https://doi.org/10.1116/6.0005068 VL - 43 IS - 6 PB - AVS ER -