05 Fakultät Informatik, Elektrotechnik und Informationstechnik
Permanent URI for this collectionhttps://elib.uni-stuttgart.de/handle/11682/6
Browse
Search Results
Item Open Access Automating deployment and testing in distributed networks of Electronic Control Units(2025) Vintonyak, RomanIncreasing complexity in automotive software has made manual testing and deployment in distributed networks of Electronic Control Units (ECUs) both time-consuming and error-prone. This thesis explores and implements a framework automating deployment and testing in distributed networks of ECUs. The proposed solution combines Ansible-based deployment with Gherkin-style test case descriptions, integrated into a CI/CD pipeline to enable consistent and repeatable testing automation. The resulting prototype called Automated Deployment and Testing of ECUs (ADATE), automates the testing of software components with automated deployment across simulated devices in a distributed environment. The framework demonstrates how automation in a distributed environment can make the testing process of ECUs both more efficient and reduce manual effort, offering a foundation for future adaptations in real-world automotive environments.Item Open Access Untersuchung der Zellausdehnung und des Gasinnendrucks zylindrischer Lithium-Ionen Zellen(2025) Hemmerling, Jessica; Birke, Kai Peter (Prof. Dr.-Ing)Item Open Access On-body visualization with extended reality : exploration and application(2025) Yu, Xingyao; Sedlmair, Michael (Prof. Dr.)Extended Reality (XR), which encompasses Virtual Reality (VR), Augmented Reality (AR) and Mixed Reality (MR), has significant potential across numerous fields by offering immersive and interactive experiences. This thesis addresses the increasing demand for effective on-body visualization technologies, particularly in the areas of biomechanical visualization, motion guidance, and feedback systems. The motivation behind this research is to leverage XR to enhance the presentation and understanding of biomechanical data and to improve performance in personal training scenarios, such as physiotherapy and exercise workouts. The thesis begins by exploring design choices for visualizing biomechanical data directly on the human body using AR, even in situations constrained by limited resources or time. It focuses on enhancing the interactivity and user comprehension of biomechanical visualizations in various everyday contexts. Through an investigation of different design options, the research identifies effective methods for presenting complex biomechanical information in an intuitive and interactive manner, making it accessible to a wide range of users, from patients undergoing physiotherapy to athletes in training. Expanding on broader applications such as physiotherapy and workouts, the research then investigates general design principles for presenting upper limb motion guidance using XR. This includes examining the impact of different perspectives, visual encoding techniques, and motion features on user performance and comprehension. The insights gathered lay a foundation for developing systems that provide clear, actionable feedback to support users in performing exercises with accuracy and efficiency. Building on these studies, the thesis applies its findings to practical deadlift training, investigating user performance and preferences for guidance visualization in this context. This study bridges the gap in previous general design implications for XR-based motion guidance systems, revealing considerations for XR systems in physically intensive exercises. Finally, the thesis presents a comprehensive review and empirical analysis of the design space for visual feedforward and corrective feedback mechanisms in XR environments. By addressing practical limitations and proposing solutions for improved system implementation, this part of the research offers a detailed framework to guide future developments in XR-based motion guidance systems. Overall, this thesis provides a thorough examination of XR-based on-body visualization for biomechanical data presentation and 3D motion guidance. It makes substantial contributions to the field, setting the stage for future advancements in XR technology for on-body visualization. The findings have broad implications for building applications in biomechanical visualization, designing effective motion guidance systems, and improving future XR-related applications and system designs.Item Open Access Assessing the use of pre-attentive visual variables for micro visualization while in motion(2025) Amzir, Amira YasminSmartwatches rely on glanceable visualizations to effectively convey information within limited screen space. This thesis aims to investigates the use of three pre-attentive visual variables - color, area, and motion - to highlight data points in common smartwatch graphs. We conducted a user study with 48 participants, evaluating four chart types: bar chart, line chart, linear progress chart, and radial progress chart. To assess whether pre-attentive processing can be leveraged for micro-visualization, we set stimulus exposure to 250 ms and examined perception under sedentary and walking conditions to determine effectiveness in mobile scenarios. Our results show that all three visual variables can be perceived with high accuracy, though performance varies depending on chart complexity. While simpler visualizations support effective highlighting, accuracy declines in more detailed time-series charts. Among the visual variables, motion and color yield the best perception rates, while area proves to be the least effective. These findings contribute to the design of efficient smartwatch visualizations that support rapid and intuitive data recognition.Item Open Access Implementing a Cholesky decomposition using SYCL(2025) Bloch, MichalPLSSVM, an LS-SVM implementation, now only uses the Conjugate Gradient algorithm for solving a set of linear equations. However, for an ill-conditioned matrix, it especially gets into trouble, as the converged solution drifts away from the actual solution due to rounding errors. Therefore, this thesis implements a different solver, e.g., the Cholesky Decomposition, which will be implemented in SYCL. We will implement multiple variations of the Cholesky Decomposition algorithm, including a blocked version, and utilize many different features of SYCL. The focus will primarily be on the fastest implementations. In the end, the fastest implementation will be integrated into PLSSVM alongside a Forward and Backward Substitution implementation for solving the set of linear equations. We will conclude with a runtime comparison between the implementations, a comparison of our best Cholesky Decomposition with the Conjugate Gradient using a dataset and a small discussion about numerical errors.Item Open Access All-inorganic CsPbI2Br perovskite solar cells with thermal stability at 250 °C and moisture-resilience via polymeric protection layers(2025) Roy, Rajarshi; Byranvand, Mahdi Malekshahi; Zohdi, Mohamed Reza; Magorian Friedlmeier, Theresa; Das, Chittaranjan; Hempel, Wolfram; Zuo, Weiwei; Kedia, Mayank; Rendon, Jose Jeronimo; Boehringer, Stephan; Hailegnanw, Bekele; Vorochta, Michael; Mehl, Sascha; Rai, Monika; Kulkarni, Ashish; Mathur, Sanjay; Saliba, MichaelAll-inorganic perovskites, such as CsPbI2Br, have emerged as promising compositions due to their enhanced thermal stability. However, they face significant challenges due to their susceptibility to humidity. In this work, CsPbI2Br perovskite is treated with poly(3-hexylthiophen-2,5-diyl) (P3HT) during the crystallization resulting in significant stability improvements against thermal, moisture and steady-state operation stressors. The perovskite solar cell retains ∼90% of the initial efficiency under relative humidity (RH) at ∼60% for 30 min, which is among the most stable all-inorganic perovskite devices to date under such harsh conditions. Furthermore, the P3HT treatment ensures high thermal stress tolerance at 250 °C for over 5 h. In addition to the stability enhancements, the champion P3HT-treated device shows a higher power conversion efficiency (PCE) of 13.5% compared to 12.7% (reference) with the stabilized power output (SPO) for 300 s. In addition, the P3HT-protected perovskite layer in ambient conditions shows ∼75% of the initial efficiency compared to the unprotected devices with ∼28% of their initial efficiency after 7 days of shelf life.Item Open Access Stochastic query synthesis for neural PDE solvers(2025) Ullrich, FinnPDEs are highly influential in physics and are describing various phenomena in the world, from wave movement to electro-magnetics. The problem arises when one tries to solve them, which requires enormous computing power for a numerical solution. To overcome the limitiations, neural PDE solvers have been proposed, using neural networks to approximate the solution trajectories. However, neural PDE solvers require training data from an computationally expensive numerical solver. Therefore, Musekamp et al. created a benchmark, which investigates active learning for neural PDE solvers. Active learning can reduce the amount of data required, while keeping the same performance. In this work, we will demonstrate a new strategy of selecting samples called stochastic query synthesis. Following this, we will remove the pool currently used and rather create a Markov chain directly from the input space containing unlabeled instances. The transition probability is based on the unadjusted Langevin algorithm, allowing us to sample by exploiting gradient information. To retrieve a better result, instead of just one chain, we will create multiple parallel chains, and only take the last state as input. We will show that this approach is equally effective as the currently implemented pool-based implementation. However, there are still performance problems that need to be solved in the future, to make it viable in practice.Item Open Access Mitigating the amorphization of perovskite layers by using atomic layer deposition of alumina(2025) Kedia, Mayank; Das, Chittaranjan; Kot, Malgorzata; Yalcinkaya, Yenal; Zuo, Weiwei; Tabah Tanko, Kenedy; Matvija, Peter; Ezquer, Mikel; Cornago, Iñaki; Hempel, Wolfram; Kauffmann, Florian; Plate, Paul; Lira-Cantu, Monica; Weber, Stefan A. L.; Saliba, MichaelAtomic layer deposition of aluminum oxide (ALD-Al2O3) layers has recently been studied for stabilizing perovskite solar cells (PSCs) against environmental stressors, such as humidity and oxygen. In addition, the ALD-Al2O3 layer acts as a protective barrier, mitigating pernicious halide ion migration from the perovskite towards the hole transport interface. However, its effectiveness in preventing the infiltration of ions and additives from the hole-transport layer into perovskites remains insufficiently understood. Herein, we demonstrate the deposition of a compact ultrathin (∼0.75 nm) ALD-Al2O3 layer that conformally coats the morphology of a triple-cation perovskite layer. This promotes an effective contact of the hole transporter layer on top of the perovskite, thereby improving the charge carrier collection between these two layers. Upon systematically investigating the layer-by-layer structure of the PSC, we discovered that ALD-Al2O3 also acts as a diffusion barrier for the degraded species from the adjacent transport layer into the perovskite. In addition to these protective considerations, ALD-Al2O3 impedes the transition of crystalline perovskites to an undesired amorphous phase. Consequently, the dual functionality (i.e., enhanced contact and diffusion barrier) of the ALD-Al2O3 protection enhanced the device performance from 19.1% to 20.5%, while retaining 98% of its initial performance compared to <10% for pristine devices after 1500 h of outdoor testing under ambient conditions. Finally, this study deepens our understanding of the mechanism of ALD-Al2O3 as a two-way diffusion barrier, highlighting the multifaceted role of buffer layers in interfacial engineering for the long-term stability of PSCs.Item Open Access Machine learning-based metabolic rate estimation from wearable sensors(2025) Olschewski, MarieAdaptive devices such as exoskeletons and prostheses can enhance human physical capabilities or replace the functionality of missing body parts. However, adjusting these devices for the specific needs of an individual remains a time-consuming and costly procedure. A key objective in optimizing these devices is minimizing the user’s energy expenditure (EE), a metric closely related to metabolic cost. Traditional methods for estimating metabolic cost, such as indirect calorimetry, are performed in controlled environments, limiting real-world applicability. This study aims to bridge this gap by exploring the use of traditional machine learning (ML) methods to estimate metabolic cost in real-time environments, utilizing wearable sensors integrated into adaptive devices. Using the dataset from Ingraham et al. (2019), which includes data from ten healthy subjects performing various exercises, the study investigates how different sensor combinations impact prediction accuracy. This thesis evaluated multiple ML models, including Random Forest (RF), Support Vector Machines (SVM), Linear Regression (LR), Decision Trees (DT), and Multilayer Perceptrons (MLP), within two cross-validation methods: Leave-One-Subject-Out (LOSO) and Leave-One-Time-Out (LOTO). Key findings from this evaluation include: In the LOSO setting, RF outperformed other models, achieving the lowest RMSE in several sensor regions, including Hexoskin, EMG Pants, and Best Combination, with the ’Best Combination’ region showing the best results. In contrast, MLP performed well in the LOTO setting, with its strongest performance observed in the ’Best Combination’ region. SVM demonstrated robust performance when all sensor data was combined, emphasizing the potential of multimodal sensor fusion. Hyperparameter tuning and sensor feature selection were crucial factors in optimizing model performance, particularly for more complex models like RF and MLP. The results suggest that while traditional ML methods can estimate EE effectively, challenges remain in refining preprocessing techniques, tuning hyperparameters, and optimizing sensor combinations. This thesis outlines the importance of model selection, sensor fusion, and parameter optimization in developing more accurate and real-time energy expenditure prediction systems for wearable technologies.Item Open Access Operationalization of automated fault description mappings in quality management systems using the natural language processing neural networks of the BERT language model(2025) Thakre, UrvashiThe automotive industry has witnessed remarkable technological advancements, leading to innovative approaches in quality management and fault categorization. This thesis focuses on improving the fault categorization process within Mercedes-Benz Operations 360 (MO360) platform by leveraging state-of-the-art transformer models like Bidirectional Encoder Representations from Transformers (BERT) and General Text Embedding (GTE) Large. Through advanced Natural Language Processing (NLP) techniques, the study addresses the challenges of mapping textual feedback to structured fault codes which is a crucial need in automotive production and customer service. The research begins with a thorough exploration of existing literature, highlighting the role of pre-trained language models in customer feedback analysis, fault diagnosis, and semantic text matching. Fine-tuned models are then developed and evaluated, achieving a training accuracy of more than 98%. Their performance is assessed using various metrics, and expert validation gives useful information about the system’s usability, accuracy, and usefulness in real life. These evaluations highlighted the system’s ability to simplify fault categorization while identifying areas for refinement, such as managing rare fault scenarios and ambiguous descriptions. An additional validation process, involving expert-labeled data, further enhanced the models’ accuracy, achieving a top-5 accuracy of 54% and a top-1 accuracy of 30.54%. Looking ahead, the thesis proposes future improvements, including multilingual support, real-time fault categorization, and explainable Artificial Intelligence (AI) features to increase transparency and trust. These findings demonstrate how NLP can significantly enhance quality management in the automotive sector, offering a robust and scalable framework that meets the industry’s evolving needs for precision and customer-focused solutions.