05 Fakultät Informatik, Elektrotechnik und Informationstechnik
Permanent URI for this collectionhttps://elib.uni-stuttgart.de/handle/11682/6
Browse
113 results
Search Results
Item Open Access Anonymisierung von Daten : von der Literatur zum Automobilbereich(2023) Herkommer, JanDie Datenanonymisierung im Automobilbereich gewinnt immer mehr an Bedeutung. Jedoch gibt es kaum Literatur und Ansätze, die sich mit der Anonymisierung von Automobildaten beschäftigen. In dieser Arbeit werden deshalb mit Hilfe einer strukturierten Literaturrecherche die aktuell verbreitetsten Verfahren und Anwendungsbereiche erörtert und die wichtigsten Erkenntnisse der Recherche zusammengefasst. So werden bei den analysierten Paper der Anwendungsbereich, die Methodik sowie der zu anonymisierende Datentyp ermittelt. DesWeiteren werden die Metriken zum Vergleich von unterschiedlichen Ansätzen betrachtet. Mit Hilfe dieser Erkenntnisse wird im Anschluss auf die Anonymisierung von Fahrzeugdaten anhand verschiedener Anwendungsfälle eingegangen und Herausforderungen und Lösungsansätze skizziert. Zuletzt wird beispielhaft ein Ansatz zur Anonymisierung von Routen implementiert, um mit Hilfe eines GPS-Sensors aufgezeichnete Fahrzeugrouten zu anonymisieren. Dabei werden zusätzliche Probleme wie der Umgang mit Messungenauigkeiten und Messfehlern sowie die tatsächlichen Auswirkungen von reduzierter Datennutzbarkeit verdeutlicht.Item Open Access Development of an Euler-Lagrangian framework for point-particle tracking to enable efficient multiscale simulations of complex flows(2023) Kschidock, HelenaIn this work, we implement, test, and validate an Euler-Lagrangian point-particle tracking framework for the commercial aerodynamics and aeroacoustics simulation tool ultraFluidX, which is based on the Lattice Boltzmann Method and optimized for GPUs. Our framework successfully simulates one-way and two-way coupled particle-laden flows based on drag forces and gravitation. Trilinear interpolation is used for determining the fluid's macroscopic properties at the particle position. Object and domain boundary conditions are implemented using a planar surface approximation. The whole particle framework is run within three dedicated GPU kernels, and data is only copied back to the CPU upon output. We show validation for the velocity interpolation, gravitational acceleration, back-coupling forces and boundary conditions, and test runtimes and memory requirements. We also propose the next steps required to make the particle framework ready for use in engineering applications.Item Open Access Adaptive robust scheduling in wireless Time-Sensitive Networks (TSN)(2024) Egger, SimonThe correct operation of upper-layer services is unattainable in wireless Time-Sensitive Networks (TSN) if the schedule cannot provide formal reliability guarantees to each stream. Still, current TSN scheduling literature leaves reliability, let alone provable reliability, either poorly quantified or entirely unaddressed. This work aims to remedy this shortcoming by designing an adaptive mechanism to compute robust schedules. For static wireless channels, robust schedules enforce the streams' reliability requirements by allocating sufficiently large wireless transmission intervals and by isolating omission faults. While robustness against omission faults is conventionally achieved by strictly isolating each transmission, we show that controlled interleaving of wireless streams is crucial for finding eligible schedules. We adapt the Disjunctive Graph Model (DGM) from job-shop scheduling to design TSN-DGM as a metaheuristic scheduler that can schedule up to one hundred wireless streams with fifty cross-traffic streams in under five minutes. In comparison, we demonstrate that strict transmission isolation already prohibits scheduling a few wireless streams. For dynamic wireless channels, we introduce shuffle graphs as a linear-time adaptation strategy that converts reliability surpluses from improving wireless links into slack and reliability impairments from degrading wireless links into tardiness. While TSN-DGM is able to improve the adapted schedule considerably within ten seconds of reactive rescheduling, we justify that the reliability contracts between upper-layer services and the infrastructure provider should specify a worst-case channel degradation beyond which no punctuality guarantees can be made.Item Open Access An analytics framework for the IoT platform MBP(2020) Kumar, AbishekThe emergence of IoT has introduced a huge amount of applications that generate massive amounts of data at a high rate. This data stream needs intelligent data processing and analysis. The evolution of Smart cities and Smart industries has resulted into an ocean of data from millions of sensors and devices. Surveillance systems, telecommunication systems, smart devices, and smart cars are some examples of such systems. However, this data itself doesn’t provide any information unless it is analysed. This results into a need of analytics tools and frameworks which can efficiently analyse this data and provide with useful information. Analytics is all about inspection, transformation and modelling of data to achieve information that further suggests and assists in decision making. In a world of IoT, analytics has a crucial role to play to improve life and better manage the infrastructure in a secure, sustainable and cost effective manner. The smart sensor network serves as the base for IoT. In this context, one of the major tasks is to develop advanced analytics frameworks for the interpretation of data provided by the sensors. MBP is a platform for managing IoT environments. Sensors and devices can be registered to this platform and the status of sensors can be viewed and modified from the platform. This platform will be used to collect data from the sensors and devices connected to the platform. There are two types of mining that can be performed on raw data, one technique analyses the data on the fly as it is received (Data Stream Mining) and the other can be performed on demand on the data collected for a longer period of time (Batch Processing). Both types of analysis has its own advantages. Lambda architecture is a data analytics architecture which allows us to perform both stream analysis and batch processing on the same data. This architecture defines some practical and well versed principles of handling big data. The pattern allows us to deal with both real time and historical data, but the analysis is performed separately and does not affect each other. In this thesis, we will create an analytics framework for the MBP IoT platform based on the lambda architecture.Item Open Access Long-term motion prediction in traffic(2020) Hengel, KatharinaThe field of Inverse Reinforcement Learning (IRL) addresses the task of finding a cost function which describes expert behavior. Since the cost function is solely computed from expert demonstrations, the sample complexity exerts influence on the performance of these algorithms. In this thesis we study the Learning to Search (LEARCH) and the maximum entropy IRL framework as example IRL techniques. Based on these two algorithms we develop a variation of the LEARCH algorithm using the idea of maximum entropy IRL. In the next step we extend LEARCH to Deep-LEARCH as well as the newly developed LEARCH variation to a equivalent Deep-LEARCH variation. Thereby we generalize the cost function to function space using Convolutional Neural Networks (CNNs). Including maximum entropy inside the Deep-LEARCH variation increases the density of the target maps of the CNN. We discover that LEARCH shows the lowest sample complexity among the investigated algorithms, while maximum entropy shows the highest sample complexity. In the deep learning setting the increased density of the CNN target maps did not improve the performance. Hence, the performance does not change, if the algorithms are extended by CNNs to the function space.Item Open Access Entwicklung von fairen und personalisierten Machine Learning Modellen(2020) Lässig, NicoDie Nutzung von Machine Learning Modellen zur Vorhersage und Unterstützung von Entscheidungen spielt in der heutigen digitalen Zeit eine zentrale Rolle. Dabei werden die Machine Learning Modelle mithilfe von bereits existierenden Datensätzen trainiert. Ein Problem davon ist, dass einige dieser Datensätze diskriminierend sind. Das bedeutet, dass zum Beispiel Personen aufgrund ihres Geschlechts eine ungleiche Chance auf ein hohes Gehalt haben. Daraus resultiert, dass die trainierten Modelle in diesem Fall die Abhängigkeit vom Geschlecht zu dem Gehalt erlernen. Es gibt bereits einige Ansätze, welche versuchen diese Unfairness auszugleichen. Die bereits existierenden Ansätze berücksichtigen dabei jedoch lediglich, dass diese Ergebnisse über den kompletten Datensatz gesehen fair sind und berücksichtigen somit nur die globale Fairness. In dieser Masterarbeit werden zwei Algorithmen (NFD-ME und PFD-ME Algorithmus) mit jeweils zwei Varianten entwickelt und implementiert, welche die lokale Fairness von einzelnen Datenpunkten betrachten. Lokale Fairness ist die Fairness in lokalen Regionen eines Punktes. Zu der lokalen Region eines Punktes zählen andere Datenpunkte, welche sich in den Attributen stark ähneln. Unsere Ansätze gewährleisten eine bestmögliche Optimierung der lokalen Fairness durch die dynamische Auswahl der lokal am besten geeigneten Modelle. Dafür werden existierende Algorithmen aus den Bereichen der fairen Modell-Ensembles, sowie der dynamischen Modell-Ensembles, kombiniert. Diese werden ergänzt durch mehrere Metriken zur Messung der Fairness. Außerdem sind diese Algorithmen nicht auf ein binäres sensitives Attribut beschränkt, sodass diese Algorithmen vielseitig anwendbar sind. Im Anschluss werden unsere Algorithmen und ihre Varianten in der Evaluation mit bereits existierenden Algorithmen aus den Bereichen der fairen Modell-Ensembles, sowie dynamischen Modell-Ensembles, verglichen.Item Open Access Entwurf und Implementierung eines Stream-basierten, dynamischen und fairen Scheduling-Verfahrens für WiFi-Netzwerkverkehr(2023) Herrmann, JonaHeutzutage sind Wireless Local Area Network (WLAN)-Netzwerke, basierend auf dem IEEE~802.11 Standard, für mobile Endgeräte besonders wichtig, da diese nur so Zugang zum Internet bekommen können. Dementsprechend gibt es solche Netzwerke fast überall wie beispielsweise in öffentlichen Verkehrsmitteln, in Hotels oder auch bei der Arbeit. Doch die Qualität und Geschwindigkeit einer Verbindung kann dabei stark variieren. Um dies zu verbessern, wird zuerst eine Analyse dieser Netzwerke vorgenommen, um die Stelle des Bottlenecks zu identifizieren. Dabei konnte gezeigt werden, dass der Bottleneck an zwei Stellen vorliegen könnte. Das wäre einmal der Uplink zum Internet und andererseits der entsprechende Downlink, wobei sich der Bottleneck am Downlink letztendlich beim Internet-Provider befindet. Mit einem neu entwickelten Stream-basierten, dynamischen und fairen Scheduling-Verfahren soll die Qualität und Geschwindigkeit im WLAN-Netzwerk verbessert werden. Dafür wird eine neue Art von Fairness definiert, sodass die Pakete von Endgeräten mit einem geringen Datenverbrauch eine höhere Priorität erhalten. Dadurch bekommen letztendlich Endgeräte, welche gutmütig sind, eine bessere Antwortzeit als diese, die eine große Datenmenge übertragen. Zum Schluss wird die Linux-Implementierung des Scheduling-Verfahrens noch unter verschiedenen Metriken evaluiert. Dabei konnte gezeigt werden, dass die gewünschte Art von Fairness damit realisiert werden kann. Dies wurde sowohl unter Laborbedingungen als auch mit realen Anwendungen des Internets erfolgreich evaluiert. Außerdem wurde durch Messungen gezeigt, dass es im Punkt Performanz keinen signifikanten Unterschied zu dem Standard Scheduling-Verfahren in Linux gibt.Item Open Access Representation learning of scene images for task and motion planning(2020) Nguyen, Son TungThis thesis investigates two different methods to learn a state representation from only image observations for task and motion planning (TAMP) problems. Our first method integrates a multimodal learning formulation to optimize an autoencoder not only on a regular image reconstruction but also jointly on a natural language processing (NLP) task. Therefore, a discrete, spatially meaningful latent representation is obtained that enables effective autonomous planning for sequential decisionmaking problems only using visual sensory data. We integrate our method into a full planning framework and verify its feasibility on the classic blocks world domain [26]. Our experiments show that using auxiliary linguistic data leads to better representations, thus improves planning capability. However, since the representation is not interpretable, learning an accurate action model is extremely challenging, rendering the method still inapplicable to TAMP problems. Therefore, to address the necessity of learning an explainable representation, we present a self-supervised learning method to learn scene graphs that represent objects (“red box”) and their spatial relationships (“yellow cylinder on red box”). Such a scene graph representation provides spatial relations in the form of symbolic logical predicates, thus eliminates the need of pre-defining these symbolic rules. Finally, we unify the proposed representation with a non-linear optimization method for robot motion planning and verify its feasibility on the classic blocks-world domain. Our proposed framework successfully finds the sequence of actions and enables the robot to execute feasible motion plans to realize the given tasks.Item Open Access Learning task-parameterized Riemannian motion policies(2021) Le, An T.Nowadays, robots gradually have more autonomy to operate alongside people not only on assembly lines, but also in daily spaces such as kitchens, museums, or hospitals. In these scenarios, a robot must demonstrate a high degree of adaptability in realtime dynamic situations while satisfying task compliance objectives such as collision avoidance. The robot skill also needs to be programmed with ease to cope with an enormous variety of task behaviors. To achieve this, we propose Task-parameterized Riemannian Motion Policy (TP-RMP) framework to address the challenges associated with learning and reproducing the skills under multiple task objectives and situations. Specifically, the task objectives are viewed as multiple subtasks, learned as stable policies from demonstrations. The learned policies are also task conditioned and able to cope with real-time changing task situations. Under the RMPflow framework, our approach synthesizes a stable global policy in the configuration space that combines the behaviors of these learned subtasks. The resulting global policy is a weighted combination of the learned policies satisfying the robot’s kinematic and environmental constraints. Finally, we demonstrate the benchmarks of TP-RMP under increasing task difficulties in terms of external disturbances and skill extrapolation outside of the demonstration region.Item Open Access Roaming with deterministic real-time guarantees in wireless Time-Sensitive Networking(2023) Haug, LucasAs Cyber-Physical Systems (CPSs) become increasingly popular in various domains such as Industrial Internet of Things (IIoT) and autonomous vehicles, the demand for deterministic real-time communication with a high reliability and bounded network delay and delay variance (jitter) has grown. In CPSs with networked sensors and actuators, these deterministic real-time bounds are often crucial in order to provide safety guarantees. Major standardization organizations like the Institute of Electrical and Electronics Engineers (IEEE) have acknowledged the necessity for deterministic networks, resulting in a set of standards known as Time-Sensitive Networking (TSN), which enable deterministic communication in wired Ethernet networks. Many CPSs, however, necessitate mobility and therefore rely on wirelessly connected devices, such as a worker wearing an exoskeleton in order to carry heavy loads. To this end, the TSN standards will also be part of the upcoming Wi-Fi 7 standard (IEEE 802.11be). Although the increased flexibility offered by the mobility of devices in these scenarios is advantageous, it presents new challenges, such as controlling access to the shared wireless transmission medium and managing handovers such that deterministic real-time guarantees are maintained. In this work, we investigate these challenges and provide novel approaches to improve the reliability, delay and jitter in wireless TSN. We first modify an already existing Integer Linear Program (ILP) to generate schedules for the Time-Aware Shaping (TAS) in a wireless network environment. The necessity for this arises because the wireless transmission medium utilizes a shared access method, whereas existing approaches are limited to point-to-point Ethernet connections. Furthermore, we provide a novel seamless handover approach for wireless TSN utilizing two wireless interfaces in a single device and extend it with a proactive handover approach in order to allow for smoother handovers with a greater reliability. In order to analyze our approaches, we extend the INET framework of the OMNeT++ simulator with an implementation of our approaches. Our evaluation shows that delay and jitter are mainly influenced by the random back-off algorithm of the channel access procedure in Wi-Fi indicating research topics for future work. Moreover, we were able to significantly improve the reliability for wireless TSN by employing our proactive handover approach in the simulation.