Refine
Institute
- Fakultät Informatik (597)
- Fakultät Elektro- und Informationstechnik (463)
- THI Business School (412)
Version
- published (858)
Data from production environments is now available in unprecedented volumes, making the problem-solving of incidents through root cause analysis straightforward. However, the root cause analysis process remains time-consuming. This study employs the Kitchenham standard systematic literature review methodology to explore how information models and deep learning can streamline this process. By conducting a comprehensive search across four major databases, we evaluate the current technological advancements and their application in root cause analysis. The aim of this study is to assesses the impact of information models for root cause analysis in a production environment. Our findings reveal that integrating knowledge graphs, association rule mining, and deep learning algorithms significantly improves the speed and depth of root cause analysis compared to traditional methods. Specifically, the use of neural networks in recent literature shows substantial advancements in analyzing complex datasets, facilitating large-scale data integration, and enabling automated learning capabilities. Comparing our findings with other recent studies highlights the advantages of using information modeling and deep learning technologies in root cause analysis. This comparison underscores the superior accuracy and efficiency of these advanced methodologies over traditional manual interpretation methods. The effective implementation of these technologies requires a robust foundation of clean, standardized data, giving rise to the concept of “Production IT.” Furthermore, it is crucial for this data to be openly available to facilitate academic research, thereby enabling the development of new methods for more efficient and effective root cause analysis.
The wireless communication technology has gained significant attention in the transportation industry over the recent years. Cellular Vehicle-to-Everything (V2X) communication facilitates the information exchange among road users (such as vehicles, pedestrians etc.) and the infrastructure with an intention to improve the overall road safety, driving comfort, traffic efficiency and save energy. Advanced use-cases aim towards enhancing key functionalities of vehicle automation by means of sensor data sharing and cooperative maneuver & trajectory planning. The introduction of the PC5 interface for sidelink (SL) communication within the mobile communication systems, supports direct exchange of messages between users, independent of the cellular network infrastructure.
Two types of radio resource allocation modes are supported in Cellular V2X: managed mode and the unmanaged mode. In the managed mode, a User Equipment (UE) remains connected to the cellular network and the process of resource allocation is coordinated by the base station. In the unmanaged mode, a UE selects its radio resources from a pre-configured resource pool without any assistance from the base station. Originally both these modes were developed by considering that the vehicles exchange periodic messages which are safety-critical in nature. The existing rule based radio resource allocation algorithms in both the modes are unable to adapt their selection parameters in the events of aperiodic data traffic patterns resulting from the diverse generation rules of different V2X messaging protocols.
We begin this PhD thesis by carrying out system level network simulations within the developed framework Artery-C, where we study the metrics and parameters that influence the performance of the rule-based radio resource allocation in the sidelink modes. In the first step, we derive the baseline conditions where each mode performs to its best efficiency. By varying the generation rules of the messaging protocols, we further analyze the behavior of the modes when V2X data traffic does not follow a specific pattern. Our studies have shown that both the modes suffer from frequent re-allocations because the messages are no longer periodic and the data sizes do not fit into the previously allocated radio resources. This results in poor utilization of the allocated resources. The unmanaged mode is particularly susceptible to radio resource collisions because the vehicles only have partial awareness about the resource selection decisions of other road traffic participants.
As a second contribution, we examine the criteria for sidelink mode selection and the possibilities for a mode switching operation within the sidelink modes and also between the sidelink and the cellular (Uu) modes. We have formulated the strategies for mode switching and calculate the latency in each phase of the mode switch procedure. Although the managed mode has shown advantages with regard to allocation and management of radio resources, it is to be noted that a vehicle cannot remain connected to a base station at all instants of time. Also, switching between different modes is not seamless considering the associated latencies in each phase. This leads us towards the goal of improving the efficiency of the allocation & scheduling of radio resources in the unmanaged mode.
After a careful review of the enhancements that can be implemented within the rule based algorithm in the unmanaged mode, it was found that there needs to be a mechanism where vehicles can continuously share their resource selection decisions, adapt their selection parameters and even re-evaluate them (if needed) within a grant period. Therefore, we investigated the Reinforcement Learning (RL) based Artificial Intelligence (AI) approaches that facilitate independent learning, adapting and decision making among spatially distributed vehicular agents. We have developed a fully decentralized multi agent networked Markovian Decision Process (MDP) model of the Cellular V2X communication network where each agent executes an AI based radio resource scheduler. By extending the actor-critic methodology of the RL, we have derived two variants - Independent Actor Critic (IAC) and Shared Experience Actor Critic (SEAC). The results of our evaluations have indicated that both these schedulers have a potential to achieve better radio resource utilization with a reduced risk of radio resource collisions among the agents. Subsequently, it brings about 15 − 20% improvement in the reliability of the communication link which we regard as a valuable contribution. To summarize, this PhD thesis investigates the performance of the rule based radio resource allocation algorithms in Cellular V2X and proposes the qualitative improvements that can be achieved by means of reinforcement learning.
Clustering and Anomaly Detection in Embedding Spaces for the Validation of Automotive Sensors
(2024)
This paper examines the relationship between corporate purpose and disaster response. To that aim, we use a unique dataset generated by Just Capital's COVID‐19 response tracker that characterizes the early responses of 928 companies included in the Russell 1000 Index generated by the COVID‐19 pandemic. We marry this data to the Business Roundtable Statement on the Purpose of a Corporation (BRT‐CPS) signatories list, which includes 210 firms that declared “a fundamental commitment to all stakeholders” as their corporate purpose. We estimate the differential impact of BRT‐CPS signees on the rest. After controlling our results with financial indicators and environmental, social, and governance (ESG) data, we convey that BRT‐CPS signees were more likely to develop strategies that benefited their surrounding communities and other stakeholders. Moreover, corporations with high ESG were less likely to have an early response that directly benefited stakeholders other than shareholders, except for layoffs and unpaid leave. Finally, we discuss the outcomes considering the symbolic management theory, corporate purpose literature, and its relation to stakeholder management.
A novel hybrid copper paste was developed for low temperature sintering of bare semiconductors. Cu(II) formate (Cu(for)) is complexed in amino-2-propanol (A2P) and added to a paste of etched brass micro flakes. A two-step sintering process is applied: The paste is printed and dried at 120 °C under formic acid (FA) enriched N2 atmosphere (FAN2) for 5 min. Afterwards, bare semiconductors are placed and sintered at 250 °C for 5 min applying a bonding pressure of 20 MPa/10 MPa. By the thermal decomposition of the Cu(for) atomic Cu is released and forms in-situ Cu-nanoparticles. An interconnect is realized with shear strength >100 MPa.
Invariant concept classes form the backbone of classification algorithms immune to specific data transformations, ensuring consistent predictions regardless of these alterations. However, this robustness can come at the cost of limited access to the original sample information, potentially impacting generalization performance. This study introduces an addition to these classes—the permutation-invariant linear classifiers. Distinguished by their structural characteristics, permutation-invariant linear classifiers are unaffected by permutations on feature vectors, a property not guaranteed by other non-constant linear classifiers. The study characterizes this new concept class, highlighting its constant capacity, independent of input dimensionality. In practical assessments using linear support vector machines, the permutation-invariant classifiers exhibit superior performance in permutation experiments on artificial datasets and real mutation profiles. Interestingly, they outperform general linear classifiers not only in permutation experiments but also in permutation-free settings, surpassing unconstrained counterparts. Additionally, findings from real mutation profiles support the significance of tumor mutational burden as a biomarker.
The integration of deep learning-based tools into diagnostic workflows is increasingly prevalent due to their efficiency and reproducibility in various settings. We investigated the utility of automated nuclear morphometry for assessing nuclear pleomorphism (NP), a criterion of malignancy in the current grading system in canine pulmonary carcinoma (cPC), and its prognostic implications. We developed a deep learning-based algorithm for evaluating NP (variation in size, i.e., anisokaryosis and/or shape) using a segmentation model. Its performance was evaluated on 46 cPC cases with comprehensive follow-up data regarding its accuracy in nuclear segmentation and its prognostic ability. Its assessment of NP was compared to manual morphometry and established prognostic tests (pathologists’ NP estimates (n = 11), mitotic count, histological grading, and TNM-stage). The standard deviation (SD) of the nuclear area, indicative of anisokaryosis, exhibited good discriminatory ability for tumor-specific survival, with an area under the curve (AUC) of 0.80 and a hazard ratio (HR) of 3.38. The algorithm achieved values comparable to manual morphometry. In contrast, the pathologists’ estimates of anisokaryosis resulted in HR values ranging from 0.86 to 34.8, with slight inter-observer reproducibility (k = 0.204). Other conventional tests had no significant prognostic value in our study cohort. Fully automated morphometry promises a time-efficient and reproducible assessment of NP with a high prognostic value. Further refinement of the algorithm, particularly to address undersegmentation, and application to a larger study population are required.
ÖPNV für Kinder
(2024)
Diese Bachelorarbeit befasst sich mit der nutzerzentrierten Konzeption hochautomatisierter Shuttlebusse im öffentlichen Personennahverkehr (ÖPNV) mit einem besonderen Fokus auf die Bedürfnisse von Kindern und deren Eltern. Ziel der Arbeit war es, herauszufinden, wie Shuttlebusse gestaltet werden müssen, um sie für Kinder optimal nutzbar zu machen und gleichzeitig das Vertrauen der Eltern in die Technologie zu gewinnen. Um die Perspektiven der zwei Zielgruppen besser nachvollziehen zu können, wurden Interviews mit Schulkindern und eine Onlineumfrage mit den Eltern durchgeführt. Basierend auf den Erkenntnissen wurden dann Prototypen entwickelt und in Nutzertests evaluiert. Die Ergebnisse der Studie mit Kindern zeigen, dass diese positiv und offen gegenüber der neuen Technologie eingestellt sind. Zudem lässt sich erkennen, dass den Kindern vor allem ein großes Medienangebot während der Fahrt wichtig ist. Die Studienergebnisse der Eltern haben gezeigt, dass durch erhöhte Level an Echtzeitinformationen, während der Fahrt der Kinder, die Zufriedenheit gesteigert wird. Zusätzlich dazu, wird der Trend erkennbar, dass durch die Bereitstellung vermehrter Informationen auch das Vertrauen in das System ansteigen kann. Die Arbeit liefert einige Empfehlungen dazu, wie die Bedürfnisse beider Zielgruppen miteinander vereint werden können.
This master thesis compares the environmental impacts of 4th-generation district heating (4GDH) and 5th-generation district heating and cooling (5GDHC) systems using different heat sources in Germany. The primary objective is to evaluate and compare the environmental performance of a geothermal-based 5GDHC system with a gas-combined heat and power (CHP) based 4GDH system.
The study follows DIN ISO 14040/14044 standards and employs the SimaPro software with the ecoinvent 3.10 database to perform the LCA. The analysis considers two scenarios: one using the German electricity grid and the other using photovoltaic (PV) panels as the electricity source. Key impact categories assessed include climate change, acidification, freshwater and marine eutrophication, human toxicity and resource depletion.
Results indicate that the 5GDHC system, despite its advanced technology, has higher overall environmental impacts compared to the 4GDH system in most categories, primarily due to the intensive resource use and higher greenhouse gas emissions during the drilling process and operational phases. However, 5GDHC shows lower impacts on human toxicity (cancer-related) and land use efficiency, highlighting areas where it performs better than 4GDH.
This comparative study underscores the need for further optimization and integration of renewable energy sources in 5GDHC systems to enhance their sustainability and reduce their environmental impacts. The findings provide valuable insights for policymakers and stakeholders in the energy sector aiming to achieve carbon neutrality and improve the environmental performance of district heating systems.
Keywords: Comparative life cycle assessment (LCA); Life cycle inventory (LCI); District heating system (DHS); 5th generation district heating and cooling system (5GDHC); 4th generation district heating system (4GDH); Geothermal energy; Combined heat and power (CHP); SimaPro; Ecoinvent
The presented work introduces a maturity model for evaluating Machine Learning implementations, with a primary focus on Production Planning and Control processes, as well as broader organizational and technical aspects in companies. This model emerges as a response to the research gap identified in the analysis of 14 existing maturity models, which served as foundational bases for the development of this novel approach. By examining success factors and obstacles at different maturity levels, categorized according to defined dimensions and overarching design fields, this model can serve as a catalyst for bridging the research gap between models demanded in practice and the scholary exploration of topics related to Machine Learning in corporate processes. Notably, the structured design of this maturity model ensures accessibility for small and medium sized enterprises (SMEs).
Lithium-Ion Batteries (LIBs) usually present several degradation processes, which include their complex Solid-Electrolyte Interphase (SEI) formation process, which can result in mechanical, thermal, and chemical failures. The SEI layer is a protective layer that forms on the anode surface. The SEI layer allows the movement of lithium ions while blocking electrons, which is necessary to prevent short circuits in the battery and ensure safe operation. However, the SEI formation mechanisms reduce battery capacity and power as they consume electrolyte species, resulting in irreversible material loss. Furthermore, it is important to understand the degradation reactions of the LIBs used in Electric Vehicles (EVs), aiming to establish the battery lifespan, predict and minimise material losses, and establish an adequate time for replacement. Moreover, LIBs applied in EVs suffer from two main categories of degradation, which are, specifically, calendar degradation and cycling degradation. There are several studies about battery degradation available in the literature, including different degradation phenomena, but the degradation mechanisms of large-format LIBs have rarely been investigated. Therefore, this review aims to present a systematic review of the existing literature about LIB degradation, providing insight into the complex parameters that affect battery degradation mechanisms. Furthermore, this review has investigated the influence of time, C-rate, depth of discharge, working voltage window, thermal and mechanical stresses, and side reactions in the degradation of LIBs.