Publikationsserver der Hochschule München
Not a member yet
547 research outputs found
Sort by
Reliable position measurement of the ball on a computer-controlled football table
Full text linkIm Rahmen dieser Masterarbeit wurde die ursprünglich 2009 entwickelte Ballerkennung eines computergesteuerten Tischkickers durch einen neuen Ansatz ersetzt. Dazu wurde eine geeignete Farbkamera mitsamt Objektiv sowie eine LED-Beleuchtung ausgewählt, beschafft und montiert. Kamera und Beleuchtung wurden durch eine dafür entwickelte Ansteuerungsplatine verbunden. Diese ermöglicht eine Synchronisierung mit der Belichtungszeit der Kamera und gleichzeitig eine davon unabhängige Helligkeitseinstellung.
Anschließend wurde ein konfigurierbares und modular aufgebautes Programm entwickelt, das die zuverlässige Verfolgung der Ballposition auf dem Spielfeld ermöglicht. Die Erkennung nutzt dabei eine Maskierung störender Bildbereiche, einen Farb- und Größenfilter sowie eine Kreiserkennung, um die Position des Balls zu ermitteln. Eine Optimierung wichtiger Funktionsparameter erfolgte mit der Brute-Force-Methode mithilfe von händisch bestimmten Positionen und Radien des Balls im Bild. Zudem wurden für das Projekt grafische Benutzeroberflächen entwickelt, die die Konfiguration erleichtern.As part of this master's thesis, a ball detection system originally developed in 2009 for a computer-controlled football table was replaced with a new approach. A suitable color camera, lens, and LED lighting were selected, procured, and installed. The camera and lighting got connected by a specially developed control board. This enables synchronization with the camera's exposure time and, at the same time, an independent brightness setting.
A modular and configurable program was developed to enable reliable tracking of the position of the ball on the playfield. The detection uses masking of interfering areas, a color filter, a size filter, as well as circle detection to determine the position of the ball. Important function parameters were optimized using the brute force method using manually determined positions and radii of the ball in the image. Additional graphical user interfaces were developed for the project to make configuration easier
Energy costs of communicating with AI
No full textThis study presents a comprehensive evaluation of the environmental cost of large language models (LLMs) by analyzing their performance, token usage, and CO2 equivalent emissions across 14 LLMs ranging from 7 to 72 billion parameters. Each LLM was tasked with answering 500 multiple-choice and 500 free-response questions from the MMLU benchmark, covering five diverse subjects. Emissions were measured using the Perun framework on an NVIDIA A100 GPU and converted through an emission factor of 480 gCO2/kWh. Our results reveal strong correlations between LLM size, reasoning behavior, token generation, and emissions. While larger and reasoning-enabled models achieve higher accuracy, up to 84.9%, they also incur substantially higher emissions, driven largely by increased token output. Subject-level analysis further shows that symbolic and abstract domains such as Abstract Algebra consistently demand more computation and yield lower accuracy. These findings highlight the trade-offs between accuracy and sustainability, emphasizing the need for more efficient reasoning strategies in future LLM developments
Radikalisierung von Kindern und Jugendlichen im digitalen Zeitalter. TikTok als Bühne für rechtspopulistische Propaganda : Eine medienpädagogische Analyse aus kindheitspädagogischer Betrachtung
No full textWie trägt die Plattform TikTok zur Verbreitung rechtspopulistischer Propaganda bei und welche medienpädagogischen Ansätze können Kinder und Jugendliche vor politischer Radikalisierung im digitalen Zeitalter schützen
Do Teaching Media Matter? A Comparative Study of Finance Education via Classroom, Livestream, Video, and Educational Games
No full textThis study examines how different instructional media—face-to-face classes, live streaming, pre-recorded videos, and educational games—affect student learning outcomes in finance education. A sample of first-year economics students was assessed on their knowledge of basic financial principles before being randomly assigned to five groups. Four groups attended the same finance course delivered through different media formats, while a fifth group served as a control and received no instruction. After the course, all students completed a second (post-course) assessment. By comparing individual pre- and post-test results, as well as learning gains across the groups, we evaluated the effectiveness of each delivery method. The results show that all four instructional formats significantly improved financial knowledge compared to the control group. Among the media types, educational games proved to be an effective and reliable tool for delivering finance content. However, the differences in learning gains between face-to-face instruction, live streaming, and pre-recorded videos were not statistically significant. These findings indicate that a range of delivery models can be used effectively in finance education. The study contributes to current debates on cost-effective teaching strategies and supports evidence-based decisions on curriculum design in digitally transformed higher education environments after COVID-19
A Digital Leadership Approach in Response to the Challenges of the World of Work 4.0 with a Focus on Social Sustainability
No full textThis thesis examines how digital leadership in Work 4.0 enhances employee well-being as part of social sustainability, addressing the research question: How can digital leadership practices foster well-being in digitally connected, hybrid work environments? Aligned with the United Nations’ Sustainable Development Goals (SDG 3: health and well-being, and SDG 8: decent work), this study underscores leadership's strategic role in promoting inclusive and sustainable workplaces. This study underscores leadership's strategic role in promoting inclusive and sustainable workplaces. Employing a qualitative approach, insights were gathered from leaders and employees across diverse industries, providing a comprehensive view of digital leadership practices. Findings reveal that psychological stability, emotional intelligence, adaptability, and inclusivity are pivotal for promoting well-being and resilience. Effective leadership styles prioritize empathy, flexibility, and open communication, mitigating challenges like emotional disconnect and team cohesion in remote settings. Practical recommendations include leadership training programs, hybrid work models to balance responsibilities, and frameworks like Key Performance Indicators (KPIs) to measure inclusivity and well-being. Despite its limited scope, the research highlights the need for future studies exploring emerging technologies' impact on digital leadership across diverse sectors
Symbolic Tools for low order Linear Fractional Transformation implemented in Matlab/MuPad
No full textIn robust control the linear fractional transformation (LFT) is a usual way to represent systems/models with uncertainties and there exist many tools for analysis and control design based on this system representation. Usually, it is straightforward to transform a system with uncertainties into LFT-form. However, there exists no unique solution and using simple approaches will result in models of very high order/complexity. Applying analysis and design methods to these high order models involves large computation times and also the accuracy of the results can be worse or even unusable. For systems with parametric uncertainties a symbolic approach using special preprocessing techniques like factorization, tree decomposition, variable splitting has shown to be to most e_ective way to end up with models of low order. For this purpose an implementation of these symbolic tools is presented based of the Matlab/MuPad symbolic engine and its applicability is shown for parametric uncertain systems, yielding LFT-models of almost minimal order
Entwicklung eines Tools zur Evaluierung der Nachhaltigkeit von Partnerunternehmen der HM Business School
Full text linkThis thesis adopts an evidence-based approach to develop a transparent and comparable sustainability assessment tool for HM Business School‘s partner companies. A mixed-method research methodology was chosen that first deductively analyzes theoretical principles from established tools for measuring sustainability. The most suitable methods were evaluated through multi-criteria analysis to develop a tailored tool for HM Business School. The research was then supplemented by an inductive evaluation of this prototype with five participants from partner companies. The test subjects applied the adapted self-assessment, followed by an interview about their user experience and the results obtained. The prototype test focused on user-friendliness, comprehensibility of questions and results, and relevance for the respective companies. It became evident that the strengths included the simple usability and opportunity for self-reflection, while the lack of industry customization and the subjectivity of the assessments were identified as weaknesses. The developed prototype showed feasibility in ist current version and could contribute to increased knowledge of sustainable business practices in the context of educational partnerships. However, adjustments would be relevant for targeted use, including the weighting of sustainability criteria, transparency of the calculation basis, and the possibility of considering benchmarks for industry-specific comparisons
Equal opportunities of Health Development in Children and Adolescents in Munich: a comparative social analysis of Neuhausen–Nymphenburg (District 09) and Milbertshofen–Am Hart (District 11)
Full text linkChancengerechte Gesundheitsentwicklung bei Kindern und Jugendlichen ist ein vielschichtiges und gesellschaftlich hochrelevantes Thema. Der Zusammenhang sozialer und gesundheitlicher Ungleichheit spielt dabei eine erhebliche Rolle. Der aktuelle wissenschaftliche und gesundheitspolitische Stand wird in diesem Bezug erläutert. Anhand einer vergleichenden Sozialraumanalyse von Neuhausen – Nymphenburg (Stadtbezirk 09) und Milbertshofen - Am Hart (Stadtbezirk 11) in München werden auf die Gesundheit Einfluss nehmende Variablen bestimmt und miteinander verglichen. Neben sozialen Strukturen der Gesundheitsversorgung, sozialräumlichen und umweltbedingten Einflussfaktoren auf die Gesundheit wurden Aktivitäten zur Prävention und Gesundheitsförderung und die Jugendhilfe berücksichtigt. Die vorliegende Sozialraumanalyse wurde mit sekundären Daten erstellt. Die sekundäre Datenanalyse erfolgte durch Gewinnung von Daten durch Veröffentlichungen und Angaben des statistischen Amts der Landeshauptstadt München, verschiedener Referate der Landeshauptstadt München, der gesetzlichen Krankenversicherungen sowie der Kassenärztlichen Vereinigung. Die Daten wurden durch Zusammentragung und kartografische Darstellungen miteinander verglichen und analysiert. Es gibt deutliche Unterschiede in der nahräumlichen Gesundheitsversorgung beider Bezirke. Alle analysierten Gesundheitsakteure sind in Bezirk 09 häufiger als in Bezirk 11 niedergelassen bzw. tätig. In Bezirk 11 besteht anhand der Sozialstrukturanalyse zudem vermutlich ein erhöhter Bedarf an gesundheitlicher Versorgung. Sozialräumliche und umweltbedingte Faktoren zeigen in beiden Bezirken Unterschiede auf. Ob diese eine messbare nachteilige oder vorteilhafte Bedeutung für die Gesundheitsentwicklung haben, muss weiterführend untersucht werden. Aktivitäten im Rahmen von Prävention und Gesundheitsförderung konnten kleinräumlich nicht erfasst werden. Da die Versorgung mit Akteuren aus dem Gesundheitswesen in dem ohnehin schon sozial schwächeren Sozialraum (Bezirk 11) wesentlich schlechter ist, haben die dort aufwachsenden Kinder und Jugendlichen geringere Chancen in ihrer Gesundheitsentwicklung, als die Kinder, die in Bezirk 09 leben. Der Zusammenhang sozialer und gesundheitlicher Ungleichheit wird in der Planung von Versorgungsstrukturen offensichtlich nicht oder nur unzureichend berücksichtigt. Die kleinräumliche Analyse von Versorgungsstrukturen in einzelnen Sozialräumen gestaltet sich ohne eine zielführende Kooperation mit der Kassenärztlichen Vereinigung sehr aufwändig und ist ohne Zusammenarbeit zudem fehleranfällig. Die Gewinnung einiger relevanter Daten zum Thema Prävention und Gesundheitsförderung bedeutet eine Zusammenarbeit mit den gesetzlichen Krankenversicherungen, denn hierzu gibt es bisher kaum vergleichbare Fakten und keine messbare Zielsetzung. Ein politisches Gesamtkonzept zur Verringerung sozialer und gesundheitlicher Ungleichheit ist bisher noch nicht erkennbar.Equitable health development in children and adolescents is a complex and socially highly relevant topic. The relationship between social and health inequality plays a significant role in this context. The current state in science and public health is explained in this regard. Based on a comparative social space analysis of Neuhausen-Nymphenburg (Municipality 09) and Milbertshofen-Am Hart (Municipality 11) in Munich,variables influencing health are determined and compared. In addition to social structures, health care as well as socio-spatial and environmental influencing factors on health, activities for prevention, health promotion and youth welfare are considered. The present social space analysis was created with secondary data. The secondary data analysis was carried out by obtaining data through publications and information from the statistical office of the City of Munich, various departments of the City of Munich, the statutory health insurance providers and the Kassenärztliche Vereinigung (Association of Statutory Health Insurance Physicians). The data were compared and analyzed through collation and cartographic representations. There are marked differences in the health care of both districts at the local level. All analyzed actors in health care are more frequently resident or operating in District 09 than in District 11. According to the social health analysis, there is likely to be an increased need for health care in District 11. Socio-spatial and environmental factors are different in both districts. It needs to be further examined whether these have a measurable detrimental or beneficial significance for health development. Activities in the context of prevention and health promotion could not be recorded at the local level. Given that health care is significantly less well provided for in the anyway socially weaker social space (Municipality 11), the children and adolescents growing up there have a lower chance of developing their health than children and adolescents living in Municipality 09. The relationship between social and health inequality is obviously not or only insufficiently considered in the planning of supply structures. The local analysis of supply structures in individual social spaces is very complex without an expedient cooperation with the Association of Statutory Health Insurance Physicians and, in addition, error-prone without its collaboration. Obtaining some relevant data on the topic of prevention and health promotion means cooperating with the statutory health insurance providers because there are hardly any comparable facts and measurable objectives. An overall policy approach to reducing social and health inequality is yet to be developed
Navigating the space between serverful and serverless : benefits, drawbacks, challenges and mitigations
Full text linkThis work explores the complexities of navigating the space between serverful and serverless for cloud-native applications. Initial deployment target choices involve requirements and assumptions, which can impact development experience, performance, and cost, as they change or no longer hold. As applications and requirements evolve, practitioners may want to reconsider deployment targets to meet performance and cost-efficiency needs. If left unaddressed, these can prevent the application from meeting its objectives, resulting in a deployment target shift that may require costly migrations if not facilitated beforehand. This research explores strategies to improve flexibility, allowing practitioners to decide about deployment later and shift to deployment targets as needed. By synthesizing various sources, this work presents a comprehensive definition of serverless and forms a new perspective, thereby broadening its applicability. Through literature review, interviews, technical exploration and performance benchmarks, two approaches to hybrid deployment are recommended: The adapted serverful-first approach leverages the adaptation of established frameworks for deployment targets, while the portable serverless-first approach focuses on serverless benefits and portability for seamless transitions across the serverful-serverless spectrum. The evaluation confirms that both approaches provide practical solutions for hybrid deployments, but practitioners need to consider factors such as traffic, usage patterns, cost and latency to select the most suitable deployment target. However, the limitations of the approaches, such as adaptation overhead and applicability constraints, still require careful consideration. Overall, this work provides a foundation for hybrid deployment strategies, guiding practitioners in navigating and optimizing the vast landscape of serverful and serverless deployment targets