HFTor - HfT Open Repository (Hochschule für Technik Stuttgart)
Not a member yet
404 research outputs found
Sort by
Enhancing Short Answer Grading With OpenAI APIs
No full textAutomated short-answer grading can accelerate and standardize the assessment of tests in higher education. The topic has received a significant boost due to the rapid development of powerful LLM models in recent years. We examine the performance of the OpenAI models GPT-3.5 and GPT-4o on the CSSAG dataset and demonstrate that GPT-4o, in particular, achieves an accuracy that falls within the range observed in assessments by different human evaluators. We pay special attention to cases where there are significant deviations from the reference assessment. Additionally, we discuss the practical implications
Kompakte Hofhäuser : Forschung für nachhaltige Urbanität hoher Dichte
Get PDFDas sog. „kompakte Hofhaus“ wurde als neuartiger Hofhaustyp an der HFT Stuttgart entwickelt und untersucht. Zu Quartieren addiert stellt er den Bewohnern bei geringem Landverbrauch urbane Lebensräume zur Verfügung und kann hohe Dichte, Nutzungsdiversität sowie ein hohes Maß an Privatheit und Wohnqualität in Innen- wie Außenräumen miteinander vereinigen, in Neubau- wie in Bestandsituationen. Hinsichtlich der Ziele knüpft diese Forschung damit unmittelbar an die Stuttgarter Weißenhofsiedlung von 1927 an
Innovation in tune: An empirical investigation of user acceptance of artificial intelligence-generated music
No full textThis study investigates public perceptions and acceptance of artificial intelligence (AI)-generated singing, addressing a critical gap in technology acceptance research. Grounded in anthropomorphism theory—the attribution of human characteristics to non-human entities— this research explores how human-like qualities of AI singing voices influence empathy and acceptance. Utilizing a quantitative research design with 310 participants, this study employs partial least squares structural equation modeling (PLS-SEM) to analyze the effects of anthropomorphic factors, such as animacy, humanlike fit, and perceived sociability, on the likeability of AI-generated singing. The findings indicate that animacy and humanlike fit significantly enhance likeability, which, in turn, influences the intention to use AI-generated singing. Curiosity and word of mouth (WoM) emerged as key drivers of acceptance. This study builds on the traditional technology acceptance models (TAM, UTAUT2) by proposing a framework that integrates human-related factors, thereby demonstrating the necessity of adapting existing models to better evaluate AI systems in creative domains. These insights contribute to advancing the understanding of human-computer interaction, particularly within the evolving landscape of AI-driven creative processes
Habitat selection of the northern lapwing (Vanellus vanellus) during breeding season in Europe.
Get PDFThe ongoing population reduction of northern lapwings across Europe has raised serious conservation concerns. One of the primary drivers of this trend is the continuous transformation of their habitats. Land-use changes, particularly the expansion of agricultural fields and intensively managed pastures, have altered the species’ breeding environments, significantly affecting reproductive success. Eggs and chicks are especially vulnerable, facing increased risks due to habitat modifications.
To investigate how landscape structure influences habitat selection, this study, conducted in collaboration with Helmholtz Centre for Environmental Research - UFZ, analyzed tracking data from 13 northern lapwings alongside land use and land cover and elevation data. A spatial complexity assessment was performed using landscape metrics, which were then integrated into step-selection function (SSF) models to quantify habitat preferences.
Findings from the SSF models indicate that, under the tested scenarios, habitats characterized by homogeneous resource diversity and spatial configuration increase the relative probability of habitat selection. These results provide key insights into how northern lapwings interact with their environment, offering a foundation for designing more effective conservation strategies tailored to their ecological needs
Performing Classical and Post-Quantum Cryptography on IoT Data: An Evaluation
No full textAbstract. The increasing number of Internet of Things (IoT) devices requires the implementation of reliable and efficient cryptographic solutions to ensure the security of data transmission and storage. Traditional cryptographic algorithms can be challenging to implement without compromising performance or security, particularly due to the significant resource constraints faced by many IoT devices. This paper aims to evaluate the performance of classical and post-quantum cryptographic libraries on IoT hardware, specifically using the PyCryptodome Python library for traditional cryptography and the Kyber 512 algorithm for post-quantum encryption. The study assesses the security features, computational efficiency, and feasibility of these libraries in resource-constrained environments. Experiments were conducted on the Raspberry Pi 5 platform to analyze encryption and decryption performance, providing insights into the trade-offs between security and processing overhead. The findings will help identify the best cryptographic solutions for IoT devices based on their performance, security, and hardware limitations. This study includes a practical implementation of the Kyber 512 algorithm on the Raspberry Pi platform, exploring the feasibility of deploying quantum-resistant cryptography in IoT environments using Python. The research examines cryptographic performance in constrained environments by measuring encryption and decryption times scientifically. The results offer valuable insights into the computational overhead and feasibility of implementing post-quantum cryptographic solutions on limited-resource hardware. Scientific graphs were created to illustrate the complexity of encryption and decryption times, providing a visual representation of performance variations in various scenarios. Understanding these graphs clarifies the trade-offs between security and computational efficiency, which are crucial for real-world IoT applications. Furthermore, the study discusses potential optimizations and future directions that could enhance the efficiency of cryptographic operations in IoT hardware. In summary, this paper contributes to the study of cryptographic applications for IoT by investigating the viability of post-quantum cryptographic algorithms in constrained environments. By considering the findings, IoT security can be strengthened against emerging cyber threats through the selection of optimized cryptographic libraries that balance security, performance, and hardware limitations
Integration of EnergyPlus into the Simulation Workflow Engine SimStadt
Get PDFUrban Building Energy Modelling (UBEM) is a key tool for sustainable urban planning, enabling cities to estimate energy demand and plan for efficiency and resilience. SimStadt is one such platform, capable of simulating thousands of buildings using the DIN V 18599 monthly balance method. While efficient, this approach cannot capture hourly dynamics, short-term cooling peaks, or the influence of control strategies.
This thesis addresses these limitations by embedding EnergyPlus, a detailed physics-based simulation engine, directly into the SimStadt workflow. A CityGML to IDF pipeline was developed to transform building geometry, usage and material data into valid EnergyPlus input files, fully integrated into SimStadt’s interface and reporting system.
Two case studies validated the solution. A single building test confirmed correct handling of geometry and schedules. A larger case study with 91 buildings in Mainz demonstrated scalability: EnergyPlus predicted approximately 32% lower aggregate heating demand than the monthly balance method, reflecting the effects of hourly solar gains and control dynamics. Residential buildings showed strong agreement (median error ~6%, 64% within ±30%), while outliers such as garages and greenhouses
highlighted categories where monthly methods are less representative.
The integration enables new insights, including hourly load profiles, peak demand analysis and cooling demand assessment. Performance tests showed runtimes of about three seconds per building, with potential for parallel execution at city scale. The work demonstrates that detailed, dynamic simulations can be embedded into a scalable UBEM platform, extending SimStadt’s capabilities and aligning it with international research trends
Nachhaltige Quartiersentwicklung – Optimierungschancen städtebaulicher Planungsprozesse
Get PDFDie Bandbreite an Themen als auch an zu beteiligenden Fachleuten ist durch die Fülle an Aufgaben inzwischen so groß, dass es eines neuen Ansatzes und Verständnisses für Planungsverfahren bedarf, um die umfassenden Aspekte in der Umstrukturierung von Quartieren transparent und nachvollziehbar in möglichst schlanken und effektiven Prozessen zu einem möglichen Optimum an CO2 Neutralität zu bringen.
Dieser Leitfaden richtet sich an kommunale Akteur:innen der Stadtentwicklung, insbesondere an Mitarbeitende in kommunalen Verwaltungen und in Planungsbüros. Er unterstützt dabei, Planungsprozesse so zu strukturieren, dass nachhaltige Quartiere nicht nur als Ziel formuliert, sondern auch wirksam umgesetzt werden können. So soll der Leitfaden eine Orientierung bieten, um Planungsprozesse effizienter und nachhaltiger zu gestalten – für lebenswerte, sozial ausgewogene und klimafreundliche Quartiere.
Der Planungsleitfaden ist im Rahmen des Forschungsprojektes „CircularGreenSimCity – Ganzheitliche, ressourceneffiziente Betrachtung von Stadtquartieren“ entstanden. Das Projekt wurde vom Bundesministerium für Wirtschaft und Energie (BMWE) im Rahmen der Förderinitiative „Energieeffiziente Stadt“ (EnEff:Stadt) unter den Förderkennzeichen 03EN3050A-E gefördert. Als Projektträger fungierte der Projektträger Jülich (PtJ)
Mobilität an Hochschulen nachhaltig gestalten
No full textMit den beiden Forschungsprojekten HFTmobil und HFTmobil2.0 beschreitet die HFT Stuttgart neue Wege in der Mobilitätsforschung. Anhand einer ganzheitlichen Befragung wurde das Mobilitätsverhalten am Campus erfasst. Die erarbeiteten Konzepte beziehen alle Hochschulangehörigen mit ein und schaffen nachhaltige Lösungen für den Weg zum Campus. Ergänzend werden die Ergebnisse dieser Mobilitätsforschung in die Stadtgesellschaft getragen und in gemeinsamen Projekten mit internationalen Partnern länderübergreifende Betrachtungen der Herausforderungen der Verkehrs- und Mobilitätswende beleuchtet
Schulungen im Industrial Metaverse
No full textMetaversum-basierte Trainingsprogramme bieten eine realitätsnahe und risikofreie Lernumgebung, die besonders im industriellen Kontext wertvolle Anwendungen findet – etwa in immersiven Schulungen und der Simulation von Arbeitsabläufen. Trotz erheblicher Potenziale bestehen Herausforderungen wie Datenschutz, technologische Akzeptanz und Integration in bestehende Systeme. Zur Untersuchung der Relevanz von Schulungen im Metaverse wurden vier Experteninterviews durchgeführt. Um die wissenschaftliche Strenge zu gewährleisten, wurde bei allen Interviews derselbe Fragebogen eingesetzt. Dies ermöglichte es, die Antworten zu analysieren und vergleichbare, verlässliche Daten zu erhalten
A Practical Evaluation of LiaScript as a Collaborative Teaching Technology
No full textCreating and maintaining teaching and learning materials in a collaborative way between several lecturers is a constant challenge. At the same time, the materials must support the learning process of the students as well as different teaching scenarios. In this paper, we evaluate the LiaScript system as a collaborative teaching technology in the context of a CS1 lecture, offering a practical perspective. We show how to use LiaScript, highlight its benefits, and address the challenges we encountered during its use