Hochschulschriftenserver der TH Wildau
Not a member yet
1283 research outputs found
Sort by
User-driven prioritization of ethical principles for artificial intelligence systems
No full textDespite the progress of Artificial Intelligence (AI) and its contribution to the advancement of human society, the prioritization of ethical principles from the viewpoint of its users has not yet received much attention and empirical investigations. This is important to develop appropriate safeguards and increase the acceptance of AI-mediated technologies among all members of society. In this research, we collected, integrated, and prioritized ethical principles for AI systems with respect to their relevance in different real-life application scenarios. First, an overview of ethical principles for AI was systematically derived from various academic and non-academic sources. Our results clearly show that transparency, justice and fairness, non-maleficence, responsibility, and privacy are most frequently mentioned in this corpus of documents. Next, an empirical survey to systematically identify users’ priorities was designed and conducted in the context of selected scenarios: AI-mediated recruitment (human resources), predictive policing, autonomous vehicles, and hospital robots. We anticipate that the resulting ranking can serve as a valuable basis for formulating requirements for AI-mediated solutions and creating AI algorithms that prioritize user’s needs. Our target audience includes everyone who will be affected by AI systems, e.g., policy makers, algorithm developers, and system managers as our ranking clearly depicts user’s awareness regarding AI ethics
Bericht Forschung und Transfer 2023
Full text linkProjekte und Publikationen der Technischen Hochschule Wildau aus dem Jahr 2023
Investigation of the relationship between overtaking distances of motor vehicles to cyclists and bicycle infrastucture characteristics using OpenBikeSensor and OpenStreetMap data from Berlin and Brandenburg
No full textSowohl das Bundesverkehrsministerium als auch das Verkehrsministerium Brandenburgs haben sich das Ziel gesetzt, den Anteil des Radverkehrs von 2017 bis 2030 zu verdoppeln. Ein wichtiger Faktor bei der Wahl des Fahrrads als Verkehrsmittel ist die subjektive Sicherheit. Diese wird stark beeinflusst durch Überholvorgänge von Kraftfahrzeugen. Seit 2020 ist in der Straßenverkehrs-Ordnung ein Mindestabstand von 1,50 m festgesetzt, den Kraftfahrzeuge innerorts beim Überholen von Radfahrenden einhalten müssen. In Forschungsarbeiten zum Überholverhalten zeigte sich, dass dieser Abstand häufig unterschritten wird und dass die Infrastruktur einen Einfluss auf den Überholabstand hat. Mit dem Überholabstandsmessgerät „OpenBikeSensor“ wurden in den letzten Jahren über 270.000 Überholvorgänge erfasst. Um die Daten auszuwerten zu können, müssen die Informationen zur Infrastruktur am Ort des Überholvorgangs aus einer anderen Quelle gewonnen werden. Ziel dieser Arbeit war herauszufinden, inwieweit die ebenfalls frei verfügbaren Daten aus der OpenStreetMap dafür geeignet sind. Dafür wurde einerseits eine Recherche im OpenStreetMap-Wiki durchgeführt, in welchem dokumentiert ist, wie Objekte zu erfassen sind. Andererseits wurden die für Berlin und Brandenburg verfügbaren Daten aus OpenBikeSensoren und der OpenStreetMap miteinander verknüpft und ausgewertet. Um die Güte von Auswertungsergebnissen einschätzen zu können, wurden sechs Analysen aus anderen Forschungsarbeiten nachvollzogen. Es zeigte sich, dass erreichbare Ergebnisse eine gute Vergleichbarkeit mit den Ergebnissen anderer Forschungen aufweisen, wenn eine hohe Anzahl von Überholvorgängen und eine detaillierte Beschreibung der untersuchten Infrastruktur vorliegen, und dass die kombinierten Daten weitreichende Analysen ermöglichen können. Die wesentliche Einschränkung liegt darin, dass für viele Straßen noch keine ausreichende Beschreibung in der OpenStreetMap verfügbar ist. Eine Tendenz zur Zunahme dieser Informationen war erkennbar.Both the German and the Brandenburg Ministry of Transport have set themselves the goal of doubling the share of bicycle traffic from 2017 to 2030. An important factor in choosing the bicycle as mode of transport is the perceived safety, which is highly affected by being overtaken by motorised vehicles. Since 2020, the German road traffic regulations have stipulated a lateral passing distance of a minimum of 1.50 m that motorists must maintain when overtaking cyclists. Studies showed that this is often not adhered to, and that infrastructure design has an impact on the lateral passing distance. With OpenBikeSensors – a device to measure lateral passing distances – more than 270,000 overtaking events have been measured in the last few years. In order to analyse this data, information on the infrastructure at the place of the overtaking event has to be obtained from an additional source. The goal of this study was to find out to what extent the likewise freely available OpenStreetMap-Data is suitable for this. On the one hand this was attained by conducting research on the OpenStreetMap-Wiki which is the place where the rules for mapping are documented. On the other hand, the available OpenStreetMap- and OpenBikeSensor-Data for Berlin and Brandenburg were linked and analysed. In order to assess the quality of possible research outcomes six analysis from other studies were reconstructed. It was found that the results achieved show a good accordance with the results of other studies if a high number of events could be used, and that the combination of the two data sets can allow a wide range of analysis. The usability of the OpenStreetMap, though, is limited by the fact that many roads are still lacking detailed information on some aspects of the infrastructure – a tendency to an increase in information density could be detected
Industrie 5.0 Lernumgebung am Beispiel der Wildauer Smart Production
No full textDie Industrie 5.0 fordert neue Lernansätze und zeitgleich auch passende Lernumgebungen. Parallel müssen diese neben den didaktischen Herausforderungen auch den Transfer- und Übertragungsgedanken auf die industriellen Anwendungen gerecht werden. Durch die täglich steigende Anzahl vielfältiger KI-Tools insbesondere textgenerierenden Tools, braucht es Systeme mit einem breiten Anwendungsbereich. Im Rahmen des vorliegenden Beitrags geben die Autoren einen Einblick in die Wildauer Smart Production, welche den transdisziplinären Gedanken von Lern- und Transferumgebungen Rechnung trägt, Möglichkeiten der Gestaltung komplexer Produktionssysteme widerspiegelt, die Integration menschzentrierter Ansätze ermöglicht und als Forschungsumgebung eingesetzt wird
Proof of Principle of Wastewater Treatment using Plasma Discharge to Reduce the Amount of Methylparaben
No full textSynthetic substances like many pharmaceuticals, preservatives or other chemical compounds are actually very difficult to handle in sewage treatment. These compounds are very stable in aqueous solution and their degradation reactions are insufficient. Therefore, to eliminate these substances from wastewater additional afford is necessary. Extreme conditions like pH value, redox potential, chemical or physical energy need to be present. With our study we try to show that the use of plasma discharge could be a solution to this problem. Using the example of methylparaben, a preservative, we could show, that the physical energy of plasma discharge is able to initialize the degradation reaction in aqueous environment. The concentration was reduced by up to 70 percent in our setting depending on the treatment duration. Overall, the system showed potential to optimize wastewater treatment. Further examinations are necessary for example regarding undesirable by-products
Efficient data acquisition for traceability and analytics
No full textImplementing processes for traceability is required in various industries to assure product quality during manufacturing, provide evidence on required processing conditions or facilitate product recalls. Commonly, radio-frequency identification (RFID) or code recognition techniques (e.g. Data Matrix) are applied to track the flow of workpieces through a manufacturing system and link processing data accordingly. Although the analysis of tracking data is well-examined, we still see a gap in the research on the trade-off between data acquisition, data analytics and data quality. Here, we present a framework to increase the value of existing data by enabling data analytics while addressing common pitfalls and reducing the costs of data management
Zum Themenschwerpunkt „Fehlerkultur in Bibliotheken“
No full textEditorial zum Themenschwerpunkt „Fehlerkultur in Bibliotheken“ aus Bibliothek Forschung und Praxis, 48(1)
Ein modernes Zentrum für Bildung und Kultur
No full textLange Zeit galt der historische Stadtkern San Salvadors – Hauptstadt des kleinen mittelamerikanischen Landes El Salvador – als hektisch bis gefährlich, und damit insgesamt als wenig attraktiv zum Schlendern und Durchführen von Gästen und Touristen. Hemdsärmelig schaffte der medienwirksame Präsident Nayib Bukele nicht nur eine Aufbruchstimmung insgesamt, sondern gleichfalls einen Neuanfang, unter anderem mit dem Neubau einer Nationalbibliothek. Er entschied sich bewusst für die Investition in eine solche leuchtturmartige Informationseinrichtung. Ausschließlich von China finanziert, verwandelte sich anfängliche Skepsis in große Begeisterung. Die Menschen nehmen das Angebot in Scharen rund um die Uhr an. Sie wollen Zugang erhalten in eine Welt, die es vordem in Mittelamerika nicht gegeben hat und ihresgleichen sucht. Diese öffentliche Bibliothek weiß mit ihrer Architektur, dem Service und der Medienausrichtung zu begeistern, was anzuhalten scheint, blickt man auf den nicht abreißenden Zuspruch, den der library turn nach sich zieht
Juridification and regulative failures. The complicated implementation of international law into national schools
No full textThis paper starts with the increasing discussions on juridification in education. Concerning theorizing on such processes, we examine the poor implementation of the UN Convention on the Rights of Persons with Disabilities Convention on the Rights of Persons with Disabilities, CRPD (2008) in the school sector of Germany. The paper considers the reasons for this failed endeavor by analyzing the complex, multilevel relations between policy and law in different national and historical contexts. With this, aspects of juridification as a process in education policy can be illuminated. In this regard, we suggest crucial aspects regarding juridification in public education: a focus on regulative failures in juridification processes, juridification’s contexts, the mandate for putting it into action, the allocation of resources, and finally, its objective and subjective rights dimensions, i.e. how an individual can claim rights within the machinery of public education
Comparison of CNN-Based Architectures for Detection of Different Object Classes
No full text(1) Background: Detecting people and technical objects in various situations, such as natural disasters and warfare, is critical to search and rescue operations and the safety of civilians. A fast and accurate detection of people and equipment can significantly increase the effectiveness of search and rescue missions and provide timely assistance to people. Computer vision and deep learning technologies play a key role in detecting the required objects due to their ability to analyze big volumes of visual data in real-time. (2) Methods: The performance of the neural networks such as You Only Look Once (YOLO) v4-v8, Faster R-CNN, Single Shot MultiBox Detector (SSD), and EfficientDet has been analyzed using COCO2017, SARD, SeaDronesSee, and VisDrone2019 datasets. The main metrics for comparison were mAP, Precision, Recall, F1-Score, and the ability of the neural network to work in real-time. (3) Results: The most important metrics for evaluating the efficiency and performance of models for a given task are accuracy (mAP), F1-Score, and processing speed (FPS). These metrics allow us to evaluate both the accuracy of object recognition and the ability to use the models in real-world environments where high processing speed is important. (4) Conclusion: Although different neural networks perform better on certain types of metrics, YOLO outperforms them on all metrics, showing the best results of mAP-0.88, F1-0.88, and FPS-48, so the focus was on these models