Logo des Repositoriums

Auflistung nach Autor:in "Riedel, Till"

1 - 4 von 4
  • Textdokument
    Adaptives luftqualitätsgewichtetes Fahrradrouting mittels Land-use Regression auf Basis offener Daten
    (INFORMATIK 2021, 2021) Janßen, Julian; Tremper, Paul; Riedel, Till
    Luftschadstoffen ausgesetzt zu sein hat langfristige negative gesundheitliche Folgen, denen besonders Fahrradfahrer im urbanen Raum ausgesetzt sind. Dabei gibt es wahrscheinlich keine unschädliche Dosis: weniger ist immer besser. Diese Arbeit zeigt, dass luftqualitätsgewichtete Fahrradrouten die persönliche Exposition gemäß dem Regressionsmodell deutlich reduzieren können, wobei die errechneten Umwege zumeist nur minimal sind. Auf Basis offener Daten wird ein neuronales Netzwerk zur Schätzung der Luftqualität trainiert. Dabei werden PM10-Daten aus mobilen Messungen als Indikator der Luftqualität verwendet. Das entstehende Land-Use-Regression-Modell bezieht dabei sowohl zeitliche als auch räumliche Features mit ein. Anschließend wird dieses Modell verwendet, um luftqualitätsgewichtete Routen zu berechnen. Dabei wird gezeigt, wie ein solches feingranulare Modell im Routing verwendet werden kann. Anhand von zufällig gewählten Start/Ziel Paaren werden die luftqualitätsgewichteten Routen mit der jeweils kürzesten Strecke verglichen.
  • Konferenzbeitrag
    Ensuring a Robust Multimodal Conversational User Interface During Maintenance Work
    (Mensch und Computer 2021 - Tagungsband, 2021) Fleiner, Christian; Riedel, Till; Beigl, Michael; Ruoff, Marcel
    It has been shown that the provision of a conversational user interface proves beneficial in many domains. But, there are still many challenges when applied in production areas, e.g. as part of a virtual assistant to support workers in knowledge-intensive maintenance work. Regarding input modalities, touchscreens are failure-prone in wet environments and the quality of voice recognition is negatively affected by ambient noise. Augmenting a symmetric textand voice-based user interface with gestural input poses a good solution to provide both efficiency and a robust communication. This paper contributes to this research area by providing results on the application of appropriate head and one-hand gestures during maintenance work. We conducted an elicitation study with 20 participants and present a gesture set as its outcome. To facilitate the gesture development and integration for application designers, a classification model for head gestures and one for one-hand gestures were developed. Additionally, a proof-of-concept for operators’ acceptance regarding a multimodal conversational user interface with support of gestural input during maintenance work was demonstrated. It encompasses two usability testings with 18 participants in different realistic, but controlled settings: notebook repair (SUS: 82.1) and cutter head maintenance (SUS: 82.7).
  • Konferenzbeitrag
    PictographAI: Interactive Generation of Stylized Pictographs for Presentations
    (Mensch und Computer 2024 - Workshopband, 2024) Makarem, Sarah; Röddiger, Tobias; Riedel, Till; Beigl, Michael
    In today’s data-driven world, effective data visualization is crucial for communication. Recent studies have shown that meaningful and relevant visual embellishments and decorations can significantly enhance data visualization memorability and comprehension. Hence, we introduce PictographAI, a generative tool integrated into presentation software to transform traditional bar charts into Pictographic visualizations. Utilizing a multimodal AI pipeline, PictographAI processes text, images, and raw data from presentation slides to automatically generate contextually appropriate pictographs. Our pipeline uses an Large language model agent, a text-guided image-inpainting model, and algorithmic post-processing to make sense of the slide contents and generate pictographs. As users update their presentation slides, the AI pipeline automates the generation of new pictographs that represent the respective contents. In this work, we demonstrate the concept and working principle that motivates the system architecture and the generative AI pipeline on a bar chart generation use case that integrates into a presentation slide creation workflow.
  • Konferenzbeitrag
    Towards Fine-Grained Sensor-Based Probabilistic Individual Air Pollution Exposure Prediction using Wind Information
    (EnviroInfo 2023, 2023) Tremper, Paul; Riedel, Till
    The estimation of pollutant exposure is highly dependent on the spatial and temporal resolution of the underlying model. This work presents a street-level Gaussian Process Regression model for urban air quality that uses a novel covariance kernel based on physical considerations to process wind information. This model can be driven by information from observations from low-cost sensor networks. We present the model, including the construction of the wind angle kernel, and discuss the inconclusive evaluation to date, the current challenges, and the way forward.