Research Interests

Research

Grasp Types

• Leveraging Publicly Available Textual Object Descriptions for Anthropomorphic Robotic Grasp Predictions [Dataset]
• A Study on the Influence of Tast Dependent Anthropomorphic Grasping Poses for Everyday Objects [Dataset]
• Incorporation of the Intended Task into a Vision-based Grasp Type Predictor for Multi-fingered Robotic Grasping [Dataset]
• Capturing Task-related Information for Text-based Grasp Classification using Fine-tuned Embeddings
• Bridging the Gap to Natural Language-based Grasp Predictions through Semantic Information Extraction [Dataset]

General HRI

• A Multimodal Teach-in Approach to the Pick-and-Place Problem in Human-Robot Collaboration
• Towards Remote Expert Supported Autonomous Assistant Robots in Shopping Environments [Concept Demonstration Video]

Others

• Our Virtual Reality Questionnaire Toolkit [GitHub]
• Our Dialogue Platform stepDP [GitHub | Documentation]

 Projects

• TRACTAT - Transfer of Control between Autonomous Agents
• CAMELOT - Continuous Adaptive Machine-Learning of Transfer of Control Situations
• MarKIeR - Machbarkeitsstudie zum Einsatz von KI-Werkzeugen für die Unterstützung der Landesmedienanstalten zur Verifizierung der gesetzeskonformen Umsetzung der Transparenzvorgaben des Medienstaatsvertrages durch Medienintermediäre
• PABeLA - Prüflabor zur automatisierten Bewertung LLM-basierter KI-Systeme in deren Auswirkungen auf Journalismus und Meinungsbildung
• MultiGrasp - Multimodal Grasp Type Prediction for Dexterous Multi-fingered Robotic Grasping (Project within the Software Campus Program)
• FEDWELL - Life-Long Federated User and Mental Modeling for Health and Well-being
• Project within RESUMAIS - Intel Center of Excellence on Responsible Human-AI Systems

Teaching

• Adaptive Human Machine Interfaces for Autonomous Systems (WS 21/22)
• Grundlagen der Medieninformatik (WS 25/26)

Supervised Theses

• "Hybrid Intelligence for Product Layout Maintenance in Retail Environments" by Hanna Schier (supervised with Martin Feick and Prof. Dr. André Miede from HTW Saar)
• "Task Intent Aware Grasp Type Prediction for Multi-fingered Robotic Grasping" by Ole Keil (supervised with Martin Feick)
• "Advancing Real-Time Class-Incremental Continual Learning through Experience Replay of Vector Quantized Prototypes" by Eiram Mahera Sheikh (supervised with Martin Feick)
• "Task-focused Grasp Classification from LLM-generated Descriptions using Transformers" by Leon Weyand (supervised with Prof. Dr. Klaus Berberich from HTW Saar)
• "Grasping Beyond Convention: Exploring Grasping Methods for Oddly Shaped Objects with Fully Interactive Surfaces" by Anam Imtiyaz (supervised with Dimitar Valkov)

Publications ()