Interactive Robotics in Medicine and Care
© Bielefeld University / Patrick Pollmeier
From Interaction to Demonstration Quality in VR
We investigate how input modality (gloves vs. controller) and its visualisation (hand vs. controller model) influence the execution of everyday tasks in VR. In a virtual kitchen, participants perform typical tasks (e.g. table setting, dishwashing, cutting, cleaning, pouring, showing/evaluating).
In addition to usability/workload, we record high-resolution motion sequences (trajectories) and interactions and analyse them using semantic segmentation (e.g. grasping → acting → putting down). This allows us to assess not only "how fast", but also "how" movements were performed and how consistent/natural they are.
Three input setups are compared:
M: Motion capture gloves with hand visualisation
H: Controller, hand is virtually reconstructed
C: Controller with visible controller model
Research questions (RQs)
- RQ1: Which input configuration offers the best user experience (usability, mental/physical load)?
- RQ2: How do modality and visualisation influence interaction behaviour and demonstration quality (e.g. execution time, unnecessary actions, semantic correctness, consistency of movements)?
