GhostAR: A Time-space Editor for Embodied Authoring of Human-Robot Collaborative Task with ...
summary
A workflow for authoring human‑robot collaboration where physical human demonstrations become editable AR avatars that act as an AR ghost time-space reference, letting users place and animate virtual robots in Unity and simulate dynamics via ROS before running real robots. A DTW-based DTW motion-mapping model matches live performance to authored ghosts and issues robot commands; trigger detection hits about 300 milliseconds detection error. A 12‑participant, three‑session study shows fast learnability and higher accuracy with practice.
AI-written summary · about
chapters (8)
- 0:00Introduction and motivation
- 1:56Human-robot collaboration challenges
- 3:46GhostAR workflow overview
- 5:31System design with example
- 8:48Technical implementation details
- 9:56User study and evaluation results
- 12:46Takeaways and future directions
- 13:07Q&A session
related talks
08mGesturAR: An Authoring System for Creating Freehand Interactive Augmented Reality Applications
09magentAR: Creating Augmented Reality Applications with Tool-Augmented LLM-based Autonomous Agents
14mImaginateAR: AI-Assisted In-Situ Authoring in Augmented Reality
41mVirtual-Physical Environment Manipulation System based on ROS
12mSwitchAR: Perceptual Manipulations in Augmented Reality
45mThe future of XR authoring:
15mVideo-Annotated Augmented Reality Assembly Tutorials
47mCreating XR Stories