Robotics & Autonomy
2026-03-01 – 2031-02-28 · Ongoing
National Research Foundation of Korea (NRF)
Background
The research considers how to organize robot workspaces in environments where people and equipment share severely limited space, including space stations. The central challenge is to reduce the installation footprint while accounting for robot motion and interference between tasks.
Current research
- Design multilevel workspaces around robot joint freedom and vertical reach.
- Learn an initial action space from demonstrations, then optimize task order and paths with reinforcement learning.
- Validate Sim-to-Real transfer through simulation, miniature systems and full-scale equipment.
Ground-based demonstration targets
| Metric | Target |
|---|---|
| Demonstration space | Robot café within approximately 3.3 m² |
| Footprint | At least 50% smaller than a conventional commercial robot café |
| Throughput per unit area | 30% improvement |
| Technology readiness | TRL 2 → TRL 5 |
Future directions for space applications
The aim is to extend workspace design and Sim-to-Real learning developed and validated on the ground to robotic tasks inside space stations.
- Application scenarios: Explore inspection and handling of supplies and equipment in narrow passages and around installed equipment.
- Environmental adaptation: Define additional research on motion and contact in microgravity, and safe coexistence with people and equipment.
- Validation approach: Extend the evaluation to space-related task scenarios through simulation and ground-based analog environments.