ICRA 2025 | Chanyoung Ahn

Day 1 – Arts in Robotics Exhibition (May 21, 2025, 14:00–18:00)
- Presented a 3D tangible display robot artwork inspired by the Korean Hahoetal mask, combining magnetic jamming mechanics with traditional design (Ahn et al., 2025).
- Enabled audiences to interactively explore changes in stiffness and form through touch, movement, and sound, transforming static art into a dynamic experience.
- Engaged with 100+ visitors over 4 hours, demonstrating both the artistic piece and a soft gripper experiment to highlight the technical aspects of magnetic jamming.
- Project website: Jamming Tangible Display

Arts in Robotics – tangible display robot artwork inspired by the Korean Hahoetal mask.

Day 2 – Workshop Presentation: Handy Moves (May 19, 2025, 09:00–17:30)
- Delivered a poster presentation:
  Chanyoung Ahn, Sungwoo Park, Donghyun Hwang,
  “Effectiveness of Kinesthetic Sensing in In-Hand Rotation of Objects with an Eccentric Center of Mass” (Ahn et al., 2025)
- Investigated how kinesthetic feedback (joint force/torque sensing) improves dexterous manipulation of objects with varying mass and eccentric centers of mass.
- Results: With PCA and RL, kinesthetic sensing improved manipulation performance by up to 2.4× on unseen objects.
- Confirmed that heavier and more unbalanced objects especially benefit from kinesthetic feedback.
- Discussed results during two 40-minute Q&A sessions; audience questions focused on real-world validation and comparisons with tactile sensing.
- Project website: Kinesthetic Rotation

Workshop Handy Moves – poster presentation on kinesthetic sensing for dexterous manipulation.

Impressions and Insights from ICRA 2025
- Gained valuable opportunities for communication with international researchers and observed current challenges in dexterous manipulation—particularly the sim-to-real gap when using tactile and multimodal sensing.
- Notable research highlights:
  - GET-Zero (Graph Embodiment Transformer): Achieves zero-shot skill transfer across 44 robot hands with different DOFs and finger lengths using graph-structured Transformer encoding of joint geometry.
  - In-Hand Translation with Shear and Normal Force Sensing: Introduces ternary shear force sensing to better capture slip dynamics, enabling zero-shot sim-to-real transfer for manipulation of diverse cylindrical objects.
- These works will directly inspire the integration of geometric structure and shear-aware tactile modeling into our lab’s multimodal sensing research.

Paper Review: Get-Zero

GET-Zero (Graph Embodiment Transformer): After studying this work—which achieves zero-shot skill transfer across 44 robotic hands via graph-structured joint-geometry encoding—I led a paper review and Q&A session with 8 participants at KIST PRIME Lab.

We present a compact, standalone system that fuses visual motion with dynamic tactile feedback, enabling audiences to interactively explore changes in stiffness and form—drawing inspiration from the traditional Korean mask, Hahoetal. More than preserving tradition, our goal is to reimagine it as a dialogue between human and art through touch, movement, and sound.

In-hand manipulation is a key capability for dexterous control, yet it becomes challenging when the mass or center of mass (CoM) of an object is not well known. Such intrinsic properties are difficult to infer precisely through visual sensing alone, which limits the reliability of manipulation strategies. This study investigates how kinesthetic sensing can support in-hand rotational tasks by enabling reinforcement learning (RL) agents to adjust to variations in object dynamics, particularly weight and center of mass. Our method incorporates both proprioceptive signals, such as joint angles, and kinesthetic data, joint forces and torques, captured from sensors embedded in a four-finger robotic hand. To reduce the dimensionality of the input space while retaining the relevant dynamics, we applied Principal Component Analysis (PCA). The resulting policy demonstrates improved adaptability. In the simulation, the manipulation success rates increased by 2.09 and 2.40 times on six and twelve previously unseen CoM configurations, respectively. In addition, kinesthetic detection improves performance 1.52 times in ten known configurations. These findings indicate that kinesthetic feedback contributes substantially to robust and generalizable in-hand manipulation. To access our data and video, please visit our project page: https://cold-young.github.io/kinesthetic_rotation/.

Paper Review: Get-Zero

References