Yuki Saigusa
Papers
4
Total Citations
60
H-Index
3
About
Yuki Saigusa is a robotics researcher advancing the frontier of dexterous manipulation through imitation learning. Their work focuses on enabling robots to perform complex, contact-rich tasks—from grasping unknown objects to nonprehensile manipulation—without explicit programming. Saigusa’s key contributions lie in developing bilateral control-based imitation learning frameworks that allow robots to learn both soft and rigid object grasping, as demonstrated in their 2023 paper (23 citations). They have also pioneered methods for variable-speed contact motion, addressing the critical challenge of operating up to the limits of control bandwidth. By incorporating self-supervised learning that accounts for motion speed, Saigusa’s research enables robots to dynamically adapt to changing environments and object physics. Their work on nonprehensile manipulation (21 citations) tackles particularly difficult tasks like sliding or pushing objects, which require deep understanding of environmental dynamics. Through these innovations, Saigusa is bridging the gap between learning-based motion generation and real-world robotic operation, achieving notable success in tasks that demand precise force adjustments and environmental adaptability.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4Imitation Learning for Variable Speed Object Manipulation.2 citations · 2021