Ryota YAMASHINA
Papers
8
Total Citations
51
H-Index
4
About
Ryota Yamashina’s research lies at the intersection of bio-inspired robotics, reinforcement learning, and human-robot interaction. His pioneering work on caterpillar robot locomotion demonstrated how Q-learning—an unsupervised reinforcement method—can enable a physical robot to autonomously discover optimal motion forms, using both objective and subjective reward signals. This foundational study, with 13 citations, opened new avenues for adaptive robot behavior. Yamashina further advanced reinforcement learning by exploring dynamic reward changes during training, mirroring human skill acquisition, and by integrating adversarial imitation learning with behavioral cloning to accelerate policy learning. In teleoperation, he proposed innovative methods that create seamless transitions between real and virtual environments and leverage illusions of human intention and time to resolve conflicts between autonomous control and operator acceptance—work that has garnered 11 citations. His contributions also include a novel involute-curve-shaped mechanism for stair-climbing robots and the use of linear temporal logic to guide reinforcement learning under uncertain event detection. Through these diverse achievements, Yamashina has significantly shaped how robots learn, move, and collaborate with humans.
Research Focus
Key Achievements
Top Papers
- 1
- 2Teleoperation by seamless transitions in real and virtual world environments11 citations · 2023
- 3
- 4
- 5
- 6Teleoperation Method by Illusion of Human Intention and Time3 citations · 2021
- 7
- 8