Yoichi Tokita
Papers
1
Total Citations
3
H-Index
1
About
Yoichi Tokita is a pioneering researcher in the intersection of bio-inspired robotics and reinforcement learning, with a primary focus on bipedal locomotion and motor control. His most influential work, "Off-Policy Natural Policy Gradient Method for a Biped Walking Using a CPG Controller" (2005, 3 citations), introduced a groundbreaking framework that combines central pattern generator (CPG) controllers—inspired by the neural mechanisms underlying rhythmic animal movements—with reinforcement learning. Tokita’s key contribution lies in developing the CPG-actor-critic model, an autonomous learning system that enables robots to adaptively acquire stable walking gaits without explicit programming. By integrating off-policy natural policy gradient methods, he advanced the efficiency and stability of learning in continuous control tasks, bridging neuroscience and robotics. Though his citation count is modest, his work laid foundational concepts for adaptive locomotion in legged robots, influencing subsequent research in bio-inspired control and autonomous learning. Tokita’s approach remains notable for its elegant synthesis of biological principles and machine learning, offering a pathway toward more resilient and adaptive robotic systems.
Research Focus
Key Achievements
Top Papers
- 1