Papers
41
Total Citations
490
H-Index
13
About
Wataru Takano is a pioneering robotics researcher whose work sits at the intersection of human motion understanding, natural language processing, and humanoid robot intelligence. His research has fundamentally advanced how robots perceive, symbolize, and generate human movement, enabling richer forms of human-robot communication. Takano's most influential contributions center on bridging the gap between physical motion and language. His hierarchical mimesis model for primitive nonverbal communication (2006, 52 citations) established an early framework for social robots capable of interpreting and mimicking human behavior. Building on this, he developed statistical models enabling bidirectional translation between whole-body motion primitives and natural language sentences (2015, 46 citations), allowing robots to both describe movements linguistically and generate motion from verbal instructions. A recurring theme across his body of work is unsupervised segmentation of continuous human motion into discrete, meaningful primitives — a capability critical for real-time robot learning. His later deep learning approaches, including the HVGH framework (2019), extend these methods to high-dimensional data using neural compression. More recently, his IMU-based annotation generation work (2020) demonstrates a commitment to practical, real-world deployment. Collectively, Takano's research has shaped how humanoid robots acquire, interpret, and communicate knowledge about human movement.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10