Home

Yaonan Zhu (朱曜南)

Project Lecturer, The University of Tokyo

About Me

I am a Project Lecturer in the Matsuo-Iwasawa Lab at The University of Tokyo, Japan. From 2021 to 2024, I was a Project Assistant Professor at Nagoya University, Japan, where I supervised or co-supervised Ph.D. and Master students. I have been working closely with Prof. Yasuhisa Hasegawa and with the research group of Prof. Liming Chen, Prof. Jan Peters, and Prof. Jian Huang. In 2021, I obtained my Ph.D. degree from Nagoya University, supervised by Prof. Yasuhisa Hasegawa. During my Ph.D. studies, I visited the Mechanical Systems Control Lab at the University of California, Berkeley, USA, where I was supervised by Prof. Masayoshi Tomizuka.

Research Focus

Embodied Intelligence Robot Foundation Models Humanoids Robot Learning Teleoperation Tactile Dexterous Hand Control Systems

Call for Students: Motivated students majoring in robotics or computer science, including visiting students/scholars, interested in embodied intelligence, humanoids, and dexterous manipulation are welcome. The research group provides advanced robotic hardware and ample local GPU/HPC resources. We also maintain active collaborations with overseas universities and research teams. Contact me.

I am open to collaborations; please feel free to contact me.

Contact Me

Email: yaonan.zhu@weblab.t.u-tokyo.ac.jp

Linkedin: linkedin.com/in/yaonan-zhu-4ab32a8b

Languages

Native in Japanese and Chinese, fluent in English. I know some French and German.

Selected Research Projects

• Foundation Models Enabled Robotics

  This research aims to enhance autonomous robotic manipulation through Human-Robot Collaboration (HRC) by leveraging foundation models to translate high-level language commands into actionable motion sequences, integrating teleoperation with Dynamic Movement Primitives (DMP) for skill learning and adaptation. [Link]

  
• Learning Locomanipulation Policies

  This research investigates the development of locomanipulation policies for legged robots, emphasizing whole-body control and reinforcement learning to achieve seamless integration of locomotion and manipulation. By leveraging Isaac Sim for sim-to-real transfer, humanoid teleoperation, and Vision-Language-Action (VLA) models, it addresses challenges in deploying adaptive and robust policies in real-world environments. The project also explores laboratory automation scenarios where humanoid robots can support repetitive, instrument-intensive, and safety-critical research workflows. [Coming Soon]

  
• Learning Enhanced Robot Manipulations

  This research investigates learning-enhanced robot manipulation, addressing challenges such as dexterous force-sensitive manipulation, teleoperation-based skill acquisition, and foundation model-enhanced perception and planning. By integrating classical control with data-driven methods, including reinforcement learning, world models, and world action models, it aims to develop generalizable and robust manipulation capabilities across diverse tasks and environments. The project further explores dexterous hand manipulation for fine-grained contact-rich skills and adaptable object interaction. [Coming Soon]

  
• Shared Autonomy and Teleoperation

  This research focuses on developing shared autonomies that enable intuitive and efficient teleoperation of robots by combining AI-driven methodologies with human input, and is conducted by an international team from Japan, Germany, and France. [Link1], [Link2]

  
• Haptic Interfaces

  This research develops a multi-point tactile interface capable of presenting shear forces alongside normal forces, allowing for real-time perception of object slippage and posture changes during dexterous in-hand manipulation, which was featured in the News Media. [Link]

  
• Tactile Sensing and Feedback

  This research leverages vision-based tactile sensors such as DIGIT and Gelsight to estimate contact force, enabling force feedback through bilateral control. It aims to expand tactile feedback possibilities by feeding back both normal and shear forces utilizing novel tactile interfaces. [Link1] [Link2]

  
• Exoskeletons and Robotic Limbs

  This research focuses on the development of exoskeletons and robotic limbs, exploring advanced control systems, biomechanics, and sensor technologies to enhance human mobility and strength, while emphasizing embodiment to create more seamless integration between users and robotic artifacts. [Link1], [Link2], [Link3]

  

Memorable Moments

We have received Best Paper Awards from IEEE International Conference on Mechatronics and Automation (ICMA 2024).

Institutional Journey

Here are the universities I have been affiliated with: