Wenli Xiao

Robotics PhD Student at CMU School of Computer Science

I'm a PhD student at CMU Robotics, advised by Prof. Guanya Shi. My vision is to build general intelligent robot that can progressively improve its capabilities in the real world to produce meaningful contributions to the society. I am generally interested in dexterous manipulation and humanoid robots.

I used to work on robot control with Prof. Guanya Shi and Prof. John Dolan at CMU during my master study, where I gained extensive experience with real-world robot and system building (Unitree G1, Booster T1, Unitree Go1/Go2, etc.).

I am currently doing a research internship at NVIDIA GEAR Lab, working with Dr. Jim Fan and Prof. Yuke Zhu, where I am leveraging industrial-level resources for scaling up foundation model learning for humanoid robots.

Google Scholar(loading...)

Twitter Github

Research Projects

HOVER: Versatile Neural Whole-Body Controller for Humanoid Robots

Co-Lead

Authors

Tairan He*, Wenli Xiao*, Toru Lin, Zhengyi Luo, Zhengjia Xu, Zhenyu Jiang, Jan Kautz, Changliu Liu, Guanya Shi, Xiaolong Wang, Linxi 'Jim' Fan†, Yuke Zhu†

HOVER: Versatile Neural Whole-Body Controller for Humanoid Robots

ICRA 2025

TL;DR: HOVER is a 1.5M-parameter neural network to control the body of a humanoid robot. It takes a lot of subconscious processing for us humans to walk, maintain balance, and maneuver our arms and legs into desired positions. We capture this 'subconsciousness' in HOVER, a single model that learns how to coordinate the motors of a humanoid robot to support locomotion and manipulation.

arXiv Website Twitter Code

AnyCar to Anywhere: Learning Universal Dynamics Model

Co-Lead

Authors

Wenli Xiao*, Haoru Xue*, Tony Tao, Dvij Kalaria, John Dolan, Guanya Shi

AnyCar to Anywhere: Learning Universal Dynamics Model

CoRL 2024, X-Embodiment workshop

ICRA 2025

TL;DR: AnyCar is a generalist vehicle dynamics model for agile mobility. It can adapt to various cars, tasks, and envs via in-context adaptation, outperforming well-tuned generalist models up to 54%.

arXiv Website Video Twitter Code