Xuxin Cheng

xuxinc [at] cs.cmu.edu

I am a 1st year PhD student at UC San Diego advised by Prof. Xiaolong Wang. I earned my M.S. in Robotics at CMU advised by Prof. Deepak Pathak. I also worked with Prof. Koushil Sreenath at UC Berkeley. I am interested in general motion generation for real robots.

I love snowboarding 🏂 and scuba diving 🤿 in my free time.

CV  |   GitHub  |   Google Scholar  |   Twitter

profile photo


* denotes equal contribution

Extreme Parkour with Legged Robots

Xuxin Cheng*, Kexin Shi*, Ananye Agarwal, Deepak Pathak
ICRA 2024
CoRL 2023 Generalist / Roboletics / Deployable Workshop (Oral)

Webpage | arXiv | Code | Video | Media | BibTeX

Legs as Manipulator: Pushing Quadrupedal Agility Beyond Locomotion

Xuxin Cheng, Ashish Kumar, Deepak Pathak
ICRA 2023

Webpage | arXiv | Video | Media | BibTeX

Deep Whole-Body Control: Learning a Unified Policy for Manipulation and Locomotion

Zipeng Fu*, Xuxin Cheng*, Deepak Pathak
CoRL 2022 (Oral, Best System Paper Finalist)

Webpage | arXiv | Video | Media | BibTeX

Reinforcement Learning for Robust Parameterized Locomotion Control of Bipedal Robots

Zhongyu Li, Xuxin Cheng, Xue Bin Peng, Pieter Abbeel, Sergey Levine, Glen Berseth, Koushil Sreenath
ICRA 2021

Webpage | arXiv| IEEE | Video | BibTeX
Media: MIT Tech Review | TechXplore | DeepTech | Inceptive Mind | Freethink | Unite.AI | Singularity Hub 

Automated Lane Change Strategy using Deep Reinforcement Learning

Fei Ye*, Xuxin Cheng*, Pin Wang, Ching-Yao Chan
Intelligent Vehicles Symposium (IV) 2020

arXiv | IEEE | BibTeX

Course Projects

Learning Precise Foot Placement Skills with HZD-based Gait Library for Legged Robots

CS285 Deep Reinforcement Learning

Computer Vison and Computational Photography

CS194 Image Manipulation, Computer Vision and Computational Photography (Spring 2020)

[Project page]  [Code

In this project collection, I implemented several state-of-the-art algorithms to manipulate photos and videos.

Development of Spherical Robot

ME102B Mechatronics Design (Fall 2020)

Completed mechanical design in Solidworks; Made parts of the robot by 3d printing and laser cut; Installed electronic components with proper specifications and assembled the robot; Designed control algorithm using PD controller and Kalman Filter to balance the robot.

Campus Life Planner and Navigator

Introduction to Software Engineering (Fall 2018)

[Code]  [Video

CLPN-BIT is a Wechat Mini-Program including features of schedule planning and campus navigation inside Beijing Institute of Technology. CLPN stands for "Campus Life Planner and Navigator".

Development of maze Robot

Intelligent Robots (Spring 2019)


Utilizing ultrasonic sensors and infrared proximity sensors, we developed an algorithm for finding the exit of a maze.

Traversability Analysis in Field Environments Using Deep Inverse Reinforment Learning


In this project we presented an approach to learn cost maps for traversable area extraction from human demonstration using Deep Inverse Reinforcement Learning, bypassing the effort of manual labeling.

source code       credits