Reinforcement learning · Research
Legged mobile manipulation
Building a unified benchmark, reproducing learning-based methods, and studying coordinated quadruped locomotion and arm control.
50SE(3) evaluation trajectories
View project Robotics · Embodied AI
Chenxi Gu顾晨曦
Legged manipulation & embodied AI.
MSc Robotics student at the National University of Singapore. My interests span legged manipulation, reinforcement learning, and embodied AI, with hands-on experience in mechanics, control, perception, navigation, and real-robot deployment.
National University of Singapore
MSc in Robotics
Aug 2026 — Jul 2027 (expected)In progress · Singapore
Xi’an Jiaotong-Liverpool University
BEng in Mechatronics and Robotic Systems
Sep 2022 — Jul 2026GPA 3.92/4.0 · Average 78%
Selected projects
ROS · Final-year project
Multi-floor autonomous navigation
Connecting mapping, relocalization, and floor-aware mission management for indoor navigation on a Unitree robot.
View projectPerception & interaction · Systems
A quadruped museum guide
Connecting visual recognition, robot movement, and spoken narration in a perception–decision–action loop.
View projectMore projects
Mechanical design · Closed-loop control
Residential cupping robot
Designing a compact arm and specialized gripper for home settings, with magnetic-encoder feedback for closed-loop stepper control.
View projectMechanics & system integration
Autonomous delivery vehicle
Building an Autoware-based delivery platform from chassis and suspension to powertrain and sensor integration.
View projectExperience
May 2026 — Present
Hangzhou, ChinaDEEP Robotics
Embodied AI Intern
- Collected and cleaned humanoid robot demonstration data for model training.
- Ran foundational π0.5 training experiments, preparing datasets, training configurations, and output checks.
- Demonstrated robots at the Global Ecosystem Cooperation Conference and represented the company during visits by Chinese Vice Premier Zhang Guoqing and Lao President Thongloun Sisoulith.
Demonstrations→Data preparation→Model training→Output review
Jul 2025 — Aug 2025
Changzhou, ChinaChangzhou Institute of Advanced Manufacturing Technology
Robotics R&D internship
- Contributed to an assistive exoskeleton, selecting motors, designing STM32F103C8T6 driver PCBs, and assembling and testing half-bridge MOS circuits.
- Implemented rotor-feedback FOC, including Park/Clarke transforms and SVPWM.
- Combined weight sensing, YOLOv8 segmentation, and PyQt5 for crab-quality inspection, and trained air-conditioner-part segmentation models for robotic picking.
Sensor feedback→Perception & control→Robot execution
Publications
arXiv · 2026 · Under review
ReCo: Response-Consistent Locomotion with Policy-Aware MPC for Legged Manipulation
IEEE EECR 2026 · pp. 243–248
A LiDAR-Anchored Synchronization Framework for Robust LVI-SLAM
RiTA · Springer LNNS 1132 · 2024 · pp. 321–335
A Residential Robot for Chinese Cupping Physiotherapy
Competitions & honors
2024
RoboMaster University Championship
National third prize · Northern regional first prize
2023
RoboMaster University Championship
National second prize · Northern regional first prize
DOBOT Cup Jiangsu Collegiate Creative Robotics Competition
Second edition · Second prize
Three consecutive Academic Excellence Awards (ranked 2nd in years 1–2 and 1st in year 3) · Graduate Best Overall Award · University Outstanding Student · Provincial Outstanding Student
Notes & tools
arXiv · Robotics
Robot Paper Digest
Following developments in embodied AI, robot learning, and control. Browse curated papers and Chinese–English abstracts by date.
ROS · Linux · Systems
Technical blog
Notes on environment setup and system experiments, alongside browser tools used in everyday development.
Skills & tools
Programming
Python · C / C++
Robotics & simulation
ROS / ROS 2 · Isaac Gym · MuJoCo
Computer vision
OpenCV · YOLOv8
Tools & systems
Ubuntu · Git · LaTeX
Languages
Chinese (native) · English · IELTS 6.5
Get in touch
Open to conversations about robotics algorithms, embodied AI roles, and project collaborations.