Qiayuan Liao

Qiayuan Liao

Incoming Ph.D. student

University of California, Berkeley

Biography

Qiayuan Liao is a Ph.D. student at UC Berkeley. He is a full-stack roboticist with a great passion for robots and can independently design, build and control a complex robot system. His research interests include legged robots, optimal control, and mechanical design.

Download my Curriculum Vitae.

Interests
  • Legged Robots
  • Optimal Control
  • Mechanical Design
Education

  • Ph.D. Student Mechanical Engineering, 2023--Present

    University of California, Berkeley

  • B.E. in Mechatronics Engineering, 2023

    Guangdong University of Technology

Skills

Mechanic

Fusion360, Solidwork, CNC, 3D Print

Control

TO, NMPC, MPC, TVLQR, PID, Kalman Filter, ROS, Gazebo

Tools

C++, Python, Git, LaTeX, OpenCV

Projects

*
All RoboMaster Legged Robots Optimal Control
All code
legged_control
legged_control

Nonlinear MPC and WBC for legged robot based on OCS2 and ros-controls.
Source
All code
cheetah_ros
cheetah_ros

Hardware/simulation interface and some basic controllers for the quadruped robot based on pinnochio and Cheetah-Software.
Source
90% Commits by me
rm-controls
rm-controls

Hardware/simulation interface based on ros-controls for developing RoboMaster robots and high-performance robots.
Source Docs
Providing whole proposal, early code, supervision
IMU-Camera Hard Trigger
IMU-Camera Hard Trigger

Transfer BMI088’s data to CAN and trigger the camera at the same time. Receive the image and sync the timestamp.
IMU Firmware Camera Driver
Real-time kernel for UP Board
Real-time kernel for UP Board

Real-time Linux kernel for UP Board, which usually used in MIT Cheetah Mini
Source
Independently
SLS 3D printer
SLS 3D printer

A desktop open source Selective‑Laser‑Sintering 3D printer using mirror galvanometer.
Source

Publications

Qiayuan Liao, Zhongyu Li, Akshay Thirugnanam, Jun Zeng, Koushil Sreenath (2022). Walking in Narrow Spaces: Safety-critical Locomotion Control for Quadrupedal Robots with Duality-based Optimization. In ICRA2023 (Under Review).

PDF

Qiayuan Liao, Zhefeng Cao, Hua Chen, Wei Zhang (2022). Real-time Trajectory Optimization and Control for Ball Bumping with Quadruped Robots. In ICRA2023 (Under Review).

PDF Cite Video

Experience

 
 
 
 
 
Undergraduate Researcher
UC Berkeley Hybrid Robotics Group (Prof. Koushil Sreenath)
Jul 2022 – Present Online

Responsibilities include:

  • Development of MPC with obstacle avoidance for the quadruped robot.
 
 
 
 
 
Undergraduate Researcher
SUSTech CLEAR Lab (Prof. Wei Zhang)
Nov 2021 – Mar 2022 Shenzhen, China

Responsibilities include:

  • Development of hardware and simulation interface of quadruped robot (Unitree’s Aliengo) based on ros-control, and Cheetah-Software.
  • Working on control the quadruped robot bumps the ball in the air to the desired position using trajectory optimization and MPC.
  • Research results submitted to ICRA2023 (under review).
 
 
 
 
 
Captain & Founder
GDUT DynamicX Robotic Team
Oct 2019 – Oct 2021 Guanzhou, China
Found and lead a team of up to 45 people in participated RoboMaster Robotics Competition which required us to design, build and control 7 different robots. More than 400 universities around the world participated and 8000 young engineers competed on the stage.
 
 
 
 
 
Emebbed Developer Intern
DJI Technology Co., Ltd.
Jan 2018 – Aug 2019 Shenzhen, China
Participant of RoboMaster High-School Students Vacation Camp, intern only in summer and winter vacation.
 
 
 
 
 
Individual developer (maker)
High & Middle schools
Jan 2015 – Aug 2018 Chaozhou, China

I have been interested in mechanics and electronics since I was a child, which led me to complete the following projects by myself:

  • Reproduced several open-source FDM 3D printers (kossel) in middle school.
  • Developed a novel desktop Selective-Laser-Sintering 3D printer in high school.