Yihua Qin

Undergraduate Student

About Me

I'm Yihua Qin, a third-year undergraduate student in Tsinghua University, where I'm currently pursuing a dual degree, a B.S. in Mathematics & Physics and a B.E. in Mechanical Engineering.

Now I am working at the RVSA Lab in Tsinghua University with Prof. Rui Chen. We are focusing on Visuo-tactile fusion imitation learning.

My research interests are Robotics, Robot learning (Imitate Learning), Computer Vision.

Education

B.S. in Mathematics & Physics, B.E. in Mechanical Engineering Tsinghua University Sep 2022 - Jun 2026

Research Experience

Research Assistant RVSA Lab Tsinghua University Jun 2024 - Present

Research Interests

Robotic Manipulation
Robot Learning
Computer Vision

Publications

Whleaper: A 10-DOF Flexible Bipedal Wheeled Robot

Yinglei Zhu¹⁺, Sixiao He¹⁺, Zhenghao Qi¹, Zhuoyuan Yong¹, Yihua Qin¹, Jianyu Chen^2*.

International Conference on Intelligent Robots and Systems (IROS2024, Oral)

Paper

ViTaMIn: Learning Contact-Rich Tasks Through Robot-Free Visuo-Tactile Manipulation Interface

Fangchen Liu⁺, Chuanyu Li⁺, Yihua Qin⁺, Ankit Kumar Shaw, Jing Xu^*, Pieter Abbeel^*, Rui Chen^*.

Robotics: Science and Systems (RSS2025, Under Review)

Video

Projects

Decision-making algorithms for robot soccer

• Utilized the publish-subscribe feature of ROS topics to control a robot for penalty kicks in the Webots simulation environment on Linux.

• Developed Python code to manage kicking coordination logic for two soccer robots during competitive matches, successfully securing first place against other teams.

• This project was carried out by me as the leader of the decision-making team of THMOS (Tsinghua RoboSoccer Team). Later, we won first place in the China Robocup2024!

LSTM-Based Adaptive Grasping Force Control Method

• Learn the three-loop control method for permanent magnet synchronous motors. Enhance the force control performance by integrating an LSTM neural network into the traditional PI grasping force control method.

• Train LSTM model in simulation environment (generating contact force-position curves for different objects) and validate its compliance and reliability in real world when grasping fragile objects with varying stiffness (such as grapes, tofu, eggs).