Undergraduate Student, Department of Automation, Tsinghua University
Hello! I am Hankun Li, an undergraduate student in the Department of Automation at Tsinghua University. My research interests lie in Embodied AI, Robot Learning, and 3D Vision. I am particularly interested in world models for robotic manipulation, visuotactile perception for dexterous hands, and diffusion-model-based imitation learning.
Previously, I worked as a research intern in Prof. Jiwen Lu’s lab at Tsinghua University, where I participated in research on 3D vision and robot manipulation. I also contributed to R2RGen, a project on real-to-real 3D data generation for spatially generalized manipulation. This experience deepened my interest in the connection between perception, data generation, and generalization in embodied systems.
In addition to research, I have worked on several technical projects, including a traditional fingerprint detection project and a light-field microscopy 3D reconstruction project. These experiences strengthened my background in computer vision, image processing, and computational modeling.
I am familiar with Python, C++, Linux, ROS 2, and LeRobot. I also have a solid mathematical foundation in calculus, linear algebra, probability, and stochastic processes.
Beyond research, I enjoy physics, music, soccer, and ultimate frisbee. I am also the bamboo flute section leader of the Tsinghua University Chinese Orchestra.
R2RGen: Real-to-Real 3D Data Generation for Spatially Generalized Manipulation
A research project on improving spatial generalization in robotic manipulation through direct augmentation of real-world 3D observation-action data.