Hi all, I am Donghyun Sung. Currently, I study robotics and maching learning as an master-degree student in DYROS Lab, Seoul National University, Korea. My recent interest is more focusing on practical aspect of grasping also called “pick and place”. Even though it has been studied long decades, it is unlikely a robot do house chores, especially putting something in right location, for me in my house. I think fundamental difference between factory setting and wild real-world is “how much a robot know envrionment in priori”. I am trying to figure out how to answer this question.
Research Interest
- Reinforcement Learning & Imitation Learning in robotics domain
- Human-like Manipulation
- Pick and Place in wild real-world
Education
- B.S 2014.03 ~ 2020.02 Mechanical Engineering, Yonsei University, South Korea.
- M.S 2020.03 ~ Dyros Lab in Seoul National University, South Korea.
Project
- Imitation Learning 2020.03 ~
- B.S Thesis
paper - Edison Contest 2019.01 ~ 2019.03
video - Creative Design Project3 2019.03 ~ 2019.06
video - Creative Design Project2 Team 8(1:02:35 ~ ) 2019.09 ~ 2019.12
video
Experience
- Study Abroad Program 2018.06 ~ 2018.12, Old Dominion University, Virginia, U.S
- Internship 2020.08.03 ~ 2021.02.02 NAVER LABS, Korea
video 02:27 Task learning based on force control
Skills
- Programming: python, c/c++
- Software/library: numpy(linear algebra), pytorch(machine learning), ROS(middle ware), Physics Simulation(mujoco, pybullet)