Skip to content
Yobel sitting on the floor and holding CATCH.

Yobel Sungkook Lee

Robotics ResearcherB.S. Mechanical Engineering, Sungkyunkwan University

I want to build many kinds of robots, each with its own character and purpose, that people want to have around and can rely on. A robot's character comes from both its body and how it moves, yet the two are usually designed separately. I want to design them together, using morphology-conditioned policies that can control many different bodies, so a new design can be tested in simulation before it is built.

So far I've designed and built a worm-inspired robot that crawls across colon tissue and a pipe inspection robot, trained reinforcement-learning controllers for legged robots, and I'm now building CATCH, a small humanoid designed to drive on tracks and stand on legs.

Email: yobel.s.lee [at] gmail [dot] com

CATCH sitting upright and waving, with a smiling face on its screen.
CATCH folded into tracked mode, seen from the side, with its arms forward and legs folded back.

CATCH, my current project, sitting upright and folded into tracked mode. Watch the progress video

Projects

  1. CATCH

    One small body designed to change shape with the task: folded low on its tracks to travel, or upright to reach out and interact with its arms and legs. I designed it in CAD, 3D printed it, and assembled it; the wiring is fully planned but not yet built.

    2026– · Independent project

    Built · Not yet powered on

    CATCH in CAD: the full assembly coming apart, electronics placed in the torso, and the wiring map of all 191 connections.

  2. WORM

    A worm-inspired colonoscopy robot that crawls through the colon by switching its adhesive pads between gripping and slipping. I led its hardware design, prototyping, and control, using an adhesive from a collaborating materials lab.

    2024–25 · Robot lead · SKKU AMRL

    Tested ex vivo · Manuscript under review

    Time-lapse of the WORM robot crawling forward step by step.

  3. Gyecko

    A 241 g gecko-inspired robot that inspects pipes from the inside, turning at a center joint to follow bends and stepping on stamp-type feet with sticky pads. Built by a team of six that I led, it crawled autonomously through dry, wet, and 30° inclined pipes, detecting obstacles along the way.

    2024–25 · Team lead · Highest Honor, Co-Deep Learning Project

    Built · Tested

    Gyecko crawling through a clear pipe, seen from above.

  4. Go2 Extreme Parkour

    A PPO-based perceptive locomotion framework for the Unitree Go2, and a study of how terrain information should reach the Actor and the Critic. In simulation, the best configuration crosses 80 cm gaps and climbs 45 cm parkour steps.

    2025–26 · Research intern · SKKU LAIR

    Simulation

    A simulated Unitree Go2 climbing a series of high steps.

Publications

  1. A strain-triggered pseudo-gel transition for instantaneous and reversible adhesion switching

    Y. Cho, M. Kim, J. Lee, H. Kim, J. Kim, Y. S. Lee, S. Yeo, J. Choi, H. Lim, K. Kim, S. Baik

    Manuscript under review, 2026

  2. Design and Implementation of AI Machine Vision Inspection System with Multi-Optical Images for OLED Film Manufacturing

    T.-Y. Kim, S. Oh, S.-M. Lee, S. Lee, S. Hwang, J. Jeong

    IEEE/ACIS International Conference on Big Data, Cloud Computing, and Data Science (BCD), 2023

Experience

  • 2026–CATCH · Independent project
  • 2025–FineVT · Robotics & ML Researcher, alternative military service
  • 2025–26LAIR, Sungkyunkwan University · Research Intern
  • 2024–25AMRL, Sungkyunkwan University · Undergraduate Researcher
  • 2024Dynamic Robotics Systems Lab, Seoul National University · Undergraduate Researcher
  • 2023AIDIN · Intern, Sensor Development Team
  • 2021–25Sungkyunkwan University · B.S. Mechanical Engineering, summa cum laude, ranked 1st of 186