Sang Hyoung Lee
Hanyang University, Korea Institute of Industrial Technology
Papers
13
Total Citations
147
H-Index
6
About
Sang Hyoung Lee is a leading researcher in robot skill acquisition, imitation learning, and deep reinforcement learning (RL), with a focus on enabling robots to autonomously learn, improve, and generalize complex manipulation tasks. His most influential work, an autonomous framework for segmenting robot trajectories (52 citations), laid the foundation for his contributions to motor skill learning. Lee’s research addresses critical challenges in manufacturing and daily-life tasks, particularly through his work on peg-in-hole tasks, where he developed a framework combining imitation learning and self-learning to master hole search and peg insertion skills (30 citations). He has also advanced visual grasping by accelerating actor-critic deep RL through state representation learning and sim-to-real transfer, achieving robust grasping of diverse unseen objects with minimal real-world data (14 and 9 citations). Lee’s notable achievements include proposing methods for incremental skill learning from demonstration, generating skill transfer orders based on motion complexity, and using probabilistic affordance for dependable behavior selection. With over 140 total citations across his top papers, Lee’s work bridges simulation and reality, pushing the boundaries of skilligent robots that can adapt to uncertainties and perturbations in real-world environments.
Research Focus
Key Achievements
Top Papers
- 1Autonomous framework for segmenting robot trajectories of manipulation task52 citations · 2014
- 2
- 3
- 4
- 5
- 6Skill learning and inference framework for skilligent robot7 citations · 2013
- 7Incremental learning of primitive skills from demonstration of a task6 citations · 2011
- 8
- 9
- 10