Papers
6
Total Citations
93
H-Index
4
About
Donghoon Youm is an emerging robotics researcher whose work sits at the intersection of legged robot locomotion, reinforcement learning, and autonomous navigation. His research focuses on developing robust control frameworks for quadrupedal robots, blending model-based and learning-based approaches to tackle some of the field's most demanding challenges. Youm's most influential contribution, "Not Only Rewards but Also Constraints," has garnered 58 citations and advances constrained reinforcement learning for legged robots, enabling controllers that balance high task performance with physically natural motion styles. His IFM framework demonstrates a sophisticated hybrid methodology, combining the precision of Model Predictive Control with the adaptability of imitation learning to achieve symmetric and robust quadrupedal gaits. Beyond control, Youm has extended his expertise to autonomous navigation, developing semantic traversability estimation and high-speed locomotion pipelines for complex, discrete terrain environments. His work on proprioceptive state estimation using neural-augmented Kalman filters reflects a commitment to reliable robot operation even when vision is compromised. Perhaps most strikingly, his RAIBO2 project showcases real-world impact — a quadruped completing a full marathon on a single battery charge. Across his portfolio, Youm is establishing himself as a versatile and practically minded contributor to next-generation legged robotics.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5
- 6