Wei-Han Wang
Papers
3
Total Citations
15
H-Index
2
About
Wei-Han Wang is a robotics researcher whose work centers on advancing reinforcement learning for autonomous robot control, with a particular focus on humanoid and bipedal systems. His key contributions lie in developing adaptive learning algorithms that enable robots to master complex motor tasks without explicit programming. Wang's most influential work, "Adaptive reinforcement learning in box-pushing robots" (2014, 8 citations), introduced an innovative adaptive state aggregation Q-Learning method that facilitates multi-agent cooperation, significantly improving learning efficiency in collaborative manipulation tasks. His research on "Gait balance of biped robot based on reinforcement learning" (2013, 5 citations) addresses the critical challenge of dynamic balance control, demonstrating how Q-learning can enable bipedal robots to maintain stability during single-leg support phases without prior knowledge of their dynamics. Wang also explored novel human-robot interaction paradigms in "Reward shaping for reinforcement learning by emotion expressions" (2014, 2 citations), where he developed a system using interval fuzzy type-2 algorithms to interpret human facial expressions as reward signals, allowing non-experts to guide robot learning through emotional feedback. This interdisciplinary approach bridges affective computing and machine learning, opening new pathways for intuitive robot training.
Research Focus
Key Achievements
Top Papers
- 1Adaptive reinforcement learning in box-pushing robots8 citations · 2014
- 2Gait balance of biped robot based on reinforcement learning5 citations · 2013
- 3Reward shaping for reinforcement learning by emotion expressions2 citations · 2014