Yu-Ming Chen
Papers
2
Total Citations
14
H-Index
2
About
Yu-Ming Chen is pioneering the intersection of reduced-order modeling and reinforcement learning for legged locomotion. His research focuses on developing task-optimal simple models that capture critical dynamics of bipedal robots while remaining computationally tractable for real-time control. In his highly cited 2024 work "Beyond Inverted Pendulums," Chen challenges conventional approaches by demonstrating that traditional reduced-order models often constrain full-body dynamics unnecessarily, proposing instead a framework for designing models tailored to specific locomotion tasks. His complementary work on reinforcement learning for reduced-order models bridges the gap between model-based safety guarantees and model-free adaptability, enabling robots to learn robust locomotion policies within physically meaningful low-dimensional spaces. With over 14 citations across his key publications in just one year, Chen's contributions are rapidly gaining recognition for addressing a fundamental tension in robotics: how to maintain the simplicity of reduced-order models while achieving the performance of full-order control. His work promises to advance both the theoretical foundations and practical deployment of agile, efficient legged robots.
Research Focus
Key Achievements
Top Papers
- 1
- 2Reinforcement Learning for Reduced-order Models of Legged Robots6 citations · 2024