Yu-Ming Eric Chen
I build physical intelligence for legged robots—merging reinforcement learning with model-based control to make high-energy behaviors reliable in the real world.
I build physical intelligence for legged robots—merging reinforcement learning with model-based control to make high-energy behaviors reliable in the real world.
I’m a roboticist and applied scientist at the Robotics and AI Institute (formerly the Boston Dynamics AI Institute). I design, train, and ship reinforcement-learning policies and model-based controllers that run on real robots—enabling landing recovery from aerial states and a unified hopping–wheelie–driving policy. To tighten the sim-to-real gap and support high-energy behaviors, I build battery–motor circuit models that capture effects like voltage sag and power limits.
Previously, I completed my Ph.D. in the GRASP Lab at the University of Pennsylvania with Prof. Michael Posa; before that, I earned an M.S. in Robotics from the University of Michigan, Ann Arbor, and a B.S. in Physics from National Taiwan University.
My research spans reduced-order models, motion planning, and control for legged robots, with a recent focus on reinforcement learning and sim-to-real transfer. I’m driven to make mobile robotic platforms interact dynamically and safely with the real world —pushing their physical capabilities and enabling useful work beyond the lab.