Deep Learning for Robot Locomotion
We explore ways in which deep reinforcement learning can be used to help underactuated and unintuitive robots learn locomotion tasks, ranging from gait discovery to path planning.
Research
Geometric Mechanics of Robot Locomotion
We explore ways in which ideas and techniques from geometric mechanics can be applied to underactuated and dynamic robot systems with an eye toward locomotion.
Exploiting Symmetries in Reinforcement Learning
We explore ways in which inherent symmetries in different systems can be used to improve or make reinforcement learning techniques more efficient.