Control of legged robots under uncertainty

Many robots like quadruped and manipulation systems are difficult to control because they undergo changing contact conditions. These discontinuous events make it difficult to use conventional algorithms for control, state estimation, learning, etc. that assume smooth systems. In this project, the student will take advantage of the structure of these hybrid dynamical systems to produce new controllers that handle uncertainty and impacts. The ideal candidate would have taken a feedback controls class and be very comfortable with linear algebra and vector calculus.