Lecture
This lecture covers the Stable Estimator of Dynamical Systems (SEDS), focusing on learning and adaptive control for robots. It explains the concept of Global Asymptotic Stability of Autonomous Dynamical Systems using Lyapunov's Theorem. The lecture delves into modeling non-linear dynamical systems with Gaussian Mixture Models, detailing the process of generating estimates and ensuring stability. It discusses the conditions for SEDS stability, the parametrization of SEDS, optimization techniques, and hyperparameters selection. The lecture concludes with a summary highlighting the automatic estimation of globally asymptotically stable dynamical systems using SEDS, emphasizing stability enforcement through Lyapunov constraints and the limitations of the approach.