Episodic Learning with Control Lyapunov Functions for Uncertain Robotic Systems
This work addresses model uncertainty in robotic systems, which is a persistent challenge for ensuring stability and safety, but it appears incremental as it builds on existing CLF methods with iterative refinement.
The paper tackled the problem of model uncertainty weakening guarantees in robotic control by developing a machine learning framework using Control Lyapunov Functions to adapt to parametric uncertainty and unmodeled dynamics, resulting in a stabilizing quadratic program model-based controller that demonstrated substantial performance improvements in a planar Segway simulation.
Many modern nonlinear control methods aim to endow systems with guaranteed properties, such as stability or safety, and have been successfully applied to the domain of robotics. However, model uncertainty remains a persistent challenge, weakening theoretical guarantees and causing implementation failures on physical systems. This paper develops a machine learning framework centered around Control Lyapunov Functions (CLFs) to adapt to parametric uncertainty and unmodeled dynamics in general robotic systems. Our proposed method proceeds by iteratively updating estimates of Lyapunov function derivatives and improving controllers, ultimately yielding a stabilizing quadratic program model-based controller. We validate our approach on a planar Segway simulation, demonstrating substantial performance improvements by iteratively refining on a base model-free controller.