Stabilizing Neural Control Using Self-Learned Almost Lyapunov Critics
This work tackles stability issues in neural control for robotics, which is crucial for practical deployment but often incremental in approach.
The paper addresses the lack of stability guarantees in learning-based control methods for robotics by developing neural control policies and Lyapunov critic functions in a model-free RL setting, achieving enhanced stability for nonlinear systems like automobiles and quadrotors.
The lack of stability guarantee restricts the practical use of learning-based methods in core control problems in robotics. We develop new methods for learning neural control policies and neural Lyapunov critic functions in the model-free reinforcement learning (RL) setting. We use sample-based approaches and the Almost Lyapunov function conditions to estimate the region of attraction and invariance properties through the learned Lyapunov critic functions. The methods enhance stability of neural controllers for various nonlinear systems including automobile and quadrotor control.