Neural Identification for Control
This addresses the challenge of controlling complex, unknown systems for robotics and automation, but appears incremental as it builds on existing neural and Lyapunov-based methods.
The paper tackles the problem of stabilizing unknown nonlinear dynamical systems by learning a control law and stable closed-loop dynamics jointly through self-supervised learning, using random control inputs as supervision and Lyapunov stability theory, and demonstrates it on tasks like pendulum balancing and vehicle path following.
We present a new method for learning control law that stabilizes an unknown nonlinear dynamical system at an equilibrium point. We formulate a system identification task in a self-supervised learning setting that jointly learns a controller and corresponding stable closed-loop dynamics hypothesis. The input-output behavior of the unknown dynamical system under random control inputs is used as the supervising signal to train the neural network-based system model and the controller. The proposed method relies on the Lyapunov stability theory to generate a stable closed-loop dynamics hypothesis and corresponding control law. We demonstrate our method on various nonlinear control problems such as n-link pendulum balancing and trajectory tracking, pendulum on cart balancing, and wheeled vehicle path following.