robust control, machine learning, non-linear systems, echo state network, uncertain systems
Abstract
This study proposes a strategy to stabilize a class of uncertain nonlinear continuous-time systems, considering the inclusion of control laws based on machine learning techniques. A two-loop approach is adopted, consisting of a robust controller and a machine learning (ML) based controller. First, the robust controller is designed to ensure the stability of the closed-loop system, regardless of the control law generated by the ML controller and exogenous disturbances. The ML controller is designed to improve the performance of the closed-loop system by reducing the effects of disturbances on the system output. A numerical example is presented to illustrate the performance of the proposed approach in controlling an unstable nonlinear open-loop system.
