Data-Driven Adaptive Control, For Unknown Non-Affine Nonlinear Systems with Varying Control Direction
Miriam Flores-Padilla, Chidentree Treesatayapun · Proceedings of the International Conference of Control, Dynamic systems, and Robotics · 2024
Non-affine systems with variant control directions are often hard to control when no information on the system is available.Hence, this article proposes a data-based adaptive control for this type of system that does not require any information on the system's mathematical model.This article uses a model estimator and a model-free adaptive controller based on Multi-input Fuzzy Rules Emulated Network.The estimator helps to obtain an approximation of the unknown and varying control direction.The estimated control direction helps the adaptive controller to have a fast response when the system's control direction changes, with no previous information on the system.We provide closed-loop stability proof according to a Uniformly Ultimately Bounded function of Lyapunov.As validation, we provide experimental results for a switching gain circuit, where the system's control direction undergoes an abrupt change.The controller could obtain a mean absolute percentage tracking error of 2.59% throughout the experiment.This proves the proposed controller's performance for systems with varying control directions.