Adaptive controller based on the affine equivalent model with disturbance rejection for discrete-time switched systems

Abstract

Switched systems often have hard-to-determine models, and the switching time is often unknown. Mismatched model dynamics and/or switching time may lead to performance degradation. Hence, the present work develops an adaptive controller for single-input single-output nonlinear discrete-time switched systems when the mathematical model, switching time and input–output disturbances are unknown. An affine equivalent model-based on Multi-input Fuzzy Rules Emulated Network characterises the unknown system. The control law is derived according to the tracking error, a proposed sliding surface and the affine equivalent model. The controller parameters were set according to the stability analyses (Uniformly Ultimate Bounded-Lyapunov method). Several experiments, including non-switched and switched systems with input and output disturbances, showed a mean perceptual error of less or equal to $1\mathrm{\%}$ for all cases, with recovery in the switch cases. This successfully proves the closed-loop performance of the proposed controller on a discrete-time switched and disturbed system.

Keywords:

• Switched non-linear systems
• MiFREN
• data-driven controllers
• disturbance compensation
• discrete systems
• experimental validation

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data of the experimental results was generated at Cinvestav Unidad Saltillo and is available from the corresponding author [[email protected]] on request.

Additional information

Funding

This research was funded by Consejo Nacional de Humanidades, Ciencia y Tecnología, CONAHCyT, Scholarship Number: 81167.

Notes on contributors

Miriam Flores-Padilla

Flores-Padilla Miriam has a Bachelor's degree in electronic engineering – at the Instituto Tecnológico de Cuidad Guzmán – and a Master of science degree in robotics and advanced manufacturing -at CINVESTAV Unidad Saltillo. Is currently in the second year of her PhD in robotics and advanced manufacturing. Her thesis topic is model-free adaptive controllers.

Chidentree Treesatayapun

Chidentree Treesatayapun has a Bachelor's degree in electrical engineering, a Master's in electrical engineering and a Doctor of Philosophy degree all from Chiang Mai University in Thailand. He is currently a research professor at CINVESTAV Unidad Saltillo, and his topics of interest are control of discrete-time non-linear systems, Fuzzy logic, and artificial neural networks.