An applied optimization-based gain-scheduled fuzzy control

A method for a constructive and fully automated design of a fuzzy gain-scheduler, which gives guarantees in terms of stability and prescribed performance, is presented. The properties are valid not only locally around an operating point, but globally inside a wide working range and without any restrictions to slowly varying scheduling variables. The design method is based on linear matrix inequality (LMI) techniques and includes an output feedback controller with a state-estimator. The dynamic output-feedback controller was applied to a real-world experiment of a liquid level control with a low-cost sensor having unmodeled nonlinear characteristics