Identification and Switched Control of an Aeropendulum System

An aeropendulum system built at the Control Research Laboratory of UNESP Ilha Solteira is presented, identified and controlled. The system has nonlinear dynamics and unknown parameters. An identification method using MatLab software is used to obtain local linear models and enable the design of automatic controllers using Linear Matrix Inequalities (LMIs). Therefore, the dynamic of the system can be approximated by local linear models at different points of operation combined by membership functions which are unknown in this case. From the approximated mathematical model of the system and the Lyapunov's stability theory, switched controllers considering decay rate and bounds on norm of the state feedback matrices K are designed. The switched control does not require knowledge of the membership functions to compose the control signal and it is an important characteristic that solve the control problem of aeropendulum modelled by Takagi-Sugeno fuzzy models. Practical applications illustrate the efficiency of the methodology used, the results of the implementations are shown and compared.