The design of controllers for nonlinear uncertain dynamical systems is one of the most important challenging tasks in control engineering. In this paper, we propose a fuzzy system for controlling a nonlinear uncertain plant. We show that alternative techniques of fuzzy control can improve or complement conventional techniques in these kind of plants. The case of use is a real pneumatic positioning system with no mechanically coupling with the final effector, and with nonlinearities and uncertainties. We used a webcam as a feedback sensor with an image processing algorithm. Conventional control techniques for linear systems such as proportional-derivative (PD), proportional-integral (PI), and proportional-integral-derivative (PID), can be applied to control the pneumatic levitation system. However, its response is uncertain for the case of vertical position setpoint variations (due to different indices of turbulence along the tube) and in object characteristics (weight, shape, roughness and size). To overcome that problem, we designed a set of fuzzy control rules considering response of the system under conventional controllers and considering the non-linear dynamics of the plant. The optimal parameters of the conventional controllers were estimated through ITAE performance index. We show the performance of a PD, PI, PID and a fuzzy controller under the same operating conditions with a fixed set point. The results obtained for the proposed fuzzy control system, demonstrates good performance in rising time, settling time, reduced overshoot and greater flexibility than conventional (PD, PI and PID) controllers.
