Neural-network-based identification and control of a thin plate using piezoelectric actuators and sensors

This paper proposes a novel neural network approach for the identification and control of a thin simply supported plate. For the control purpose, the piezoelectric sensors and actuators are attached on a flexible structure. The motion behaviour of a two-dimensional model of piezoelectric materials bounded to the surface of the plate is analytically investigated. A novel linear differential inclusion is developed for a class of multilayer feedforward networks. With this technique, it is shown that the plant identified by the neural network can be represented as a linear time-invariant system. On the basis of the identified model, advanced linear control theory can be directly applied to design the stabilizing flexible structure controller. Extensive simulations are conducted to show the effectiveness of the proposed method