Model-Following Controller for Nonlinear Plants using RBF Neural Networks

In this paper, a design of control systems for nonlinear plants is proposed. It is derived that the control systems of any nonlinear plants can be reduced to a simple second-order model and shown that the control systems so designed yield no offsets caused by load disturbance. For the identification of nonlinear plants, radial basis function neural networks, which are known for their stable learning capability and fast training, are used. In the simulation study of nonlinear plants, it was observed that the error between the plant output and the reference model output is negligibly small. Moreover, in the experimental study of an actual pneumatic cylinder, it is shown that, under varying conditions, the tracking response obtained from the proposed design scheme is robust to the load disturbance.