Control of nonminimum phase systems using neural networks

A new control method using a neural network is proposed for control of a nonminimum phase system. The proposed controller has a hybrid architecture of a multilayer neural network and a linear feedback controller. The multilayer neural network adaptively transforms a control action of the linear feedback controller to follow the system outputs with the nonminimum phase characteristic to those of a reference model with a minimum phase characteristic in each sampling instant. Computer simulation results show that the proposed controller effectively controls the non-minimum phase systems with nonlinear disturbances and time delay properties.