Simple Recurrent Neural Network Control for Non-minimum Phase Nonlinear System
非最小相位非线性系统的简单递归神经网络控制

We discuss a uniform structure of simple recurrent neural networks, based on which a novel neural control system is developed. With the introduction of the weighted control information into the neural controller's cost function, the method is an extension of the common neural networks controller proposed before. The stability of the whole neural control system is demonstrated and its effectiveness is verified via simulation.