Characteristic of adaptive type neural network direct controller with separate learning rule of each layer

A previous my paper proposed a new neural network learning rule for three layer nonlinear neural network. It was called separate learning rule of each layer. This learning rule is that the neural network weights between one layer and next layer are only changed at same time and other neural network weights are not changed. One of advantages of the proposed learning rule is to realize easier discussion of the neural network controller stability condition. This paper presents several simulation results and discusses the characteristic of the adaptive type neural network direct controller with the separate learning rule of each layer.