Universal Learning Network(ULN) and its application to control systems are discussed. In ULN, any kinds of nonlinearly operated nodes with a continuously differentiable function are connected to each other by multi-branches that may have arbitrary time delays including zero or minus ones. A generalized learning algorithm is proposed, which can be applied to any kinds of networks including static or dynamic networks, feedforward or recurrent networks, time delay neural networks and networks with multi-branches. One of the most important features of ULN is the use of the higher order derivatives. As for the application of ULN, control problems such as robust control and chaotic control are studied using second order derivatives and it is shown that the second order derivatives are effective tools to realize the sophisticated robust control and chaotic control in the nonlinear systems.
|ジャーナル||Research Reports on Information Science and Electrical Engineering of Kyushu University|
|出版ステータス||出版済み - 3 1 1997|
All Science Journal Classification (ASJC) codes
- コンピュータ サイエンス（全般）