A Double Parasols Neural Network and its application to nonlinear discrete time systems

A new neural network, double parasols neural network, is proposed here which is similar to a multilayered neural network, but is derived from an interconnected neural network. EBP-EWLS method can be used as a learning algorithm. The proposed network is considered as a generalization of a regressor model for linear systems to nonlinear systems. Using this network, identification and control are examined for nonlinear discrete time systems to present a validity of the proposed method. Control method used here is a target following control (TFC) which is also the original work of authors.