A State-Space-Based Recurrent Neural Network for Dynamic System Identification

摘要

In this work a new type of recurrent network for modelling the input-output behaviour of a general class of discrete nonlinear systems is presented. The proposed network is a generalisation of the Elman and Jordan networks, based on the state-space description of a nonlinear system. Simulation results of the identification of several stable systems are shown to demonstrate the capabilities of the networks.