On-line identification of multivariable nonlinear system using neural networks

In this paper, an on-line identification method based on recurrent neural networks (RNN) proposed for multivariable nonlinear systems. This work is an extension of an on-line method for single-input single output system. The large number of input-output vectors is being considered. As the complexity and nonlinearity of the systems is treated. The effectiveness of the proposed algorithm applied to two examples of multivariable nonlinear dynamic systems is demonstrated by simulation experiments. The results of simulation showed that the use of the neural networks is helpful for adaptive strategy design.