Abstract
A block-oriented nonlinear system consists of a series of blocks. These blocks represent both memoryless nonlinearity and linear dynamics that comprise the overall input-output dynamics of the system. Under this category, the nonlinear systems are represented as Wiener-, Hammerstein-, Wiener-Hammerstein, and Hammerstein-Wiener type of models, etc. Since methods used to estimate the parameters vary with different types of model, structure identification becomes an essential step in identification. In this article, a strategy based on the results of consecutive relay feedback experiments is proposed to select a model among those block oriented ones aforementioned. After the model structure has been determined, estimation of parameters can then proceed. A few example processes are used to illustrate the identification by this proposed method.
