Intelligent modelling, observation and control for nonlinear systems

This paper considers the problem of the stable and convergent identification of a static nonlinearity where the structure and the parameters of the linear section of the plant are known exactly. In this first step of the derivation of one intelligent tool the mathematical procedure is shown in detail. In the following steps this first solution is generalized and practical examples are given. Additionally it is demonstrated that a direct design of the nonlinear controller is also possible when the knowledge of the real plant is sparse. These intelligent methods are very valuable control tools, when the plant has relevant nonlinearities. But these tools should be used only when the plant is actually a nonlinear one. Therefore linear control tools for the self-optimizing or the adaptation remain important control tools, when the plant could be approximated by a linear model.