Robust control of nonlinear systems using norm-bounded neural networks

A new method for designing robustly stable closed-loop systems which contain neural networks is presented. The class of plants considered constitutes a set of unknown but invertible nonlinear systems. In this method, neural network outputs are treated as system uncertainty and are combined with other plant uncertainties so that a robust controller can be designed. A procedure for determining how large the neural network´s output must be and an algorithm for confining the network´s output to be less than this bound is presented. A previous result in robust control is expanded upon for use in this procedure