Several Extensions in Methods for Adaptive Output Feedback Control

Abstract

Several extensions to neural network (NN)-based adaptive output feedback control of nonlinear systems are developed. An extension from linearly to nonlinearly parameterized NNs is given for a direct adaptive output feedback approach. An extension that permits the introduction of e-modification in both the direct adaptive approach and an error-observer-based approach is also given. Finally, for the case of nonaffine systems, we eliminate a fixed-point assumption that has appeared in earlier work.