Nonlinear adaptive control using neural networks and multiple models

The principal contribution described here is concerned with combining linear and nonlinear models to improve the performance of essentially nonlinear dynamical systems even while assuring their stability. The system under consideration is defined, and some preliminaries about neural networks and growth rates of signals are given, which is central to the proof of stability. Following this, the well-known results in robust linear adaptive control are reproduced for easy reference. The key result of stability analysis when multiple models are used is presented. This is a very general result, and the special case when some of the models are actually neural networks is included.