H∞ neural network adaptive control

This paper introduces an H_∞ Adaptive Control architecture using neural networks for systems whose uncertainty has an unknown structure. This architecture merges ideas from robust control theory such as H_∞ control design, the Small Gain Theorem, and ℒ stability theory with Lyapunov stability theory and recent theoretical achievements in adaptive control to develop an adaptive architecture for a systems whose uncertainty satisfies a local lipschitz bound. The method permits a control designer to simplify the adaptive tuning process, band limit the adaptive control signal, and treat unmatched uncertainty in a single design framework. The design framework is similar to that used in robust control, but without sacrificing performance. All of this is accomplished while providing notions of transient performance bounds dependent on the characteristics of two linear systems and the adaptation gain.