Neural-networks-based Adaptive Control for an Uncertain Nonlinear System with Asymptotic Stability

摘要

This paper proposes a neural-networks(NN)-based adaptive controller for an uncertain nonlinear system with asymptotic stability. While the satisfactory performance of the NN-based adaptive controller is validated well in various uncertain nonlinear systems, the stability is commonly restricted to the uniformly ultimate boundedness(UUB). To improve the UUB of the NN-based adaptive control to the asymptotically stability(AS) with continuous control, the existing NN-based adaptive controller is augmented with a robust-integral-signum-error (RISE) feedback term, and overall closed-loop stability is rigorously analyzed by modifying the typical stability analysis for the RISE feedback control. To demonstrate the effectiveness of the proposed controller, numerical simulations for a fault tolerant flight control with a nonlinear F-16 aircraft model are performed.