Minimum seeking neural network for direct feedback control

Abstract

This paper presents a faster, robust learning algorithm for a neural network controller design. The learning scheme is regarded as finding the optimal weights via the proposed minimum seeking scheme. The only information required is the system output measurement.