Output-feedback based model following nonlinear adaptive control using neural netwok

Abstract

This paper deals with an adaptive control problem based on output-feedback. The objective of this problem is that the output of nonlinear system using this adaptive control technique should follow a given command, which is continuous and differentiable in time. In this approach, relative degree of the output is assumed to be known. This approach introduces error state observer and single hidden layer neural network to the adaptive control structure. The update law of the neural network weights is obtained from ultimate boundedness of error signal through the direct method of Lyapunov stability. This technique is applied to an example of `Van der pol' problem to demonstrate effectiveness of this technique.