Adaptive neural regulator and its application to torque control of a flexible beam

This paper proposes an adaptive regulator using neural network. For a controlled object with linear and nonlinear uncertainties, the conventional optimal regulator is designed based on a known linear part of the controlled object and the uncertainties included in the controlled object are identified using the neural network. At the same time, the neural network adaptively compensates a control input from the predesigned optimal regulator. In this paper, we show how the output of the neural network compensates the control input based on the Riccati equation, and a sufficient condition of the local asymptotic stability is derived using the Lyapunov stability technique. Then, the proposed regulator is applied to the torque control of a flexible beam. Experimental results under the proposed regulator are compared with the conventional optimal regulator in order to illustrate the effectiveness and applicability of the proposed method.