No abstract

In this paper, we consider the real-time neural-network-based control of a gyro-mirror line-of-sight stabilization platform. This is an example of a multivariablc nonlinear servomechanism encountered in certain practiced applications, and the appropriate neural-network-based modelling and controller design are discussed. In the paper, real-time experimental results in applying the proposed controller to a pilot-scale gyro-mirror platform are presented to demonstrate its effectiveness. These experiments also serve to verify the analytical results in a proptotype real-time application.