Performance of a discrete Linear Quadratic Gaussian controller with a realistic noise model for adaptive optics

Abstract

Modern adaptive optics systems represent a significant
challenge for automatic control. The output disturbance
must be compensated in real time despite significant measurement
noise and system delays. This paper presents the formulation of a
discrete Linear Quadratic Gaussian controller with a realistic
model of the measurement noise. The cases of high, medium, and
low measurement noise are considered. The influence of the
measurement noise on the discrete Linear Quadratic Gaussian
controller in adaptive optics systems is discussed. The analysis of
numerical stability of the discrete controller is provided.