Control system design for retinal imaging adaptive optics systems

This paper presents a solution to control system design issues for adaptive optics systems used in retinal imaging. In this paper, the control problem for adaptive optics systems is generalized to that of shape control for a flexible membrane representing a deformable membrane mirror. Due to the dynamic nature of the aberrations in the eye, the shape control problem addressed is the tracking of an unknown and time-varying shape for a distributed membrane (i.e., desired shape of the mirror). The design of a controller to achieve the shape control objective is based on a model of a distributed parameter system representing the mirror membrane. To accomplish this task, a multivariable centralized controller is utilized. Since the desired shape of the mirror is unknown and time-varying, the controller is tuned online to converge to the controller needed to achieve regulation. This is done iteratively, by taking advantage of the Q-parameterization of all stabilizing controllers, so that the controller will converge to the ideal controller. The online tuning is used to compensate for the lack of information of the desired shape for the deformable mirror.