Design and analysis of adaptive optics control systems

An adaptive optics system consists of optical components whose characteristics are controlled in a closed loop feedback system. Adaptive optics systems have important applications in astronomy and laser beam propagation.; Traditional adaptive optics control systems are designed and analyzed using single input single output (SISO) techniques. However, recent advances in robust multi-input multi-output (MIMO) control system theory have enabled us to design higher performance closed loop systems and improved analysis techniques. In this dissertation, we develop and apply robust MIMO modeling, identification, and control techniques for adaptive optics systems.; In particular, we define and perform a complete procedure for model identification of adaptive optics systems. We define temporal and spatial performance metrics for the adaptive optics system which generalize to the traditional performance metric. We design a sequence of controllers that are based on SISO techniques but outperform the current state of the art SISO design. We also develop a MIMO controller based on the H-infinity optimization technique which demonstrates superior robust performance over all of the previous designs, with comparable complexity achieved through MIMO model reduction.; One major source of instability and performance limitations in adaptive optics systems is the misalignment between the deformable mirror and the wavefront sensor. We develop a framework to use MIMO structured singular value theory to predict and verify the amount of misalignment that will cause instability. We also develop a nonlinear adaptive filter and demonstrate that accurate estimates of misalignment are possible. To the best of our knowledge, our work presents the first results for the development of real-time control techniques for correction of optical misalignment. Experiments are included which demonstrate the validity of the above techniques.; This dissertation establishes the foundation for improved robust dynamic control of adaptive optics systems, and provides methodologies for performance prediction and estimation for fundamental limitations that are likely to occur. It reduces the gap that exists between the current SISO control techniques practiced by the optical engineers and the robust MIMO control methods developed by the control systems community, and will result in improved control design for adaptive optics systems.