Introduction
Adaptive Optics is a steadily growing field, especially in areas such as Astronomy, Ophthalmology, and Telecommunications.  It is a method of compensating for regions of distortion in an optical system by means of a distortable optical element such as a deformable mirror.  By changing the shape of a deformable mirror, a distorted phase front imaged onto it can be “flattened”, thereby allowing the Strehl ratio of the system to be greatly enhanced.  Often it is possible to obtain diffraction-limited performance in adaptive optics systems, even if initial Strehl ratios are quite poor.
 Existing methods of implementing adaptive optics are quite complex.  One of the most troublesome aspects is getting the information regarding the phasefront so that corrections can be made to improve it.  The methods often involve the use of imaging detectors (such as CCD’s), frame grabbers, and image analysis software.  These systems are quite complex and require a large amount of computer processing.  For most applications of adaptive optics, the distorting media is constantly changing, and corrections to the phasefront have to be made frequently.  Due to the complexity of the image acquisition and analysis, this can be somewhat time consuming, and the ability to correct for distortions that occur at faster rates is often limited.  These methods also typically require knowledge of a previous state of the system, as a direct analysis of an interferogram image only delivers information regarding the locations of fringes (like having a topographical map without a reference altitude).  To know whether an actuator on a deformable mirror must move up or down to correct a phasefront requires knowledge of where the fringe was located in a previous analysis loop.  A memory bank of the history of the actuator movements and phasefronts is needed.  Under some circumstances, this history may need to be several frames in duration.
    The complexity of these setups lend themselves to solutions involving VLSI.  By implementing signal processing electronics integrated within the detector array, a significant reduction in the amount of computer processing and software needed to drive the system should result.  This also leads to a substantial increase in refresh rates, and an improvement to the signal to noise ratio.

Abstract
Methods
Results
Discussion
Ackowledgements

Send correspondence to: winsor@stsci.edu

07 December, 2001