Attractive concepts for space-based telescopes with large apertures include thin stiffened membranes and assemblies of lightweight mirror segments. Either may be quite stable at rest, but disturbances can produce small distortions that would degrade optical performance if uncorrected, and means must be provided to accurately sense and position to the tolerances required for phase coherence. Hartmann sensing or shearing interferometry have generally been used in the past with ground-based adaptive optics systems, and phase retrieval techniques have recently been proposed for NASA_s NGST program, but all require a relatively bright distant object, the first two effectively measure the local tilt instead of the displacement itself, all require intensive computations to produce the needed result, and phase retrieval may not converge to the correct result. Unlike those techniques which seek to determine and then to correct the wavefront, this proposal is based on the all-but-forgotten "multidither" approach that unambiguously maximizes the intensity of the image itself by individually "tagging" corrective elements with small high-frequency piston oscillations and then adjusting those segments until those frequency components vanish. Sirius proposes to resurrect this promising technique, using high-speed electronics, logic, and detectors to overcome its early faults and to exploit its strengths.
