Efficient computation of minimum-variance wave-front reconstructors with sparse matrix techniques.

摘要: The complexity of computing conventional matrix multiply wave-front reconstructors scales as O(n3) for most adaptive optical (AO) systems, where n is the number deformable mirror (DM) actuators. This impractical proposed systems with extremely large n. It known that sparse methods improve this scaling least-squares reconstructors, but techniques are not immediately applicable to minimum-variance now favored multiconjugate (MCAO) multiple sensors (WFSs) and DMs. Complications arise from nonsparse statistics atmospheric turbulence, global tip/tilt WFS measurement errors associated laser guide star (LGS) position uncertainty. A description given how can still be applied by use a approximation turbulence recognizing terms arising LGS uncertainty low-rank adjustments evaluated using inversion lemma. Sample numerical results AO MCAO illustrate made has negligible effect on estimation accuracy, time compute reconstructor natural system O(n3/2) only few seconds = 3500, reduce computations factor 8 sample 2417 DM actuators 4280 subapertures. With extrapolation 9700 17,120 subapertures, reduction approximately 30 or 40 1 predicted.