arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 0911.0822astro-ph.IM
收藏

利用数据挖掘预测自适应光学中的波前以减小伺服滞后误差

Prediction of wavefronts in adaptive optics to reduce servo lag errors using data mining

Akondi Vyas, M B Roopashree, B Raghavendra Prasad

摘要:

自适应光学中的伺服滞后误差导致波前畸变的不准确补偿。已经尝试预测未来的波前,使用数据挖掘的波前的最近的过去,以减少这些错误。蒙特卡罗模拟进行实验获得的数据,密切遵循柯尔莫哥洛夫相位特性。据报道,数据挖掘后,波前校正的6%的改善。数据挖掘分为三个步骤:(a)数据立方体分割(b)多项式插值和(c)波前估计。重要的是要优化段的大小,给出最佳的预测结果。最佳可预测未来的优化有助于选择合适的曝光时间。

英文摘要:

Servo lag errors in adaptive optics lead to inaccurate compensation of wavefront distortions. An attempt has been made to predict future wavefronts using data mining on wavefronts of the immediate past to reduce these errors. Monte Carlo simulations were performed on experimentally obtained data that closely follows Kolmogorov phase characteristics. An improvement of 6% in wavefront correction is reported after data mining is performed. Data mining is performed in three steps (a) Data cube Segmentation (b) Polynomial Interpolation and (c) Wavefront Estimation. It is important to optimize the segment size that gives best prediction results. Optimization of the best predictable future helps in selecting a suitable exposure time.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑