Statistical approach of truncation order selection for evaluating wavefront aberration effect in cold atom...

Area of Science:

• Atomic physics
• Optical metrology
• Precision measurement

Background:

• Wavefront aberration (WAE) significantly impacts cold atom gravimeter (CAG) accuracy.
• Determining the Zernike polynomial truncation order (TO) for WAE assessment in CAGs is challenging due to limited parameter adjustments.
• Conventional methods struggle to obtain sufficient data for reliable TO determination.

Purpose of the Study:

• To develop and validate a robust method for determining the optimal truncation order (TO) of Zernike polynomials for wavefront aberration (WAE) assessment in cold atom gravimeters (CAGs).
• To enhance the accuracy and reliability of WAE correction in CAGs by balancing model complexity and predictive performance.

Main Methods:

• Proposed a combined approach using Leave-One-Out Cross-Validation (LOOCV) for pre-screening candidate orders and F-statistic for validating higher-order model significance.
• Conducted experiments using modulated Raman beam aperture under three distinct wavefront conditions.
• Implemented joint fitting of data from multiple wavefront conditions to improve evaluation uncertainty.

Main Results:

• The proposed LOOCV and F-statistic method accurately determined the Zernike polynomial TO in conventional CAGs, outperforming traditional approaches.
• Joint fitting of data across different wavefront conditions reduced evaluation uncertainty by approximately 64% on average compared to individual fitting.
• Demonstrated the capability to balance model complexity and predictive performance in WAE assessment.

Conclusions:

• The developed method provides an accurate and reliable means for TO determination in WAE assessment for CAGs.
• Joint fitting significantly enhances the precision of WAE evaluation, leading to more accurate gravimeter measurements.
• This approach offers a crucial advancement for improving the performance and accuracy of cold atom gravimeters.