Vibration identification based on Levenberg-Marquardt optimization for mitigation in adaptive optics systems

Area of Science:

• Optics and Photonics
• Control Systems Engineering
• Mechanical Vibrations

Background:

• Vibrations pose a significant challenge in achieving high performance in adaptive optics systems.
• Effective vibration mitigation is crucial for applications requiring precise optical adjustments.

Purpose of the Study:

• To propose and validate a novel method for identifying vibration models in adaptive optics systems.
• To demonstrate the effectiveness of the identified model in suppressing various types of vibrations.

Main Methods:

• Utilized the Levenberg-Marquardt nonlinear least squares algorithm for vibration model parameter acquisition.
• Implemented linear quadratic Gaussian (LQG) control for vibration mitigation based on the identified model.
• Conducted experimental validation to assess the performance of the proposed method.

Main Results:

• Significantly suppressed vibrations, achieving a root-mean-square residual vibration level below one microradian.
• Demonstrated effective mitigation of a wide range of vibrations, from low-frequency perturbations to high-frequency peaks.
• Showcased superior performance compared to classical control strategies.

Conclusions:

• The proposed vibration model identification method enables substantial vibration suppression in adaptive optics.
• The method is effective across a broad spectrum of vibration frequencies, enhancing system precision.
• This approach offers a significant advancement over traditional vibration control techniques.