Scanning Micromirror Calibration Method Based on PSO-LSSVM Algorithm Prediction.

Area of Science:

• Micro-opto-electro-mechanical systems (MOEMS)
• Precision engineering
• Optical metrology

Background:

• Scanning micromirrors are essential in MOEMS but susceptible to environmental and fabrication errors affecting measurement accuracy.
• Calibration of scanning micromirrors and their associated measurement systems is critical for reliable performance.
• Existing calibration methods may not fully address the complex influences on micromirror deflection.

Purpose of the Study:

• To develop and present a novel calibration method for scanning micromirrors.
• To establish a precise correspondence between actual micromirror deflection angles and measurement system outputs.
• To enable accurate prediction of micromirror tilt angles for improved MOEMS applications.

Main Methods:

• Implementation of a particle swarm optimization-least squares support vector machine (PSO-LSSVM) regression model.
• Utilizing the PSO-LSSVM to predict the tilt angle based on measurement system output.
• Establishing a predictive relationship between the actual deflection angle and the measured output.

Main Results:

• The developed PSO-LSSVM model successfully predicts micromirror tilt angles.
• A high decision factor (R2) of 0.9947 was achieved for the model on the x-axis.
• Demonstrated the effectiveness of the proposed calibration method in enhancing measurement accuracy.

Conclusions:

• The PSO-LSSVM-based calibration method offers a robust solution for improving scanning micromirror accuracy.
• This approach effectively mitigates errors arising from fabrication and environmental factors.
• The high R2 value indicates the significant potential of this method for practical MOEMS applications.