Evaluating Uncertainty of Microwave Calibration Models With Regression Residuals

Area of Science:

• Microwave Engineering
• Metrology
• Statistical Modeling

Background:

• Multivariate microwave calibration models are crucial for accurate measurements.
• Quantifying uncertainty in these models is essential for reliable results.
• Existing methods may not fully address uncertainties arising from regression residuals.

Purpose of the Study:

• To develop and present a sensitivity analysis method.
• To introduce a Monte Carlo algorithm for uncertainty evaluation in multivariate microwave calibration.
• To assess the performance and limitations of these algorithms using synthetic data.

Main Methods:

• Sensitivity analysis applied to calibration models.
• Monte Carlo simulation for uncertainty quantification.
• Use of synthetic data to test algorithms, including scenarios with correlated errors.

Main Results:

• The developed algorithms effectively evaluate uncertainty in microwave calibration models.
• Performance and limitations of the algorithms were explored.
• The evaluated uncertainties can be combined with prediction intervals for total measurement uncertainty estimation.

Conclusions:

• The proposed sensitivity analysis and Monte Carlo algorithm provide a robust framework for uncertainty evaluation in microwave calibration.
• These methods are valuable for improving the reliability of microwave measurements.
• The approach aids in estimating the total uncertainty of calibrated measurements.