Adaptive framework for uncertainty analysis in electromagnetic field measurements.

Area of Science:

• Metrology
• Electromagnetics
• Data Science

Background:

• Uncertainty in electromagnetic field (EMF) measurements can lead to inaccurate risk assessments.
• The standard Guide to the Expression of Uncertainty in Measurement (GUM) often uses static models and neglects prior information, yielding non-robust uncertainties.
• Robust uncertainty quantification is essential for reliable EMF exposure assessment.

Purpose of the Study:

• To develop a principled and systematic framework for EMF measurement uncertainty analysis.
• To fuse information from current measurements with prior knowledge for improved accuracy.
• To provide a more robust alternative to the conventional GUM approach.

Main Methods:

• Developed a framework integrating current measurements and prior knowledge.
• Utilized a likelihood function based on kernel mixtures for dynamic data adaptation.
• Employed importance sampling to incorporate flexible prior information.
• Validated the framework using a broadband radiation meter and isotropic field probe.

Main Results:

• The proposed framework significantly outperforms the GUM approach in uncertainty analysis.
• Achieved a 28% reduction in measurement uncertainty compared to the GUM method.
• Demonstrated the framework's ability to dynamically adapt to data and incorporate prior information.

Conclusions:

• The novel framework offers a more robust and accurate method for EMF measurement uncertainty analysis.
• Integrating prior knowledge and current data leads to substantial improvements over traditional methods.
• This approach enhances the reliability of EMF exposure assessments and risk evaluations.