Evaluating Correlation Between Measurement Samples in Reverberation Chambers Using Clustering

Published

October 17, 2023

Author(s)

Robert Jones, Robert Horansky, Carnot Nogueira, Catherine Remley

Abstract

Traditionally, in reverberation chambers (RC) measurement autocorrelation or correlation-matrix methods have been applied to evaluate measurement correlation. In this article, we introduce the use of clustering based on correlative distance to group correlated measurements. We apply the method to measurements taken in an RC using one and two paddles to stir the electromagnetic fields and applying decreasing angular steps between consecutive paddles' positions. The results – using varying correlation threshold values – demonstrate that the method calculates the number of effective samples and allows discerning outliers, i.e., uncorrelated measurements, and clusters of correlated measurements. This calculation method, if verified, will allow non-sequential stir sequence design and, thereby, reduce testing time.

Conference Dates

June 16, 2023

Conference Location

San Diego, CA, US

Conference Title

101st ARFTG (Automatic Radio Frequency Techniques Group) Microwave Measurement Conference

Pub Type

Conferences

Download Paper

https://doi.org/10.1109/ARFTG57476.2023.10278668

Local Download

Keywords

Correlation, Pearson correlation coefficient (PCC), reverberation chambers (RC), mode-stirring samples, correlative distance, clustering analysis, adjacency matrix.

Citation

Jones, R. , Horansky, R. , Nogueira, C. and Remley, C. (2023), Evaluating Correlation Between Measurement Samples in Reverberation Chambers Using Clustering, 101st ARFTG (Automatic Radio Frequency Techniques Group) Microwave Measurement Conference , San Diego, CA, US, [online], https://doi.org/10.1109/ARFTG57476.2023.10278668, https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=936694 (Accessed December 30, 2024)