NIST logo

Uncertainty analysis for autocorrelated measurement data


Develop approaches to assess uncertainties for various autocorrelated measurement data.


When repeated measurements are autocorrelated, it is not appropriate to use the traditional approach to calculate the uncertainty of the average of the measurements, which assumes that the measurements are statistically independent. We need to develop approaches for assessing uncertainties for various autocorrelated measurement data including stationary process data and non-stationary process data. Some tools such as Allan variance also need to be investigated.

Major Accomplishments:

Published two major papers in Metrologia on calculations of uncertainty of the mean of stationary measurement data and Allan variance of time series models for measurement data, respectively. Presented invited talks on the topics in several international conferences.

Lead Organizational Unit:


Nien-fan Zhang

100 Bureau Drive, M/S 8980
Gaithersburg, MD 20899-8980