TOPICS
Search

Allan Variance


The Allan variance sigma_y^2(tau) is a measure of the stability of a time series y(t) of fractional-frequency deviations from a frequency source. Divide the time series into consecutive, nonoverlapping intervals of duration tau, and let y^__k(tau) be the arithmetic mean over the kth interval. The Allan variance at averaging time tau is

 sigma_y^2(tau)=1/2<[y^__(k+1)(tau)-y^__k(tau)]^2>.

Here the angle brackets denote averaging over k. The square root sigma_y(tau) is the Allan deviation. Unlike the ordinary variance, the Allan variance is formed from first differences of adjacent block arithmetic means and can remain useful for some nonstationary noise models.


See also

Allan Deviation, Standard Deviation, Time Series, Variance

Explore with Wolfram|Alpha

References

Allan, D. W. "Statistics of Atomic Frequency Standards." Proc. IEEE 54, 221-230, 1966. https://doi.org/10.1109/PROC.1966.4634.

Cite this as:

Weisstein, Eric W. "Allan Variance." From MathWorld--A Wolfram Resource. https://mathworld.wolfram.com/AllanVariance.html

Subject classifications