Allan deviation calculation for time-stamped data with missing datapoints and/or outliers

[aewallin]

Implement allan deviation calculation for time-stamped data with missing datapoints and/or outliers (removed by MAD test).

Reference:
Sesia et al. "Application of the Dynamic Allan Variance for the Characterization of Space Clock Behavior" (IEEE TRANSACTIONS ON AEROSPACE AND ELECTRONIC SYSTEMS VOL. 47, NO. 2 APRIL 2011)

Matlab code for this is also available.

[ChesterCamm]

Has this been implemented?

[aewallin]

There is gradev()
https://github.com/aewallin/allantools/blob/master/allantools/allantools.py#L1244

More documentation and tests for this would be useful.
I am not sure if the implementation of gradev() follows the paper mentioned above.
Also, I'm not sure if Stable32 can be used to generate comparison-results given some test-dataset with gaps.

[jondoesntgit]

Is there any known method to do this with gaps in frequency data (not phase data)?

_{Sent with GitHawk}

[aewallin]

There's a reference to another Metrologia paper in the discussion of #60.

I don't think there's a 'golden' solution for missing frequency data. From what I recall the solutions either pretend there's no gap at all, or perhaps fill the gap using some guess/model of what the value might be.

[jondoesntgit]

Good to know.

If you have any other references that would be useful about gaps in frequency data, I might include them in a paper that I'll eventually publish on an algorithm we developed at Stanford to fill in gaps in frequency data and infer the Allan Deviation from those gaps.

We use this library all the time for our research in fiber optic gyroscopes, which measure a rotation rate (frequency-type data), which is integrated into an angle. The IEEE 952-1997 standard uses Allan Deviation frequency-data graphs to characterize the performance of the units under test.

[aewallin]

here is a recent paper on missing data points:
https://www.researchgate.net/publication/334315256_The_Corrected_Allan_Variance_Stability_Analysis_of_Frequency_Measurements_With_Missing_Data