Compute various Allan deviations for a constant-rate time series, therein normal, modified and timed Allan deviations of time series in one function.
Inputs are a simple struct containing time series and rate, the outputs contain the assigned Allan deviations.
Calculations of modified and timed Allan deviations for very long time series might become slow. It is advisable to uncomment .msig* and .tsig*, only after calculations of .sig*, .sig2* and .osig* have been proven sufficiently fast.
No pre-processing of the data is performed.
For constant-rate time series, the deviations are only calculated for tau values greater than the minimum time between samples and less than half the total time.
v3.0 faster and very plain code, no plotting; various Allan deviations
can be calculated; script and sample data are availabie on
(Normal, overlapping and modified Allan deviations are calculated in one
function, in strong contrast to MAHs approach of splitting up among various
functions. This might be beneficial for individual cases though.)
v2.0 and others
v1.71 'lookfor' gives now useful comments; script and sample data are
availabie on www.nbi.dk/~czerwin/files/allan.zip
v1.7 Improve program performance by mainly predefining matrices outside
of loops (avoiding memory allocation within loops); no changes to manual
early program core by Alaa MAKDISSI 2003
(documentation might be found http://www.alamath.com/)
revision and modification by Fabian CZERWINSKI 2009
For more information, see:
 Fabian Czerwinski, Andrew C. Richardson, and Lene B. Oddershede,
"Quantifying Noise in Optical Tweezers by Allan Variance," Opt. Express 17, 13255-13269 (2009), http://dx.doi.org/10.1364/OE.17.013255
The screenshot shows the use of allan v3.0 in
 Fabian Czerwinski, Andrew C. Richardson, Christine Selhuber-Unkel, and Lene B. Oddershede, “Allan Variance Analysis as Useful Tool to Determine Noise in Various Single-Molecule Setups,” Proc. SPIE, Vol. 7400, 740004 (2009), http://dx.doi.org/10.1117/12.827975