[Isocops] Fwd: Error propagation

Nathan Schwadron nschwadron at guero.sr.unh.edu
Mon Oct 25 14:27:37 EDT 2010



Begin forwarded message:

> From: "DeMajistre, Bob" <Bob.DeMajistre at jhuapl.edu>
> Date: October 25, 2010 11:04:15 AM EDT
> To: Nathan Schwadron <nschwadron at guero.sr.unh.edu>
> Cc: Ken Fairchild <fair-play at comcast.net>, Roland Vanderspek <rvdspek at bu.edu>, Paul Janzen <paul.janzen at umontana.edu>, "EXT Tapley, Mark" <mtapley at swri.edu>, Geoff Crew <gbc at space.mit.edu>, "Maher A. Dayeh" <maher.aldayeh at swri.org>
> Subject: Error propagation
> 
> I’ve got an operating version of the full covariance propagation working for hilo_reflux – the first stage where the non-diagonal elements play. 
> 
> I’ll be committing to shortly for everyone’s inspection. Note that unless you use the new –c option, hilo_reflux will run exactly as it has before. The new software passes all the regression tests. The –c option is described by running “hilo_reflux —help”.
> 
> I’ve enclosed some plots that show how the full propagation works on orbit 0011.
> The first plot shows the variances (the diagonal of the covariance matrix). The black trace shows the original error propagation (what we’ve been using). The red trace is the variance BEFORE hilo_reflux is run (for reference). The yellow trace is the variance from the new error propagation.
> The second plot shows the fluxes and uncertainties – to give a perspective on how big the changes are.
> The third plot shows the relative uncertainty for the three cases on the first plot.
> All of these plots are just different views of the same information. For ESAs 3-5, the varainces are almost identical to what was calculated before. For ESAs 1 and 2 , the new variances are actually closer to the pre-reflux uncertainties (with some notable excursions). In ESA 6, the new variance sits in between the old calculation and the pre-reflux values. I haven’t looked in detail into this behavior yet. 
> 
> Again, I’d like someone to look at the code changes I’ve made. I’m still a newb. 
> 
> Bob

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.sr.unh.edu/mailman/private/isocops/attachments/20101025/157bbd4e/attachment-0002.html 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: eprop_o0011.pdf
Type: application/pdf
Size: 83489 bytes
Desc: not available
Url : http://lists.sr.unh.edu/mailman/private/isocops/attachments/20101025/157bbd4e/attachment-0001.pdf 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.sr.unh.edu/mailman/private/isocops/attachments/20101025/157bbd4e/attachment-0003.html 


More information about the Isocops mailing list