Changes include:
- Added an improved method for generating random vectors that allows for less variation during interpolation. This means that the recommended grid size can be made much smaller (now 8 times the obscuring distance, down from 20).
- Refined the calculations around what grid size was necessary and the effect that changing the offset vector has on worst-case ability to reduce uncertainty area.
- Added an example.
- Wording and the usual editorial cleanup stuff.
My demonstration page is now live here, using the updated method:
<http://held-location.sourceforge.net/js_geoshape/maptest.html>
--
I've also developed a few tools for visualizing the interpolation process using Processing. The following are probably of most interest:
Random vector grid visualization
<http://held-location.sourceforge.net/randomVectorGrid/randomVectorGrid.jar>
Shows how random vectors are chosen and interpolated. Compares both interpolation methods.
Square peg interpolation visualization
<http://held-location.sourceforge.net/interpolation/interp.jar>
Shows how interpolation is used between two different points in the circle. Compares interpolation methods and demonstrates how the square peg method is used. I found the loops that form to be interesting.
--Martin
p.s. I'd link to the applets directly, but I couldn't be sure that they are safe - the applets caused every browser I have to die horribly. It seems like a Java problem. Strip the last part of the path off if you are game.
On 2011-01-14 at 16:25:38, IETF I-D Submission Tool wrote:
> A new version of I-D, draft-thomson-geopriv-location-obscuring-02.txt
> has been successfully submitted by Martin Thomson and posted to the
> IETF repository.
_______________________________________________
Geopriv mailing list
Geopriv@ietf.org
https://www.ietf.org/mailman/listinfo/geopriv
