WiGLE.net

Hi everyone,

I was wondering how wigle.net calculates/estimates the location of a network and found a paragraph in the FAQ:

How does triangulation work?
The "triangulation" is actually just an average of the latitudes and longitudes gathered using the signal strength (squared) as a weight. This is probably more correctly called "weighted-centroid trilateration." This assumes that signal strength will change at the inverse square of the distance. This is reasonable as long is you don't get a one-sided view of the network (i.e. only sample it from one side) since it will be skewed in that direction.

Based on that information I searched for some research papers to understand the methods/algorithms better.

D. Han et al. (Researchgate Source) differs between weighted centroid and trilateration:



Based on their differentiation would you say that wigle.net uses a weighted-centroid approach or a trilateration?
I'm asking because I have to write a paper about it

Many thanks in advance!

Well, that's awkward - they've assigned apparently-exclusive meanings to things that aren't mutually exclusive. When I wrote that section (and when we defined our technique) I meant to point out:

1. Trilateration based on n-dimensional observations (not "triangulation") is the correct description for combining observations using averages to calculate an n-dimensional centroid
2. During trilateration, we weight individual observations based on signal-strength squared (RSS^2) to approximate the likely signal-strength centroid more accurately.

Our technique probably predates that paper by 8 years or so, so I can't speak to that terminology without reading the whole thing (shout out to my man, Aaron). I'd look at the general definitions of "centroid," "trilateration," and "weighted average" to understand why we call it what we call it. Of course corrections and improvements in terminology are extremely welcome if you find we're describing it incorrectly, but I'm disinclined to read that passage as mutually exclusive without context. That article is also dated, and may either intentionally or unintentionally suggest that the techniques aren't normally combined.

Cheers,

-ark

Could you please open source the trilateration algo?

I'm afraid that's impractical - it's not a general purpose tool.

we do however open-source a super-quick trilateration algo for the network view in the client:

https://github.com/wiglenet/wigle-wifi- ... .java#L403