Input as cartesians?

[JulianSchuette]

Hi,

just to make sure, I assume TrilaterationFunction expects the position argument to be in cartesian coordinates, not in lat/lon or anything else?

Cheers,
Julian

[lemmingapex]

Julian,

Yes, you are correct. Positions should be valid cartesian coordinates. In full: Inputs identified as tuples (lists) of n real numbers should exist in a euclidean space of dimension n.

If you are using lat/lon, you will need to convert to a cartesian coordinate system. Depending on the range of latitudes and longitudes inputs, and the desired accuracy, you will want to be cognisant of this conversion.

If accuracy is not a concern, you can model the earth as a sphere:
http://stackoverflow.com/questions/1185408/converting-from-longitude-latitude-to-cartesian-coordinates/1185413#1185413

If you have large lat/lon ranges or you are using an ellipsoid model of the earth (like WGS-84), you will want to look into solutions that account for this.

Does this answer your question?

• Scott

[JulianSchuette]

Perfect, Scott. Thanks a lot.
I am using pretty small lat/lon scales (<100m) and can accept minor conversion errors, so I will go the simple route first.

Cheers.

[kliewkliew]

Using the linked SO equations, the input positions would be composed of x and y (throwing away z)?

[lemmingapex]

Why would you throw away z?

[kliewkliew]

The positions are arrays of two doubles but there are three axes in the SO answer. I don't understand how to coerce the SO answer into the required input positions.

[lemmingapex]

@kliewkliew https://github.com/lemmingapex/Trilateration/blob/master/src/test/java/com/lemmingapex/trilateration/TrilaterationTest.java#L255-L267

[kliewkliew]

Thanks! Much appreciated