A question about CoordinateConversions about open-dis-java

[icemagno]

I see. Then I think we can close. Many thanks and BRAVO ZULU.

[leif81]

Hi @icemagno

There is a unit test here that does the round trip check. A small amount of acceptable error had to be identified to get it to pass. I did not see an issue with elevation though in the sample data. Maybe you found a corner case?

https://github.com/open-dis/open-dis-java/blob/master/src/test/java/edu/nps/moves/disutil/CoordinateConversionsTest.java

[leif81]

Do you mean this value? -9.313225746154785E-10

That's a very small number right, close to 0.

[leif81]

I plugged your values into the unit test in PR #72 and it passed.

Maybe i'm overlooking something, but I think the coordinate conversion is working OK.

[icemagno]

Well I think -9 meters is a little far from zero. In my case I have to simulate airplanes and this may crash someone in land or a mountain. Or just cause a missile to miss the target.

Im using RPR FOM and data comming from other source as LAT, LON, ALT. So I must convert it to Spatial data (XYZ) and send to RTI. Then I need to get it back to the original values in others federates and this is when I get the different values. So my original source tells me the airplane is landed at zero meters and the federates understand -9 meters for example.

I've checked others values and the error is different every time. For example there is no error with elevation 10. You can try a loop passing values from zero to 100 and check the recalculated value to verify.

My test case was just convert to XYZ and back again to LAT LON ALT using the output values.

The coordinates are ok indeed but depending on the elevation value passed as source we may get less or more error in decimals leading to the "GPS imprecision effect" when the object position is not accurate in small scales.

I mean... if a helicopter is standing at ground and constantly passing its position as LAT,LON,ALT to be converted to XYZ and then converted again we may see the helicopter changing its position when we change only its elevation. I'm working on military simulations and don't know if these small errors are allowed.

If you tell me it is ok then I'm fine. Sorry about my english. You can ask me to clarify if you don't get something.

[leif81]

Unless I'm misreading the value, the elevation is not -9, it's -0.0000000009. It's presented in scientific notation.

[leif81]

If you want to be certain, plug your values into the unit test and see if it passes.

[joshserrin]

Sounds like your federate code is parsing -9.313225746154785E-10 as -9 instead of -9E-10. Doesn't appear to be an issue with open-dis.