[Question] GPS Acuracy?

I added gps locate into my program that is able to continue after the program has been stopped, and every time requests the cords it goes haywire and moves to a weird spot but If I use a non wireless turtle and manually enter the cords it continues just fine. Is GPS locate accurate or is the coordinates outputted different then minecrafts?

https://github.com/H...vancedorefinder

gps is 100% accurate, else it errors
you probably entered the coords wrong for the gps host
or your using reis minimap wich confuses the Y and Z coords

How do you know what to round the cord to because I was using f3, I might have got the cords wrong because of it being near or on 0

Nvm the GPS api is stupid it has the cords mixed up it uses z for the height instead of y

No, it doesn't. If you read BigShinyToys' tutorial, forget everything he said about the coordinates.

GPS API uses whatever coordinate system you give it. All it does is perform some mathematical operations on the location and distance information it receives. You could use X for the vertical coordinate, it wouldn't care. Make sure you set up your GPS hosts with the x, y, z coordinate system you want them to use.

I have it working correctly now, but sometimes when you start it the cords are off my one block anyone know why?

Henness, on 18 December 2012 - 05:33 AM said:

I have it working correctly now, but sometimes when you start it the cords are off my one block anyone know why?

Yes, are your gps hosts placed close to each other? Like in BigSHinyToys' tutorial? Because if that's the case, it's possible that there are some rounding errors when you're far away. That's because the distances between host1 and client and between host2 and client are barely different.

Orwell, on 18 December 2012 - 05:39 AM said:

Henness, on 18 December 2012 - 05:33 AM said:

I have it working correctly now, but sometimes when you start it the cords are off my one block anyone know why?

Yes, are your gps hosts placed close to each other? Like in BigSHinyToys' tutorial? Because if that's the case, it's possible that there are some rounding errors when you're far away. That's because the distances between host1 and client and between host2 and client are barely different.

Ya that's how I have it set up. When I was setting it up what I was thinking there could be a problem with that. Because I've done some triangulation to find strong holds before and I was allays off by about 50-100 if you don't get points far enough apart the error gets really big.

This does beg a question then. If BigShinyToys' tutorial is not good, what is a GOOD tutorial or what are the real optimal formation of the GPS hosts? Also, what happens if I add a 5th host and a 6th etc.. Theoretically the accuracy should be going up with every new sampling point of reference, but I don't know if the GPS code even recognizes more than 4 hosts.

Helfire, on 18 December 2012 - 06:46 AM said:

This does beg a question then. If BigShinyToys' tutorial is not good, what is a GOOD tutorial or what are the real optimal formation of the GPS hosts? Also, what happens if I add a 5th host and a 6th etc.. Theoretically the accuracy should be going up with every new sampling point of reference, but I don't know if the GPS code even recognizes more than 4 hosts.

1. I believe his tutorial is good, it's just a minor catch.
2. You get higher accuracy from:
[indent=1]a. Placing them further away from each other.
b. Using an asymmetric set up.
c. Using more hosts.[/indent]
3. Yes, I seem to remember that the gps code handles more than 4 hosts.

My biggest problem with his tutorial is the decision to use an unintuitive and arbitrary coordinate system (that he happens to personally prefer) rather than basing it off of the system Minecraft uses. This adds confusion for people using his tutorial putting in coordinates from the Minecraft system, then expecting to get sane coordinates out of gps.locate.

Lyqyd, on 18 December 2012 - 07:08 AM said:

My biggest problem with his tutorial is the decision to use an unintuitive and arbitrary coordinate system (that he happens to personally prefer) rather than basing it off of the system Minecraft uses. This adds confusion for people using his tutorial putting in coordinates from the Minecraft system, then expecting to get sane coordinates out of gps.locate.

I agree, and Minecraft's axis choice is just fine for euclidean space, so I don't see the problem with y being height.

I put my gps hosts (actually rnhst hosts, but they serve gps to authorized IDs) at +/- 2 on the x and z, and get perfect accuracy out to the standard rednet range limit.

BST's tutorial is fine, as long as you use a consistent coordinate system. But you'll be better off using a program that sets up the array automatically so that it's always consistent even if it's wrong.