# Why are the ISS attitude values shown on these two web sites so different?

Here are screen captures from two sites that seem to be showing approximately live - or at least changing - telemetry from the ISS. They both say the spacecraft attitude is shown in degrees.

Looking more closely with fresh data, there is a relationship between the two sets of attitudes, as if they had different definitions of the spacecraft axis. The first one has Roll: -175.93 while the second shows Yaw: -4.08 - they sum to 180.

note: the situation has been status quo for at least the past two weeks, it's not a transient.

25-Feb-2017 06:18 UTC

## below: from http://www.telemetry.space/

25-Feb-2017 06:18 UTC

• I'd change the title to "Why (actually) are the..." but I promised to stop using the 'a-word'. I am looking for the reason these two sites are different. Is one of them wrong? – uhoh Feb 25 '17 at 7:00
• This looks rather weird. Looking at the current value, pitch is the same, while yaw and roll are swapped and measured in opposite directions (one has an offset of 180°). On the other hand, Error and Rate are identical on both pages - this doesn't make sense to me. – asdfex Feb 25 '17 at 12:09
• I'd like to finally get this sorted/corrected. Unfortunately the above site, made by Tietronix, is no longer functional so I can't use that as a comparison point. Since this is really a multipart question, I'm not really sure how or which Stack site to ask on. 1. What are the current, correct yaw, pitch, and roll values for the ISS (for verification). 2. How are the quaternion "components" 0–3 in NASA's USLAB000018–21 supposed to be translated into a quaternion and/or Euler angles. What do you guys think: ask those here, on StackOverflow, split across both, or neither: ask NASA (again)? – Yuri Gadow Jul 26 '17 at 14:31
• @YuriGadow hello again! It certainly can't hurt to ask a question here any time, and the atmosphere here is quite helpful and constructive so if the question should be changed or broken apart, people may just make such recommendations and you can edit accordingly. In SO things move very quickly so if it's not really on-topic it won't be so well received. The other thing you could do would be to ask in the [meta section](space.meta.stackexchange.com) of this site and tag it with discussion. That's where questions about questions are usually asked. – uhoh Jul 26 '17 at 14:46
• Thank you @uhoh, I've created a new question, hopefully making it clear the programming bit is secondary to data interpretation bit! space.stackexchange.com/questions/22413 – Yuri Gadow Jul 26 '17 at 16:11

The conversion of the attitude data as it appeared (undocumented) from Lightstreamer to roll, pitch, and yaw was the biggest pile of guesses and assumptions atop assumptions I made back when Matt and I were coding this. It is almost certainly wrong if it is varying from the ISS Live! website.

We've been working on a complete rewrite of the code and it'd be nice to actually understand what the data format is and how to properly convert it, rather than just throwing formulae at the data until it sorta fits; I'll have a look through newly found documentation and review that code again.

Feel free to chat with us in Slack or comment on the card: https://trello.com/c/NLeP5EXq/57-attitude-values-differ-from-isslive

• @uhoh My website? 'Tis open source. Anyway, the math I can likely sort, what I need to find time and patience for is to again try to reverse engineer the data from Lightstreamer / NASA, I had found no documentation on what that is and I'd rather not go chasing errors in quaternion math if it's not even a quaternion or I've just guessed the attributes' order wrong. It's all open source, and we are happy to help with rough pull requests—jump in if you're interested in this. – Yuri Gadow May 10 '17 at 18:53

The ISS Live page is using the ISSACS system (ISS Analysis frame) which is the normal coordinate frame for this data.

Just doing the transforms in my head, it looks like the other web site is using a frame where the Y axis is the same as ISSACS, but the frame is pitched so that +X points to nadir and +Z points to ISS aft. Thus pitch is the same in both frames, but ISSACS +roll becomes -yaw in the new frame, and ISSACS +yaw becomes -roll.

For a given data snapshot

ISSACS / "Pitched ISSACS"

Yaw -4 / -.55

Pitch -1 / -1

Roll .61 / -176

GIMPed picture of the "Pitched Frame"

• I'm not so interested in the "how" as I am the "why". What is GIMPed? Is this a legitimate, accepted, alternate coordinate system for the ISS, or just a bug in their quaternions or other math? (I peeked at the code a little bit). – uhoh Feb 25 '17 at 17:17
• I've no idea. The "pitched frame" is counter-intuitive for me, but maybe it makes sense to someone. – Organic Marble Feb 25 '17 at 17:29
• That "X' axis choice on telemetry.space really looks like they are assuming ISS is not orbiting but falling out of the sky directly down.... – SF. Feb 26 '17 at 1:00
• Thanks @OrganicMarble, some actual documentation—even if not related directly to the data format Lightstreamer sends—is most welcome. I’ll be reading through the doc you linked when I check over the code again! – Yuri Gadow Mar 14 '17 at 12:23
• @asdfex I don't :) But, the reason is likely that the current attitude numbers come from Lightstreamer bundled in a single quaternion (that was my best guess anyway) while the error and rate numbers are separate records that we didn't have to guess at how to pull apart and convert. – Yuri Gadow Mar 14 '17 at 12:44