Jump to content

[old thread] Trajectories : atmospheric predictions


Recommended Posts

Alright, either I misunderstood this fully (quite possible) or I broke it (this is of course with FAR+DR):

Well like I said above entering backwards isn't supported at the moment unfortunately, but when you show entering pointing prograde it looks like you have the MJ cross aligned on the runway, not the marker from this mod so it looks like you probably ended up where this mod predicted.

@Alshain, yes parachutes would alter where you ultimately come down, however I don't think it is realistic to try and get this mod or any mod for that matter to try and predict exact landing location for a pod with chutes accounted for and everything. However if he is able to get the ability to calculate the reentry profile while reentering facing retrograde working, that will give an accurate enough prediction that you can compensate by placing the marker slightly ahead of the desired landing zone. With practice I'm sure you could get extremely accurate with targeted landing zones even if chutes aren't accounted for.

Link to post
Share on other sites

Wow, lot of posts since my last visit.

I am indeed working on this backward reentry thing, and also on various other sources of inaccuracies (if someone tried a Jool aerobraking at 10km/s, he knows what kind of inaccuracies, if he didn't die in the resulting unplanned atmospheric capture). Actually, it's kind of working now (code pushed on github), but not much tested. I'll try to release that soon.

And I confirm the red "cross" (with three branches) on the screenshots above does not come from this mod.

Link to post
Share on other sites
Wow, lot of posts since my last visit.

You got featured by the devs, that's what happens.

I am indeed working on this backward reentry thing, and also on various other sources of inaccuracies (if someone tried a Jool aerobraking at 10km/s, he knows what kind of inaccuracies, if he didn't die in the resulting unplanned atmospheric capture). Actually, it's kind of working now (code pushed on github), but not much tested. I'll try to release that soon.

Good to hear.

And I confirm the red "cross" (with three branches) on the screenshots above does not come from this mod.

I believe that is from MechJeb

@Alshain, yes parachutes would alter where you ultimately come down, however I don't think it is realistic to try and get this mod or any mod for that matter to try and predict exact landing location for a pod with chutes accounted for and everything. However if he is able to get the ability to calculate the reentry profile while reentering facing retrograde working, that will give an accurate enough prediction that you can compensate by placing the marker slightly ahead of the desired landing zone. With practice I'm sure you could get extremely accurate with targeted landing zones even if chutes aren't accounted for.

I don't think it's realistic... in stock. It's very realistic to imagine that in FAR. If you know when/where your chute will open and what kind of braking power it has, and if and when it will be cut (prior to landing), it's all just math after that. I'm not saying it needs to be top priority by any means, but it can be done, especially with the RealChute mod if your arming the parachute to auto-deploy.

Edited by Alshain
Link to post
Share on other sites
Wow, lot of posts since my last visit.

I am indeed working on this backward reentry thing, and also on various other sources of inaccuracies (if someone tried a Jool aerobraking at 10km/s, he knows what kind of inaccuracies, if he didn't die in the resulting unplanned atmospheric capture). Actually, it's kind of working now (code pushed on github), but not much tested. I'll try to release that soon.

And I confirm the red "cross" (with three branches) on the screenshots above does not come from this mod.

that Red Cross looks like a mechjeb landing target marker.

Link to post
Share on other sites
Wow, lot of posts since my last visit.

I am indeed working on this backward reentry thing, and also on various other sources of inaccuracies (if someone tried a Jool aerobraking at 10km/s, he knows what kind of inaccuracies, if he didn't die in the resulting unplanned atmospheric capture). Actually, it's kind of working now (code pushed on github), but not much tested. I'll try to release that soon.

And I confirm the red "cross" (with three branches) on the screenshots above does not come from this mod.

Thanks for your work!

I used the three-branches red cross once the cross of your mod has moved far away (you see it on the right side) to show where it should have stayed since I was oriented prograde. I started both tries with the cross of your mod exactly on the runway.

Should have stated that at the beginning, sorry for the confusion (yes it's the MJ one).

Link to post
Share on other sites
You got featured by the devs, that's what happens.

Yeah, didn't see that coming, and also didn't thought the mod would have so much success when I created this thread. Well, I'm glad if it helps :-) And thanks everybody for the support and kind words.

If you know when/where your chute will open and what kind of braking power it has, and if and when it will be cut (prior to landing), it's all just math after that. I'm not saying it needs to be top priority by any means, but it can be done, especially with the RealChute mod if your arming the parachute to auto-deploy.

I don't think chutes are a priority indeed, because most of the time you can deploy them at the last moment, so they don't dramatically change your landing site. Also, don't imagine prediction is more accurate than it is, it's kind of subject to the butterfly effect (chaos theory), so depending on your craft, your piloting, and starting conditions, prediction can be quite far from where you'll actually land. Especially with wings, but that's observable even with a pod. Probably farther than the difference you'll get by deploying your chutes at 100m of the ground.

Edited by Youen
Link to post
Share on other sites
Thanks for your work!

I used the three-branches red cross once the cross of your mod has moved far away (you see it on the right side) to show where it should have stayed since I was oriented prograde. I started both tries with the cross of your mod exactly on the runway.

Should have stated that at the beginning, sorry for the confusion (yes it's the MJ one).

OK, so it does look like either prediction was wrong, or it was accurate only for a very precise entry profile (see my previous post about chaos theory). Maybe it's just a plain huge bug, if you can send me material to reproduce that, such as your .craft file (preferably using stock pieces or at least not too many mods) I'll take a look.

Link to post
Share on other sites
OK, so it does look like either prediction was wrong, or it was accurate only for a very precise entry profile (see my previous post about chaos theory). Maybe it's just a plain huge bug, if you can send me material to reproduce that, such as your .craft file (preferably using stock pieces or at least not too many mods) I'll take a look.

Thanks, sent you a PM.

Link to post
Share on other sites

I went ahead and tested your latest code on the github repository for backwards (180 AoA) descents. Landings seem to fall consistently short by a tiny amount (~10km) (not using parachutes) but are still very close to initial prediction. It might just be piloting, but thought I'd share my result.

Edited by Czerky
clarity
Link to post
Share on other sites

Predictions are still just that: Predictions. The "butterfly effect" with FAR enabled can and will mess up your actual impact compared to AT's prediction. However, it does get you into the right neighbourhood quite well, which is admirable enough. Considering that ferram never wanted to do aerobraking predictions just due to the inherent inaccuracies, the result is really amazing.

The biggest strength however is the spaceplane guidance. A really nifty addition that helps with various types of powered landings not covered by MechJeb.

Link to post
Share on other sites

First, I wanted to say that I love this mod. I only just started using it this week, and though I find that the landing spot drifts a lot when descending into Kerbin's atmosphere, it works fine for landing on Mun and Minmus (I just saw the post about mounting a probe backwards, so I will try that and see if it fixes my Kerbin descent).

One suggestion:

A feature that would be useful would be to see the altitude of a point on a trajectory. Would it be possible to add a key-activated toggle that displays the altitude of when the mouse touches a trajectory or orbit line in map view? This would make it possible to see, for example, where my returning craft will be when the chutes auto-deploy at 20,000m.

Link to post
Share on other sites
Predictions are still just that: Predictions. The "butterfly effect" with FAR enabled can and will mess up your actual impact compared to AT's prediction. However, it does get you into the right neighbourhood quite well, which is admirable enough. Considering that ferram never wanted to do aerobraking predictions just due to the inherent inaccuracies, the result is really amazing.

The biggest strength however is the spaceplane guidance. A really nifty addition that helps with various types of powered landings not covered by MechJeb.

Which is why several times I've suggested in the MJ thread for Sarbian is that the correct solution to the FAR problem is to accept the inaccuracies for what they are. IRL, it's impossible to make the kinds of accurate predictions that players expect.

Link to post
Share on other sites

Version 0.2.2 is out. Some issues are fixed, and it allows setting an attack angle of 180° for backward reentries. It also occurred to me that you should keep correctly oriented, even if you don't have wings, FAR forces are not necessarily symmetrical, so keep the wings symbol of the nav ball aligned with the horizon just to be sure... For backward flight, it should be upside down (the orientation you'll have if you start prograde and make a half-looping)

Link to post
Share on other sites

Holy crap, I actually did it! I landed at the KSC in a capsule decoupled from orbit (and deadly reentry to boot). I also did it on the first try no less. And I think it's safe to say it was largely thanks to this mod.

Javascript is disabled. View full album

Now I will note that the accuracy wasn't perfect, I couldn't blindly trust it, as the ground-marker (cross) drifted quite a ways across the continent as I reentered (first drifted short of the target, then drifted back toward the target), but between my meager RCS fuel and some adjustment of my angle of attack, it was enough to hit my target regardless of the unexpected drifting. It's good enough to get the job done. I've waited a long time for a mod like this.

Some suggestions for the future:

  1. It would be nice if it could update your ship's drag, not based on the ship's current stage/state, but rather a future stage. That way you could get your orbiter into a desired trajectory without having to first decouple the capsule from it yet. (But I'm not even sure if this is possible from a technical standpoint, maybe not).
  2. Figuring out what causes the ground-target to drift would be nice. The accuracy proved to be good enough, but more accuracy would always be welcome to have.

Also I still have a question though: On the descent profile, what do "entry", "high", "low" and "ground" mean exactly?

Edited by Entropius
Link to post
Share on other sites
Also I still have a question though: On the descent profile, what do "entry", "high", "low" and "ground" mean exactly?

You can set your angle (either relative to the prograde direction, or relative to the horizon) for different altitudes. Say you're going to land with a spaceplane, maybe you want to descend fast (assuming you don't use deadly reentry), so you can set 0 or even a negative AoA for "entry" and "high", then give time to your plane to slow down by setting a small horizon angle for "low" (for example -5°), and finally set horizon to 0° for "ground" so that you arrive with a low vertical speed. The mod assumes your craft is oriented as specified during the prediction, and interpolates between each angle for intermediate altitudes. This is exactly what the square indicates on the nav ball if you set the impact position as target. The circle on the other hand adds an offset showing the general direction you should point at to correct the trajectory (if you have drifted from it) ; but that's only a general indication, you don't actually have to point at it (for example if it's way above the square, you can just point a little bit above the square which should make the circle go down slowly until you're back on a good trajectory). Oh, and I'm not responsible if you break your craft by using aggressive AoA, you can blame ferram4 for that though ;-)

Link to post
Share on other sites
Hi there,

i tested it yesterday, and i found out that the firespitter airbrakes are not accountet in the prediction. (It makes no difference if the brakes are set or not to the plotted tajectorie)

My guess is that this part adds a force of its own. If it's not part of the supported mods (which at this time are FAR and NEAR), you can indeed expect that it won't be used at all. I'll add that as a feature request on the bug tracker.

Link to post
Share on other sites

I've been looking for something like this for some time now as I've always been pretty awful at predicting where I'm going to land :)

I had a quick go with it last night sending a simple 1 man pod+chute for reentry and trying to hit the KSC. Set up for 0 angle of attack at all ranges and followed the velocity vector in (bottom of capsule first) and landed on the other side of the mountain range. Is this because it was expecting me to go in top first?

Also tried out using the stock explorer that comes with SP+ and got no trajectory prediction at all no matter what I set. Any ideas on what to do? I'll try a fresh install when I get some free time to see if that will help.

Anyway thanks for this great mod :)

Link to post
Share on other sites
Youen, check out Diazo's Vertical Velocity controller mod for a great example of changing "forward direction" in-flight.

http://forum.kerbalspaceprogram.com/threads/50736-0-24-2-(Jul-19-14)-Automate-Vertical-Velocity-and-Altitude-Control

I don't think that will be necessary now that the angle can be set from -180 to 180. It would actually be more complicated than just saying where is "forward", because you also need to say where is "up" : flying with wings vertical is not the same as wings horizontal (and I'm sure it has an impact even if you don't have wings). At this time, you can use the "control from here" to change the reference frame (should work on command modules and docking ports), and the angle from -180° to 180° arround the "right" axis. But if you have a use case where it's not enough, let me know.

Link to post
Share on other sites
Hi there,

i tested it yesterday, and i found out that the firespitter airbrakes are not accountet in the prediction. (It makes no difference if the brakes are set or not to the plotted tajectorie)

My guess is that this part adds a force of its own. If it's not part of the supported mods (which at this time are FAR and NEAR), you can indeed expect that it won't be used at all. I'll add that as a feature request on the bug tracker.

They do not. Firespitter Airbrakes operate by adjusting the part's maximum_drag and minimum_drag properties. The stock drag system does the rest.

Link to post
Share on other sites
This thread is quite old. Please consider starting a new thread rather than reviving this one.

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...