Unpiloted Spaceplane Landing From Orbit

[Superfluous J]

A conversation in this thread gave me this idea. I thought landing a space plane without controlling it was impossible. I am apparently wrong. Now I'm curious...

The Winners!

[table][tr][td]Vertical Landing (Parachute replacers)[/td][td]Horizontal Landing (gliders)[/td][/tr][tr][td]1st Place! - kujuman - 9 parts[/td][td]1st Place! - sploden - 9 parts

2nd Place - zorque - 10 parts

3rd Place - bsalis - 63 parts[/td][/tr][/table]

The Challenge

Build a space plane that can land from orbit with NO INPUT FROM THE USER. The plane can have Kerbals and/or probes on it, but you can't touch your controls once you've entered the atmosphere. The winner will be the plane with the smallest number of parts that can consistently land on flat terrain.

Rules/Restrictions

1) 100% stock parts. You can use any parts you want EXCEPT PARACHUTES. You don't even need to use wings though I think they're probably required...

2) No mods that modify the physics of the game.

3) The plane must start in an at least an 80km (periapsis) orbit around Kerbin. Hyperedit may be used to put your ship into this 80+ km orbit around Kerbin. Alternatively, you can use any lifter you choose to get your plane into space, or even use Alt-F12 to cheat your plane up there with no gravity and free fuel (Just turn gravity back on before starting!). Any lifter you DO use doesn't count towards your parts, only the number of parts on the plane that enters the atmosphere counts. This is not a "reach space" contest, it's a "survive re-entry" contest.

5) You can control the plane in ONLY two ways, ONLY until it is no longer in space (below 70km):

5a) To slow it down (turning and thrusting retrograde)

5b) Aligning for atmospheric entry (but only ABOVE 70km)

6) You may not decouple, undock, or disconnect any pieces once your periapsis drops below 80km. The plane in orbit must be the plane that descends and lands.

7) As soon as the plane drops below 70km, take your hands off the controls and hope it lands.

8) You can retry if your plane hits water or a mountain (Aim for the big continent or possibly the poles) but don't retry 50 times hoping to get lucky. This is a design challenge and the results should be reproducible. If any pieces fall off or explode, your plane FAILED to land and you must try again.

9) To keep the spirit of the challenge, no Kerbal can get out and push at any time during this challenge. It's an interesting idea, but there are to many questionable loopholes it raises.

10) Post your designs, either as a ship file or as a really good picture of the plane.

The plane that can consistently land on flat ground, that has the lowest number of parts, wins! In case of a tie, the plane posted first wins.

(this is my first challenge, so please let me know everything I did wrong so I can fix it here and not make the mistake again in the future!)

REVISION HISTORY:

I rewrote rule #6 as it was poorly worded. The OLD Rule #6 is below:

6) You may not decouple, undock, or disconnect any pieces once you've left the 80-90km orbit. The plane in orbit must be the plane that descends and lands.

I also added a new Rule as #9, and moved the old #9 to #10.

I rewrote rule #3 to remove the 90km upper limit. Upper limits on the orbit only make this easier. If you want to start from a 100x100km orbit, go for it! The OLD Rule #3 is below:

3) The plane must start in an at least an 80km (periapsis) and at most 90km (apoapsis) orbit around Kerbin. Hyperedit may be used to put your ship into this 80-90km orbit around Kerbin. Alternatively, you can use any lifter you choose to get your plane into space, or even use Alt-F12 to cheat your plane up there with no gravity and free fuel (Just turn gravity back on before starting!). Any lifter you DO use doesn't count towards your parts, only the number of parts on the plane that enters the atmosphere counts. This is not a "reach space" contest, it's a "survive re-entry" contest.

Edited December 15, 2013 by 5thHorseman

[Hodo]

Is it even possible? I mean it can be done with KOS, Mechjeb or some other flight computer but just letting it coast down and land on its own. That maybe impossible.

[Pds314]

Interesting... I have a few ideas:

Using a sufficient number of wings in the correct orientations, use "lift" to cause the drag on the plane to slow it down to landing speed.

Use an absurdly tough lander made out of girders, I-beams and structural pylons.

Use a combination of the above.

Use a basic jet that provides not quite enough thrust and will always face down thanks to the aerodynamics of the plane to effectively land it in arbitrarily low gravity.

Land it directly into a mass-driver rocket to slow it down (stupendously tough to do).

[BagelRabbit]

I'm debating whether to:

a) land a spaceplane by pointing it at the horizon, turning on the SAS and hoping for the best, or

have a ballistic lander with plenty of pylons and girders.

Both are viable, and I'll test them both out... I have confidence that at least one will work.

[Kasuha]

Is there any requirement on descent profile? Does it have to be standard deorbiting, or is it ok if the requirement is that it kills all its horizontal speed before entering atmosphere?

... that's just pure theoretical question, I don't have any particular design that would require it on mind.

[Superfluous J]

I'm debating whether to:

a) land a spaceplane by pointing it at the horizon, turning on the SAS and hoping for the best, or

have a ballistic lander with plenty of pylons and girders.

Can a craft made out of pylons and girders reach the ground and not destroy those pylons and girders? The entire craft must survive re-entry AND landing.

Is there any requirement on descent profile? Does it have to be standard deorbiting, or is it ok if the requirement is that it kills all its horizontal speed before entering atmosphere?

You can do anything you want so long as you don't control the ship once it drops below 70km, and so long as you don't decouple anything once the periapsis drops below 80km. So if you use a ton of fuel to kill your velocity to 0, you gotta bring that tank down with you. Or decouple it while it's still burning in a reproducible way which... well good luck with that.

[Kasuha]

So if you use a ton of fuel to kill your velocity to 0, you gotta bring that tank down with you. Or decouple it while it's still burning in a reproducible way which... well good luck with that.

Technically it's about as hard as landing on Tylo - instead of landing on real surface, you land on imaginary surface 80 km above Kerbin sea level.

And since it is clearly possible I don't see much reason for actually doing that if I can lift my probe vertically and just deploy it above 80 km.

According to my experiments it's not very important how you deploy that ship anyway.

[Superfluous J]

Technically it's about as hard as landing on Tylo - instead of landing on real surface, you land on imaginary surface 80 km above Kerbin sea level.

And since it is clearly possible I don't see much reason for actually doing that if I can lift my probe vertically and just deploy it above 80 km.

According to my experiments it's not very important how you deploy that ship anyway.

I'm not sure I understand what you're saying? Lifting the ship/plane/whatever to 80km isn't enough. Before you start this whole thing, you must be in ORBIT with a periapsis above 80km.

That's rule/requirement #3 above.

Edited December 11, 2013 by 5thHorseman

[kujuman]

hummmm....

I'd suggest maybe trying a helicopter type (or seedpod) landing, but that might require FAR to actually make it work correctly.

I'd also advocate allowing FAR as a separate category, as the "lift" physics in stock are awful.

comes back after proof of concept...

Ok, done, mostly. 7 parts + 2 more for tiny OMS, probably can strip it down a few more parts.

completely stable, so no need to test from reentry. slows to ~9m/s at ground level. still a bit too fast for not breaking the winglets off, but another few test runs should solve it.

Update 2: currently testing w 3 adv canards to save on parts. might be able to get down to 2 and use a sepretron from deorbit burn, meaning a total part count of 4

Edited December 12, 2013 by kujuman

[makinyashikino]

Abyssal Lurker once landed a plane in one of his videos without any input, it doesn't meet the guidelines of not having pieces fall off, but it does sort of prove that if designed better, can be done. It's here if you want to see it.

[Kasuha]

I'm not sure I understand what you're saying? Lifting the ship/plane/whatever to 80km isn't enough. Before you start this whole thing, you must be in ORBIT with a periapsis above 80km.

That's rule/requirement #3 above.

My argument is that difference between "being at orbit between 80 and 90 km" and "sitting 80 km above surface with 0 surface or vertical speed" is a few tons of fuel which can be decoupled without breaking any rules. If I am allowed to hyperedit my ship to that initial orbit, I should be allowed to hyperedit (or lift) the ship to that sitting duck position as well. Because that makes no difference on the ship's final ability to land, it'll only save some time.

[Superfluous J]

I'd also advocate allowing FAR as a separate category, as the "lift" physics in stock are awful.

I'm not against that. Consider it a thing and I'll update the first post when I have a bit more time on my hands.

Abyssal Lurker once landed a plane in one of his videos without any input, it doesn't meet the guidelines of not having pieces fall off, but it does sort of prove that if designed better, can be done. It's here if you want to see it.

Thanks for that!

My argument is that difference between "being at orbit between 80 and 90 km" and "sitting 80 km above surface with 0 surface or vertical speed" is a few tons of fuel which can be decoupled without breaking any rules. If I am allowed to hyperedit my ship to that initial orbit, I should be allowed to hyperedit (or lift) the ship to that sitting duck position as well. Because that makes no difference on the ship's final ability to land, it'll only save some time.

I'm not comfortable saying you can consistently do it, but I'm willing to consider it. The rules are complicated enough and it seems that to allow for this I'd have to add several more rules.

Something like:

Alternatively, you can use HyperEdit to simply place yourself at a height of 80-90km over Kerbin in a sub-orbital trajectory of any kind. If you do this, however, you may not touch the controls at all after you've used HyperEdit.

[antbin]

I'd try for this challenge, but I reckon any aircraft that will glide slowly at sea level in KSP's drag model will probably glide slowly at high altitude as well. And I don't have the patience to wait to descend at 10 m/s from 20,000m.

[kujuman]

Can a Kerbal get out and push to lower Pe for a min theoretical 3 part lander?

Edit: Completed challenge with 9 part ship, no kerbal pushing.

Javascript is disabled. View full album

Edited December 12, 2013 by kujuman

[Sirine]

I'd try for this challenge, but I reckon any aircraft that will glide slowly at sea level in KSP's drag model will probably glide slowly at high altitude as well. And I don't have the patience to wait to descend at 10 m/s from 20,000m.

Agree. After 30mins of waiting. End up landing on water, or worst; hit a mountain.

[Blaster]

This would be a good challenge to try making a non-parachute drop pod with.

[Kasuha]

Okay, did it with eight parts. In the end I believe I followed your rules right - standard descent, last decouple above 80 km, at 70 km I was already out of battery power so I couldn't do anything even if I wanted. The only thing I did was to give it some rotation to make sure the right end enters the atmosphere first.

Funny thing was that it was slowing down so efficiently I did not see any reentry effects at all.

And yes, it was rotating the whole descent. That's not very apparent from still images.

Javascript is disabled. View full album

Edited December 12, 2013 by Kasuha

[Superfluous J]

I'd try for this challenge, but I reckon any aircraft that will glide slowly at sea level in KSP's drag model will probably glide slowly at high altitude as well. And I don't have the patience to wait to descend at 10 m/s from 20,000m.

The two submitted entries got to the ground in about 5 minutes from entering atmosphere.

Can a Kerbal get out and push to lower Pe for a min theoretical 3 part lander?

Hm. I think that meets the rules but I honestly didn't think about it. Let me think some more.

Edit: Completed challenge with 9 part ship, no kerbal pushing.

Our first winner!

Okay, did it with eight parts. In the end I believe I followed your rules right - standard descent, last decouple above 80 km, at 70 km I was already out of battery power so I couldn't do anything even if I wanted. The only thing I did was to give it some rotation to make sure the right end enters the atmosphere first.

Funny thing was that it was slowing down so efficiently I did not see any reentry effects at all.

And yes, it was rotating the whole descent. That's not very apparent from still images.

Sadly this does not count. You decoupled - I believe - AFTER your periapsis was under 80km. This breaks rule #6 (which I think I could have worded better):

6) You may not decouple, undock, or disconnect any pieces once you've left the 80-90km orbit. The plane in orbit must be the plane that descends and lands.

If your periapsis is below 80km you've left the 80-90km orbit. I should have used the word periapsis and for that I apologize. I have modified the first post to make this more clear.

Edited December 12, 2013 by 5thHorseman

[Superfluous J]

I'm not comfortable saying you can consistently do it, but I'm willing to consider it. The rules are complicated enough and it seems that to allow for this I'd have to add several more rules.

Something like:

Alternatively, you can use HyperEdit to simply place yourself at a height of 80-90km over Kerbin in a sub-orbital trajectory of any kind. If you do this, however, you may not touch the controls at all after you've used HyperEdit.

I considered this and I've decided that no, I don't like it. If you want to do a trick to decouple a still-burning engine to slow yourself down even further, go ahead. However consider that an additional challenge to get what you want while remaining within the rules that really intended you to enter the atmosphere at near-orbital velocity.

[zorque]

I refuse to do a heli-spinner for this. That's not the spirit of the challenge to me.

Here's an actual plane that can do it with a horizontal landing.

Key steps: Boost into orbit at ~100 KM, separate all but lander and retro burn package, retro burn at beginning of big ocean before KSC, ditch retro burn pkg, line up to 30 degrees down at 80KM, let go (forgetting to set brakes).

Steps detailed in pics titles.

-Mike

EDIT: Sorry, 28 parts. Wasn't really trying for part cound, just do-ability.

Javascript is disabled. View full album

Edited December 12, 2013 by zorque

[kujuman]

Even if we can't use Kerbals to get out and push, I have an improvement by one part. Calculations show I have enough delta-v by replacing the tank+ant with a sepratron that I can still complete the deorbit, assuming that the figures given by the KSP orbit mechanic are accurate (http://forum.kerbalspaceprogram.com/threads/4374-KSP-Orbit-Mechanic-1-2a-Optimize-Your-Orbits)

I might bother to build and fly it, but it's pretty much assuredly going to work. Also waiting on final ruling on getting out to push before I fly again. Low altitude testing shows that the craft is just as stable without the fuel tank and engine on it though.

For everyone attempting, a circular 81km orbit to 81km x 40km requires about 35.6 m/s. 81x81 -> 81x55 requires only a 22.3 m/s burn. So for the first orbit a single septratron will work if the rest of the ship is about 1.49t or less, the second your ship can be up to 2.51t

[kujuman]

I refuse to do a heli-spinner for this. That's not the spirit of the challenge to me.

Neat. I figured this was doable as a plane, but eh, too much work

I'd actually consider this the first success, or at least separate the "planes" vs the "parachute replacers" aka heli-pods

[Sirine]

I refuse to do a heli-spinner for this. That's not the spirit of the challenge to me.

Here's an actual plane that can do it with a horizontal landing.

Agree. Like your design.

[Superfluous J]

I refuse to do a heli-spinner for this. That's not the spirit of the challenge to me.

Here's an actual plane that can do it with a horizontal landing.

Key steps: Boost into orbit at ~100 KM, separate all but lander and retro burn package, retro burn at beginning of big ocean before KSC, ditch retro burn pkg, line up to 30 degrees down at 80KM, let go (forgetting to set brakes).

While your plane is far closer to what I envisioned making this challenge, it actually also breaks rule #6 that I obviously worded horribly. You decoupled your retro burn package after your periapsis dropped below 80km (but while your plane was still over 80km, which the old wording of rule #6 implied).

You've inspired me to create 2 winner ladders, though, one for vertical landing and another for horizontal (rolling) landing, because they are very different things that require very different approaches.

I might bother to build and fly it, but it's pretty much assuredly going to work. Also waiting on final ruling on getting out to push before I fly again. Low altitude testing shows that the craft is just as stable without the fuel tank and engine on it though.

I modified the rules, adding a "no kerbal EVA to push" rule. It is fraught with loopholes so I figured I'd just nip it in the bud.

Using a decoupler to give you the final push to atmo is fine, so long as your periapsis is 80,000 meters or more when you do it. It may take a few orbits to come down, though

Edited December 12, 2013 by 5thHorseman

[kujuman]

Using a decoupler to give you the final push to atmo is fine, so long as your periapsis is 80,000 meters or more when you do it. It may take a few orbits to come down, though

Can I claim 7 parts and then show proof in about 2 hours? Because I know decoupler spam