Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[LoSboccacc]

it does has limitations, but sound miles ahead of 'things in water to try docking'

[Graknorke]

You ran out of electricity, miauw xD

Now, today I had, as part of my volunteer job, patrolling (=wandering around looking useful) the aerospace part of the science museum. As I was looking at the pitiful excuses for interactive booths, I thought KSP would be so perfect as a set of kiosks.

Eg, there was a 'Dock with ISS!' thing that was three shapeless plastic lumps in water, controlled with a joystick. There was also a 'launch a rocket!' thing that was nothing but a button, then autopilot. :/
I see KSP doing an awesome job (with some modifications for kiosks, like a reset ability, multiple cameras at a time to reduce learning curve for docking....) in the museum, making rocketry fun and alive, not the boring and dead panels of generic text, the bland models of lunar landers...
Actually, I saw actual RCSes on the lander today. They had chutes to prevent the propellent from damaging thhe ship.

TLDR: KSP FOR AEROSPACE MUSEUMS! PLEASE! D: It's even scientifically accurate!

Somewhat accurate. Can only model one source of gravity at a time.

I'm not sure that that's correct.
Else entering a different body's SOI wouldn't be as smooth as they are.

[LoSboccacc]

yes it is correct.

SOI interface is a change in force, so it won't change trajectory abruptly.

[Skyrunner]

I think the 'one body at a time' is an unsolvable technical limitation. O.o
I remember Wiki saying some smart guy prove you can't predict the course of an orbiting object influenced by more than one gravity field at a time.

[Pyre]

That and it wouldn't add much to game play even if you could predict your trajectory. Sure you get Lagrangian points, but most people wouldn't be able to take advantage of them, and even if you could, your missions to other planets would take (in-game) decades instead of years.

[Putnam]

I think the 'one body at a time' is an unsolvable technical limitation. O.o
I remember Wiki saying some smart guy prove you can't predict the course of an orbiting object influenced by more than one gravity field at a time.

This is why universe sandbox freaked me the hell out when I realized what it is.

[Skyrunner]

Why would it? O_o I'm sure the Universe Sandbox does the math in an n-body simulation... doesn't it?

[DrPoo]

No, it uses magic.

[Starver]

I think the 'one body at a time' is an unsolvable technical limitation. O.o
I remember Wiki saying some smart guy prove you can't predict the course of an orbiting object influenced by more than one gravity field at a time.

I think the big problem with the Three[1]-Body problem is integrating the effect gravity of each component upon each other for any progression of 't'.

By making two 'minor' assumptions (to whit, that user craft/debris does not significantly impinge on any gravity well[2], and that you ignore the potential for cumulative perturbations in all natural bodies' orbits[3]) you could keep a decently realistic orbital ballet going but can now abandon the idea that a spacecraft is only under the influence of one body at a time.  Let the spacecraft 'suffer' a true summation of forces from every 'permanent' body in the system, while safely ignoring the inconsequential reciprocal dynamics.  This way it would be able to take advantage of Lagrange Points and perhaps add some additional (or at least less abstracted) slingshotting possibilities.

It's still a shifting map of gravity gradients that could tax the "projected orbit" calculator, and perhaps a compromise approximation of making it the two[4] most significant bodies to the craft (for any given region of space) would give practical (if not exactly accurate) L-points while reducing the pressure on the path calculator.


But I'm sure each level of compromise has already been considered by the developers, already.  (I don't frequent their forums, so I'm in the dark about any actual Word Of God statements on this subject.)



[1] Or any N, where N>2.

[2] Or you could fade off such an effect at 1km or so, so that a minute force might eventually draw pretty-much relatively unmoving items together when in a particularly slow part of a sufficiently eccentric orbit, but ignore it otherwise.

[3] If there's a definite resonance between orbital periods, then perhaps a non-elliptical orbit could be hard-coded, perhaps even a 'horseshoe orbit', like some actual solar system asteroids and such) but everything's still on 'rails' for all practical purposes, just slightly modified rail-paths.

[4] Or maybe three.  The actual closest body, the body around which this first orbits and the closest satellite of this first body.  But the closest body and whichever is most significant (assuming a choice) between {body1.parent && body.child.nearest} could give you the best two from which Lagrange Points could form from, albeit then without perturbation from additional satellites of the parent in the chosen pair or from the 'grandparent', if any.  (Here I'm assuming the current distribution of system masses stays much the same as the current layout, without any supermassive outer moons able to supersede over ultra-minor inner moons, at any point...  But then such a situation would probably make problematic my basic assumption that one could 'railroad' the orbits, and wilfully remain blind to the more realistic possibility of system instabilities.)

[LoSboccacc]

The main problem is that you can't apply force for each frame that way.

It works in orbiter, but it breaks as soon as you add warp effects.

Semplifing to a single body means they could just store the orbital parameters and put the craft on rails. After the orbit is on rail they can get the precise craft position at any time in past of future whitout having to calculate the effect of forces at each intermediate step

[sneakey pete]

Yes, the warp and the orbital predictions are the reasons that they use a SoI system, not because its impossible to code for a n body problem

[Starver]

Not that I've seen much need for a calculation of a craft's past position in the current iteration...  There's the 'already passed' half of any plotted orbital path but right now that could just as easily be the segment from +0.5 to +1 orbits 'forward' (ignoring any encounter/dis-encounter transitions) as extrapolating backwards.

The same treatment (tied to the moving volume of space[1] of the same two (or three) most significant influences) might not be able to produce something similar (not always being able to create a true closed orbit) but extrapolating backwards could be done for that, the same as extrapolating forwards, at least to a point where (if one ignored change of frames) an equal amount (by time and/or distance) of forward orbital and rearward orbital brings the tracks to the same celestial 'longitude' from the primary orbital body, even if they don't meet there.  The transition point to now being in two(/three) different influences (losing one moon and gaining another, or gaining the sun instead of one/vice-versa) could probably be similarly treated as the single-change-of-influence transitions as of now, i.e. do not draw the path beyond that encounter-point and instead draw the new 'continuation' on the moving frame that is destined to meet the current one.

(I can see what I'm explaining quite clearly in my mind's eye, but I'm not sure it's properly survived the transition to explanatory text.  Apologies if I've made the result too convoluted, but I think you should be able to work out what I mean, even if you don;'t agree with it..!)

But, yes, the maths is more complex to create the 'railroad' for coasting (and warping) on. Still simpler than adding a propulsive delta-V component to the predicted path.


Anyway, let it be the way it is, I'm just postulating.  Perhaps out of my nether regions, for which I might need to apologise...

[1] A constantly rotating frame of reference, around the barycentre of the two bodies concerned (typically NAMND to the centre of the larger, rotating with the same periodicty as the smaller's orbital plan), upon which a more complex 'orbital ellipse' would be plotted that could be like an oscillating spiral if sent far enough in either direction.  Plotted on whatever frame the normal orbits-cam uses (rotationally static to the stellar background, I think, but centred around the main orbital body) it might look strange[2], but as understandable by the experienced orbital navigator as existing "slingshot" orbit->moon encounter->moon-de-encounter->onwards tracks.

[2]

Spoiler: Thusly, without the annotation? (click to show/hide)

That's a perfectly closed path.  A slightly different trajectory would create a path that (for any given "currently at" point in the orbit) will deviate both prograde and retrograde legs of the 'orbit' by increasing amounts.  Maybe at some point breaking out into a different sphere of influence.  But even if it's tight enough that it stays within those bounds[3], the diametrically 'opposite' point in the orbit[4] could be a disconnection between the two equi-opposite 'leg-ends'.

[3] In that picture it might depend on no major Martian/Venusian-orbit planet finding itself assigned as more of an influence than the Earth is, as it transition via the long, far left side of that frame's horse-shoe, under my limited-but-N>1 influence scheme.  In fact, I'm not sure if this particularly extreme example would work under my proposed scheme...  But it'd be nice if it could!!!

[4] Imagining the tracked object is currently on the far left stretch (effectively retrograde, by that frame's standards, but celestially still prograde), the 'opposite' point on the orbit would actually be on the lower, sunwards, stretch, on exactly the same side.  Move the object position round one way and the equi-opposite point would (at a different speed) retreat in the other direction, of course.

[miauw62]

What is the advantage of Nuclear engines over normal ones?

[pilgrimboy]

I was wondering if any of these features are currently in the game.

Take on Missions to attract interest in your Space Program. Or use the Mission Planner to create your own missions.
Manage your Space Program. Hire astronauts, train them, research and improve parts.
Build Space Stations, and surface bases on other worlds.

They are on the planned features. Do they keep that page updated?

I'm just wondering if KSP has moved into being a game rather than just a ship designer, launcher.

Really interested in it once it is developed a little more.

[miauw62]

Well, there is a carreer option when creating a new save. Not sure if it does anything. I would find it frustrating to have limited money/stuff myself. And with mods there is tonnes of stuff to do, like map all the planets, build Kethane drilling whatevers everywhere...