What do you think about automated launches?

[JAFO]

10 hours ago, MailletC said:

I was thinking more like the MechJeb Ascent Guidance module, but using it only after doing enough testing to design a custom launch profile.

Honestly, this sounds like GravityTurn would be most suited to what you're after.

[RizzoTheRat]

I tend to use MechJeb's launch autopilot for later game launches where the part count means I'm getting a fair bit of lag.  However I've found in 1.0.5 it's not very good at it, it will quite often flip out a rocket that I can launch manually without difficulty.

I also like automatic execution of nodes (either via MechJeb or RemoteTech, if I do them manually I'm everlastingly failing to drop out of time compression quickly enough and having to go back to a save.  Plus where's the fun in tweaking the manouver nodes to get a precise altitude on a circularisation or altitude adjustment burn, I'm much rather just automate simple but time consuming stuff like that.

One day I'll get round to playing with KOS, I like the idea of programming your own autopilot, although I'm no programmer and haven't studied any control theory since university 20+ years ago so it could be a steep learning curve, although rather than launches I'm more interested in programming landings so I can crank up the delay on RemoteTech (when Curiosity landed on Mars the landing took 7 minutes, but the signal delay was 14 minutes) 

 

 

Some people get very hot under the collar about automation and MechJeb, probably because they fail to realise that other people want different things out of the game, for some it's the thrill of seeing a craft they've designed launch perfectly with almost no additional control inputs, for others the launch is just a necessary evil and they're far more focussed on doing the mission to another planet.  There there's others like Overlander who would probably say "What are launches?"  

 

 

[moogoob]

I launch just about everything using Mechjeb. Yeah I can do it myself, and probably better than MJ could. But the repeatability can't be beat when engineering something to get to space.

[glen.mack]

On 22/02/2016 at 9:44 AM, RizzoTheRat said:

One day I'll get round to playing with KOS, I like the idea of programming your own autopilot, although I'm no programmer and haven't studied any control theory since university 20+ years ago so it could be a steep learning curve, although rather than launches I'm more interested in programming landings so I can crank up the delay on RemoteTech (when Curiosity landed on Mars the landing took 7 minutes, but the signal delay was 14 minutes

 

I doubt it will be that steep if you have programming experience, even with landings. Unless you're aiming for a spot to an accuracy of a few meters then it can get pretty tricky. I've seen some people on the forums working on collision detection and terrain scanning, and I know one of the demo programs has some slope detection. Transmission delay was something I actually forgot to talk about in the video.

[Mad Rocket Scientist]

I think that video is incredible.  Great job on getting a precise landing with the booster and capsule!

[Guest]

On ‎2‎/‎22‎/‎2016 at 4:18 AM, JAFO said:

Honestly, this sounds like GravityTurn would be most suited to what you're after.

Honestly, that mod looks great! I'll have to run a test with it. Thanks JAFO.

[Talavar]

I play KSP as a coffee game, so I religiously use Mechjeb. I can already do it manually, but I've already done it so many times, that I'm not worried about testing myself any more. Also, it's nice to have mechjeb for "hold vertical" when making a hover craft, or helicopter with mod parts.

Edited February 25, 2016 by Talavar

[stosyfir]

Once in a while I'll pop on mechjeb's ascent autopilot on a launch vehicle Ive already flown myself. Really depends on the mission and whether or not Im feeling lazy. Its useful enough if Im just hauling up a space station module - I always perform the rendezvous and docking myself though (some may call me crazy but thats the fun part)

[regex]

On 2/20/2016 at 10:40 AM, glen.mack said:

How would you implement an autopilot feature if it were up to you?

I wouldn't.  A big part of KSP for me is actually flying the rocket.  I've used MechJeb in the past and the game just felt kind of ... hollow with it.

[Frank_G]

I like to design, construct, fly, do EVAs and and... Why should i let the machine play my game? Every orbit a true achievement, each 30 minute transfer burn a nice relaxing tea break. I enjoy every bit of KSP

[Geschosskopf]

On 2/20/2016 at 0:40 PM, glen.mack said:

New Shepard has flown and landed twice. According to the wiki, more sophisticated autopilot features will not be part of the main game. How do you feel about automated launches and landings?

Best way to do things, and in fact most real rockets use it.

KSP's in-game informational systems are analogous to the cockpit instrumentation of WW1 airplanes.  Sure, you can use it to fly stuff yourself and I often do, but I find that it clashes mentally with the whole rockets thing.  To the point of breaking immersion.  Besides, I usually launch lots of rockets at once.  So I more often use MJ.

[tater]

Until very recently I had never used anything like mechjeb. I got Gravity Turn recently, and have been messing with it to consistently test designs (a kopernicus rescale, though, not stock). Honestly, once a design is decent launches are pretty much hands-off anyway. Regardless, I think it's interesting to experiment with it (it's entirely novel to me at the moment).

All that said, I think that there should be options to do everything in KSP, not only "autopilot" (which in effect is every real launch we watch), but having kerbal pilots actually able to pilot. I consider the important "piloting" to be setting up maneuver nodes, though the most fun piloting to me is propulsive landings. That is something I'd basically always want to do myself. 

I'm thinking of career mode wrt kernels as functional pilots, because as a "manage a space program" game, I think that having the ability to hand certain routine tasks off allows you to do different things with less tedium. Setting up an infrastructure to assemble larger ships in orbit, for example, then letting the staff at least place things nearby each other, or docked to a station so I can do final assembly would be interesting. Resupply missions (I always use life support)... I dunno, I think that it would be interesting to see, even if I'd want to do it myself most of the time.

 

[Temeter]

Automatisation surely has it's place, I like Mechjeb a lot.

That said, playing manually is a big element of KSP, and can give you an understanding of the games system - and real space travel - you might otherwise never learn. So I mostly use MJ to do routine tasks and some more advanced planning. Sure I could - and did - e.g. manually create interplanetary transfers, estimating the phase angles by eye and lots of experimentation, but it gets repetetive at times, and then MJeb is a great help.

Otoh, automation might be either your highest level of play or something completely to avoid. That's one of the cool things about KSP, it gives you an incredible amount of freedom and how to approach your tasks.

Edited February 25, 2016 by Temeter

[Gojira1000]

On 2/21/2016 at 10:02 AM, Bill Zarr said:

 In a game I'd rather be playing Neil Armstrong style, heroically nailing a moon landing on manual than to be spam in a can watching a blinking autopilot.

I

 

That was Buzz Aldrin, and he's coming to your house to punch you.

[tater]

The Commander flew the LEM, not the LM Pilot. Strange, but true. Maybe Dr. Rendezvous did the orbital piloting?

[Kobymaru]

I love to fly stuff, it's a great source of enjoyment for me. What drives my playing, however, is "advancement". To a smaller extent simple grinding parameters like Career Money and Science points, to a greater extent advancements in "technology" (my building skills), procedures and flying skills.

Doing everything manually is absolutely fine - in the beginning. Sure, the first times in a career where you get to orbit or do your first Mun landing, everthing is new and exciting. But when you have done over 200 launches, often enough comepletely identical - the excitement wears off.

I like to advance in mission complexity: build bigger ships, more landers, more functionality, kolonization... The beginnings of these complex missions are the same: launch a lot of stuff into orbit, rendez-vous, dock, transfer burn, ... This gets boring over time.

So I am absolutely for automation in the late game.

[justidutch]

I have probably launched close to 1,000 times in the 3.5 years I've played this game.  Not once, ever, have I thought a launch to be boring.  Or repetitive.  Even if I happen to launch the same rocket (very rare) over and over, I am always trying to get better and better flight profiles.  I find this fun; I must be defective.  When playing career I find it very rare to use the same rockets more than a couple launches before I am able to get newer and better parts, and therefore designing new craft.  Which demand different flight profiles, which I find challenging to figure out.  Every single time.  Part of the fun lies in that I don't always get it right, sometimes going too fast and wasting fuel that way, or going too slow, or my gravity turn was too turned too early, etc etc.

I can understand how some would find this ridiculous, but I enjoy it.

In sandbox I've built stations that required 16 launches of the exact same vehicle (segments of a ring), and flown them manually each time, because they were massive and hard to get it right.  Then, getting the intercept and docking the massive ships were very very intense, hard, and therefore fun. For me.

Automation in KSP, for me, would be like drinking non-alcoholic beer.  What the heck would be the point?

[Bluegillbronco2]

On 2/21/2016 at 1:51 AM, Evanitis said:

I learned a lot from MJ, but these days I use it less and less. I'm tempted to change it to KER, but I'd really miss the Manuver Planner tool. Creating a circularization node, or one that zeroes relative velocity with one click is -so- much more convenient than fiddling with the stock axises. I also love having an ingame porkchop-plotter (that has more functionality than the most popular web-based one).

I sometimes use it's ascent autopilot feature too, but not for making my job easy. I found that I can launch -anything- manually, but designing something that even MJ can get to orbit is an engineering challenge for me (maybe it can launch anything too, but I never figured out how to utilize it many settings).

I'd also really like to get into kOS scripting one day, but with my zero~ish programming experience, that would require a week when I don't have anything else to do - and I don't think I'll have such luxury in the foreseeable future.

That's why i have both much jeb and Kerbal Engineer. I get the persice maneuver editor and flight recorder of mJ. And I get the extremely helpful and fully customizable readouts of KER.

[Norcalplanner]

Maybe I'm in the minority here, but I launch rockets using a hybrid approach. I use MJ's SmartASS function in surface mode and click my mouse a lot instead of pressing "d" a lot. The thrill of manual launches combined with the precision of automation is my happy place.

[Bill Zarr]

9 hours ago, Gojira1000 said:

That was Buzz Aldrin, and he's coming to your house to punch you.

No, you are wrong. It was Neil Armstrong who piloted the landing, (he had a background as a test pilot, Aldrin did not.) If you had bothered to use a search engine you could have easily found this out. Wikipedia gives this information, but since that sucks as a reliable source, and I can't point to it out in a book over the internet I'll direct you to a nasa pdf  of an interview with Armstrong (see page 83.)

BRINKLEY:  I guess we should mention — we jumped the story a little — but when the landing occurred, you had to pilot to make sure it didn't go into a crater. How damaging would that have been if you couldn't and you landed in — would that have been the end of the mission?

ARMSTRONG:  We could have tried to land there, and we might have gotten away with it.  It was a fairly steep slope and it was covered with very big rocks, and it just wasn't a good place to go.  You know, if I'd run out of fuel, why, I would have put down right there, but if I had any choice of a promising spot, I was going to take it. There were some attractive areas far more level, far less occupied by boulders and things, a half mile ahead or so, so that's where I went. I wanted to make it as easy for myself as I could on that first — there's a lot of concern about coming close to running out of fuel, and I was very cognizant of that.  But I did know that if I could have my speed stabilized and attitude stabilized, I could fall from a fairly good height, perhaps maybe forty feet or more in the low lunar gravity, the gear would absorb that much fall.  So I was perhaps probably less concerned about it than a lot of people watching down here on Earth.  That's not to say I wasn't thinking about it, though, because I certainly was, but I thought it was important to try to get it down smoothly on the first try. We didn't know how that landing was actually going to go until that point.  So I wanted to make it as gentle as I could.

[Gojira1000]

1 hour ago, Bill Zarr said:

No, you are wrong. It was Neil Armstrong who piloted the landing, (he had a background as a test pilot, Aldrin did not.) If you had bothered to use a search engine you could have easily found this out. Wikipedia gives this information, but since that sucks as a reliable source, and I can't point to it out in a book over the internet I'll direct you to a nasa pdf  of an interview with Armstrong (see page 83.)

BRINKLEY:  I guess we should mention — we jumped the story a little — but when the landing occurred, you had to pilot to make sure it didn't go into a crater. How damaging would that have been if you couldn't and you landed in — would that have been the end of the mission?

ARMSTRONG:  We could have tried to land there, and we might have gotten away with it.  It was a fairly steep slope and it was covered with very big rocks, and it just wasn't a good place to go.  You know, if I'd run out of fuel, why, I would have put down right there, but if I had any choice of a promising spot, I was going to take it. There were some attractive areas far more level, far less occupied by boulders and things, a half mile ahead or so, so that's where I went. I wanted to make it as easy for myself as I could on that first — there's a lot of concern about coming close to running out of fuel, and I was very cognizant of that.  But I did know that if I could have my speed stabilized and attitude stabilized, I could fall from a fairly good height, perhaps maybe forty feet or more in the low lunar gravity, the gear would absorb that much fall.  So I was perhaps probably less concerned about it than a lot of people watching down here on Earth.  That's not to say I wasn't thinking about it, though, because I certainly was, but I thought it was important to try to get it down smoothly on the first try. We didn't know how that landing was actually going to go until that point.  So I wanted to make it as gentle as I could.

I forget not everyone gets humor. http://factually.gizmodo.com/buzz-aldrin-once-punched-a-moan-hoaxer-in-the-face-1605920370

[Bill Zarr]

35 minutes ago, Gojira1000 said:

I forget not everyone gets humor.

Looks more like you forgot people don't like to be insulted.

I'm not sure why you think I'd want to be made the butt of your joke. The implication that I'm either to stupid to know who the piloted the first moon landing or am some kind of moon hoaxer conspiracy theory loony is just insulting. In case you didn't know, generally people don't find it amusing if the joke goes "you're an idiot and should be punched in the face".  If you insult me in a joke that's based on a complete lie then yea I'm going to point out that factual inaccuracy because that's an easy way to kill the joke. Obviously that upset you since you resorted to insulting my sense of humour.

Perhaps you should consider not making "jokes" at other peoples expense, and not further insulting them when they don't find your first set of insults to be humorous.

 

 

[DerekL1963]

On 2/21/2016 at 1:13 PM, Dres said:

For me, the game would be unplayable without some autopilot mod.

This.   With my eye-hand coordination, without MJ, I'd simply have to delete KSP and walk away.  (And before some idiot pops up with the same nonsense they always do...  No, it won't "get better if I practice", it's a physical limitation and that's all you need to know.)

And there's no One True Way to play KSP anyhow - if there, there wouldn't be mods.  The game would be locked down to force us onto that One True Way.   The only people who advocate One True Way are people trying to puff themselves up over others.   Some people like being Neil Armstrong...  Me, I like being George Low, and Bob Gilruth, and Werner von Braun, and Chris Kraft, and Gene Kranz.    When MJ touches down a lander I've designed as part of a mission plan I've conceived, I'm rightfully satisfied and proud of myself - the machinery worked just like *I* planned.   If someone thinks I'm wrong for feeling so, or cheating for using MJ, or one of the dozen or more the inane variants of the same stupid claim...  they're the ones that are the less for it and their ignorance is not my problem.

[RizzoTheRat]

This thread got me looking at kOS.  Automating launches is a lot harder than doing it manually

[Wallygator]

22 minutes ago, RizzoTheRat said:

This thread got me looking at kOS.  Automating launches is a lot harder than doing it manually

^^^  This.

I think I'll take a crack at kOS.  The idea of planning and implementing a detailed launch sequence in conjunction with an over-engineered flying monstrosity seems like it would feel more like a "real space program" program!

Hmmm... Obtaining kOS now... stay tuned...