Jump to content

[1.12.x] Anatid Robotics / MuMech - MechJeb - Autopilot - [2.14.3] [4th March 2023]


sarbian

Recommended Posts

It already happened several times, usually going to norther landing places, and always missing by a similar distance. Meanwhile I have done several other landings without problem.

It seems that the problem was starting to low. TRying again from 25km worked as expected.

dDPVP8K.jpg

Edited by Tacombel
Link to comment
Share on other sites

41 minutes ago, Tacombel said:

It already happened several times, usually going to norther landing places, and always missing by a similar distance. Meanwhile I have done several other landings without problem.

It seems that the problem was starting to low. TRying again from 25km worked as expected.

I've seen that from varying heights.  My general conclusion is that MechJeb needs to be able to move the ship to point towards the burn within a certain amount of time to be accurate - I typically see it with ships with low amounts of control torque on SAS.  The other situation I see for it is if the TWR significantly drops during the landing process.

Link to comment
Share on other sites

4 hours ago, Philaphlous said:

I did notice the first time after installing the mod it reverted to 0 then afterwards it was like -189....lol I'll try that! Do you just put 0 in the launch to plane of target or do you also put 0 in the inclination right below the altitude?

For me, this field is always fill with randon number (for me), i just put 0 for the launch to plane of target and the inclination is just computed by mechjeb so 6° for minmus.

Link to comment
Share on other sites

On 7/9/2019 at 7:32 AM, LTQ90 said:

in the ascent Guidance windows, before clicking on "Launch into plane of target", i ALWAYS put 0 in the textbox at the right of this button. and all is ok. 

This is the correct answer.

I'm not really certain that box is useful at all in stock launches right now for launch to plane.  It should probably always be zero, and that textbox and all the code behind it should most likely get removed (although not from the rendezvous function, just from launch-to-plane).

Link to comment
Share on other sites

Just started playing again ... updated all my mods etc, got rid of some old out of date ones.     I'm having a curious problem with MechJeb's Landing Guidance ... the cursor to select a landing is missing.   You can still select a landing spot, but that then also is missing the target cursor?

Link to comment
Share on other sites

49 minutes ago, Mordrehl said:

Just started playing again ... updated all my mods etc, got rid of some old out of date ones.     I'm having a curious problem with MechJeb's Landing Guidance ... the cursor to select a landing is missing.   You can still select a landing spot, but that then also is missing the target cursor?

I think a lot of us are seeing this problem. It's intermittent for me. Today it worked for probably the first time I remember in a couple weeks. I can't figure out a method to reproduce it or work around it. Sometimes it works, but lately (since BG) it seems less reliable.

Link to comment
Share on other sites

37 minutes ago, Tonka Crash said:

I think a lot of us are seeing this problem. It's intermittent for me. Today it worked for probably the first time I remember in a couple weeks. I can't figure out a method to reproduce it or work around it. Sometimes it works, but lately (since BG) it seems less reliable.

I'll try a game with only mechjeb running to see what happens when I get around to it ... thinking it may be a couple of mods that are 1.6.1 only such as umm KWRocketry or Spacey ... think those are the ones from memory.

Link to comment
Share on other sites

On 7/9/2019 at 12:30 PM, DStaal said:

I've seen that from varying heights.  My general conclusion is that MechJeb needs to be able to move the ship to point towards the burn within a certain amount of time to be accurate - I typically see it with ships with low amounts of control torque on SAS.  The other situation I see for it is if the TWR significantly drops during the landing process.

Current landing AP is pretty wrong. Landing should be a continuous burn with no course corrections. Instead, if there is error then it needs to compute how much corrective burn it needs but then spread it out over the remaining burn time. And the current targeted system IS capable of that but only within the final stage of the burn and if the burn is interrupted for any portion of time it will want to do another course correction before moving on to the final burn stage.

The landing AP also has no real concept of vertical velocity management which is why landing from too low an altitude tends to result in crashes. It spends too much time cancelling the horizontal velocity and unless its gravity turn orients downwards enough to cancel out the vertical before the suicide timer goes negative then it will probably crash.

I've been sitting on some code for the AP that takes that into account and pitches up just enough to keep its downwards momentum under control. It works pretty well in scaled systems but not so much in stock so I never made a pull request on it. And it doesn't work for targeted landings at all. 

Link to comment
Share on other sites

59 minutes ago, Starwaster said:

Current landing AP is pretty wrong. Landing should be a continuous burn with no course corrections. Instead, if there is error then it needs to compute how much corrective burn it needs but then spread it out over the remaining burn time. And the current targeted system IS capable of that but only within the final stage of the burn and if the burn is interrupted for any portion of time it will want to do another course correction before moving on to the final burn stage.

The landing AP also has no real concept of vertical velocity management which is why landing from too low an altitude tends to result in crashes. It spends too much time cancelling the horizontal velocity and unless its gravity turn orients downwards enough to cancel out the vertical before the suicide timer goes negative then it will probably crash.

I've been sitting on some code for the AP that takes that into account and pitches up just enough to keep its downwards momentum under control. It works pretty well in scaled systems but not so much in stock so I never made a pull request on it. And it doesn't work for targeted landings at all. 

I don't think keeping the downward momentum under control is what you want.  You want to leave the burn for as late as possible.  The ideal landing is a gravity turn in reverse.  Without knowing exactly how the rocket will perform you can't do that precisely, though, so you're going to need to allow some margin.

Link to comment
Share on other sites

18 minutes ago, Loren Pechtel said:

I don't think keeping the downward momentum under control is what you want.  You want to leave the burn for as late as possible.  The ideal landing is a gravity turn in reverse.  Without knowing exactly how the rocket will perform you can't do that precisely, though, so you're going to need to allow some margin.

Your second and third sentences contradict each other ;)

Leaving your burn as late as possible is the opposite of a gravity turn in reverse. Apollo's lunar landers did the latter. Typically  starting with 115 x 15km orbits and braking at the low point. Pitching up just enough to manage their descent so that both vertical and horizontal are mostly canceled as they approach their landing sites. (keep in mind that their orbital velocity was starting at ~2km/s while doing this)

The difference between the two (shallow vs steep) is that shallow landings have lower gravity losses. 

But I guess for most KSP stock landings it's not usually an issue since our landers don't usually have to worry about delta-v that much or conserving any of it for the final approach. 

Nevertheless, it is more efficient to do a shallow rather than steep landing. 

Link to comment
Share on other sites

13 hours ago, Tonka Crash said:

I think a lot of us are seeing this problem. It's intermittent for me. Today it worked for probably the first time I remember in a couple weeks. I can't figure out a method to reproduce it or work around it. Sometimes it works, but lately (since BG) it seems less reliable.

Yup wierdest thing ... 

Did my first Mun landing last night, with the cursor not working.    Had not shut down KSP, it was running since last night ... and went and did another mun landing and the darn cursor is working all on it's own again.

Link to comment
Share on other sites

One thing that hasn't been clear to me is how does MechJeb account for terrain elevation in plotting it's descent profile, if at all. This is only a problem for me on the Mun. I usually do most of my landings from a station in an 11km orbit, so the starting orbit parameters are nearly identical. Sometimes the lander may come in almost horizontal finally cancelling horizontal speed within meters of final touchdown. An unexpected crater rim can be catastrophic. Other landings I'm descending vertically for close to a thousand meters. The elevation also seems to throw off the accuracy. When targeting a base, I have to iterate on the landing coordinates to find a set that will get me close to the desired touchdown point. 

One thing I'd like to see is once the horizontal speed is mostly canceled in the final descent is some attempt to maneuver closer to the target coordinates instead of just doing a vertical descent. I end up doing trying to do this by hand a lot of the time with mixed results, or use the Translatron function to "hover taxi" to where I want after touchdown. 

Link to comment
Share on other sites

9 hours ago, Starwaster said:

Your second and third sentences contradict each other ;)

Leaving your burn as late as possible is the opposite of a gravity turn in reverse. Apollo's lunar landers did the latter. Typically  starting with 115 x 15km orbits and braking at the low point. Pitching up just enough to manage their descent so that both vertical and horizontal are mostly canceled as they approach their landing sites. (keep in mind that their orbital velocity was starting at ~2km/s while doing this)

The difference between the two (shallow vs steep) is that shallow landings have lower gravity losses. 

But I guess for most KSP stock landings it's not usually an issue since our landers don't usually have to worry about delta-v that much or conserving any of it for the final approach. 

Nevertheless, it is more efficient to do a shallow rather than steep landing. 

I'm not contradicting myself.  On launch a rocket goes straight up, then soon tips with the rotation of the body and accelerates to orbit.  This is the lowest energy trajectory, you want to do the exact reverse to land with the least fuel.  The first burn puts your periapsis a bit above the point where you want to have killed your horizontal velocity (the reverse of your circularization burn.)  You then come in, lighting your rocket at the point it will kill your horizontal velocity over your target.  This of course means you end up lower than the orbit you were in, hence the reason your periapsis is higher than your objective.  As you approach your target you tip more vertical as you can't just suddenly turn vertical when you complete your burn.  (Now, if you have a small rocket with the vastly-overpowered rotation capability of many such rockets in KSP you would be better off not rotating until you're over your target.)  You still have substantial vertical velocity at this point, optimum is you shed it in a suicide burn.

This isn't what Apollo did because they took the safe route.  Their approach burned considerably more fuel, though.

Link to comment
Share on other sites

7 hours ago, Loren Pechtel said:

I'm not contradicting myself.  On launch a rocket goes straight up, then soon tips with the rotation of the body and accelerates to orbit.  This is the lowest energy trajectory, you want to do the exact reverse to land with the least fuel.  The first burn puts your periapsis a bit above the point where you want to have killed your horizontal velocity (the reverse of your circularization burn.)  You then come in, lighting your rocket at the point it will kill your horizontal velocity over your target.  This of course means you end up lower than the orbit you were in, hence the reason your periapsis is higher than your objective.  As you approach your target you tip more vertical as you can't just suddenly turn vertical when you complete your burn.  (Now, if you have a small rocket with the vastly-overpowered rotation capability of many such rockets in KSP you would be better off not rotating until you're over your target.)  You still have substantial vertical velocity at this point, optimum is you shed it in a suicide burn.

This isn't what Apollo did because they took the safe route.  Their approach burned considerably more fuel, though.

https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19700024568.pdf

https://www.hq.nasa.gov/alsj/nasa58040.pdf (same document but not text searchable; just a scan of the original document)

"The purpose of the DO1 is to efficiently reduce the orbit altitude from approximately 60 nautical miles to 50 000 feet for PDI. Performance of contir~uous powered dcscent from altitudes much greater thwl 50 000 feet is inefiicient, and a PDI at lower than 50 000 feet can C become a safety hazard (ref. 3)"

I don't know what your source of information is but I trust the original documents describing the why and how of the lunar landings. And that document says that Apollo's descent trajectory was chosen for efficient propellant usage.

Link to comment
Share on other sites

Wanted to  give a bit feeback, works so far ok, but i've noticed, with airplanes autoguidance and landing ( not the rocket auto landing mode), when it flies to line up with the runway it never goes in a state for intercepting glidescope  going left to right and back, and when it coasts towards the  landing point, it lets the plane just drift off....

 

Second, for the Rover autopilot module, works okay by logic, but when trying to add waypoints to follow from map view, its locations are not accurate with the mouse. for example i can click 15 km further on the map to designate a waypoint, but the actual marker ( and its data) are way off. However when in normal view of ur vessel, zooming out as far as suitable to get a overview of the planet, i can set waypoint just fine and very accurate. just annoying is the point ( as this is tested today at minmus) that the drawn line between the points, is most of the time dug into the ground. position a lil bit higher is a fix.

 

Other than that nothing yet to report, but playing carreer only so i have not unlocked all features, yet.

 

Link to comment
Share on other sites

gxCCh48.jpg

Was someone complaining about accuracy of the Landing Guidance module? I landed the probe at Mun East Crater first, then manually input its coordinates into the Landing Guidance and landed this manned lander right next to it. Surely no reasonable person could ask for more.

MechJeb Dev #886 Sarbian  *  KSP 1.7.3.2594 (WindowsPlayer x64) en-us

Edited by El Sancho
national security
Link to comment
Share on other sites

12 hours ago, El Sancho said:

gxCCh48.jpg

Was someone complaining about accuracy of the Landing Guidance module? I landed the probe at Mun East Crater first, then manually input its coordinates into the Landing Guidance and landed this manned lander right next to it. Surely no reasonable person could ask for more.

MechJeb Dev #886 Sarbian  *  KSP 1.7.3.2594 (WindowsPlayer x64) en-us

It works for many ship designs - but not for all of them.  If you try to land with low attitude control and low TWR you're more likely to have issues.  Most of my ships land like yours - but I have occasional designs which will miss the landing point by up to 20km.

Link to comment
Share on other sites

2 hours ago, DStaal said:

It works for many ship designs - but not for all of them.  If you try to land with low attitude control and low TWR you're more likely to have issues.  Most of my ships land like yours - but I have occasional designs which will miss the landing point by up to 20km.

Agreed! In that case the solution is to start from a higher orbit, in my experience. I can't really explain it in theoretical terms, but I had one lander design that would land within 20 meters from a 300km orbit, but if I started from 100km up, it could be off by 200 meters.

@Gordon Dry - Yes, absolutely! Landing in atmosphere has always been a Charlie Foxtrot.

 

Link to comment
Share on other sites

My apologies if this has been answered elsewhere, but how do I remove the career tech lock? I tried "MechJeb Embedded Universal" as suggested in the OP but this did not have any effect (it seems "MechJeb Embedded Universal" is really old also).

Link to comment
Share on other sites

8 minutes ago, Clinton said:

My apologies if this has been answered elsewhere, but how do I remove the career tech lock? I tried "MechJeb Embedded Universal" as suggested in the OP but this did not have any effect (it seems "MechJeb Embedded Universal" is really old also).

hi Clinton.

go in your gamedata/mechjeb2/parts and edit the file named MechJebNoCommandPod.cfg

inside this you can unlock all tech with this statement :unlockTechs = start

Hope it helps you.

Edited by LTQ90
error
Link to comment
Share on other sites

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...