Arrowstar
-
Posts
2,549 -
Joined
-
Last visited
Content Type
Profiles
Forums
Developer Articles
KSP2 Release Notes
Bug Reports
Posts posted by Arrowstar
-
-
Today I've built KSPTOT v1.6.10 pre-release 7. This pre-release adds a bunch of new functionality and includes some bug fixes and minor performance improvements, all of which are in LVD. Here's the change log:
- Noticeable performance improvement when using numerically propagated celestial bodies.
- LVD: Added new ground track tab to main LVD UI's display area.
- LVD: Added sky box code for more "cinematic" display of 3D trajectories. Set option to enable in View Settings.
- LVD: Added additional camera controls functionality (Dolly horizontally/vertically camera).
- LVD: New steering model: tabular quaternion interpolation.
- LVD: New throttle model: tabular throttle interpolation.
- LVD: Update version of IPOPT optimizer to v1.1.6.
- LVD: Add new SQP optimizer algorithm.
- LVD: Bug fixes and performance improvements.
Please let me know if you find any bugs. Thanks! Happy orbiting!
-
On 2/11/2024 at 7:07 PM, D_X_Racer_2 said:
I'm running into a problem with the new PR6. After opening the program I can access every module except LVD.
When I select LVD I get the message "Starting Launch Vehicle Designer. Please Wait...". After several seconds there is a brief flash of a window opening and then closing immediately. Then several seconds after that I get a message "An error occured while opening Launch Vehicle Designer."
I downloaded the PR6 file and runtime twice to make sure it wasn't a bad download. I also tried installing on an older notebook that never had this installed just to make sure there wasn't any conflict with previous runtime installs.
The error log is here
======================================== _ __ _____ _____ _______ ____ _______ | |/ // ____| __ \__ __/ __ \__ __| | ' /| (___ | |__) | | | | | | | | | | < \___ \| ___/ | | | | | | | | | . \ ____) | | | | | |__| | | | |_|\_\_____/|_| |_| \____/ |_| ======================================== KSPTOT v1.6.10 PR6 MATLAB 23.2.0.2485118 (R2023b) Update 6 DATE: 2024/02/11 18:48:15 ======================================== Warning: No celestial bodies in the loaded bodies.ini file are using numerical propagation. > In CelestialBodyIntegration/integrateCelestialBodies In CelestialBodyData In projectMain Unable to resolve the name 'internal.matlab.inspector.PropertyEditedEventData'. Error in matlab.graphics.internal.propertyinspector.generatePropertyEditingCode (line 7) Error in matlab.graphics.internal.CameraToolBarManager/setcoordsys (line 870) Error in cameratoolbar (line 90) Error in ma_LvdMainGUI_App/ma_LvdMainGUI_OpeningFcn (line 1379) Error in ma_LvdMainGUI_App>@(app)ma_LvdMainGUI_OpeningFcn(app,varargin{:}) (line 5427) Error in appdesigner.internal.service.AppManagementService/runStartupFcn (line 135) Error in matlab.apps.AppBase/runStartupFcn (line 68) Error in ma_LvdMainGUI_App (line 5427) Error in newMainGUI_App/LVDButtonPushed (line 306) Error in appdesigner.internal.service.AppManagementService/executeCallback (line 172) Error in matlab.apps.AppBase>@(source,event)executeCallback(ams,app,callback,requiresEventData,event) (line 62)Is there a new step in the setup that I missed?
No, this is a new one certainly. Can't say I've seen this before. I'm going to push a PR7 out pretty soon (maybe even tomorrow), so can you let me know if the issue persists with the new build? If it does, I'll investigate further. Thanks!
-
23 hours ago, EwTE_H said:
Mockingbirdnest says that as principia is turned on or off the celestial bodies are likely to move quite a bit. If I use stock gravity, will I find myself going to the place where the celestial body is not here anymore?
Ah, I didn't realize you were using Principia. In that case, yes, you can turn on the n-body calculations for the celestial bodies as you had it. Just realize that it's going to be very slow. Sorry about that.
-
23 hours ago, EwTE_H said:
So I should use stock-like gravity on celestial bodies and n-body gravity on the spacecraft? If the destination is one of the Joolian moons where n-body gravity on celestial bodies is necessary, what should I do?
Correct, that should work just fine!
-
14 hours ago, EwTE_H said:
What do you mean? Should I give up n-body and return to patched conics?
Ah, no, sorry, I think I see the confusion. Numerically propagating the orbits of the celestial bodies is NOT the same as using n-body gravity on your spacecraft. To enable n-body gravity, you need to do the following:
- In your initial state, set the bodies you want to use for 3rd body gravity.
- For each event, set the Force Model to use 3rd body gravity.
That should be all you need!
-
1 hour ago, EwTE_H said:
Thank you but I still have a few questions.
How to turn on numerical propagation only for some celestial bodies? Does it affect accuracy to turn off some or all of that?
Where does the number 10 come from?
Honestly, just leave the numerical propagation to two body for everything. Your simulations will run much faster.
You typically want to scale your objective function so that the value is between -1 and 1. Since final vehicle mass is the objective function, and the final vehicle mass is around 4, a scale factor of 10 is appropriate since it'll make it 0.4, which is between -1 and 1.
-
On 1/31/2024 at 9:02 PM, EwTE_H said:
Here https://livefile.xesimg.com/programme/python_assets/c79165d17bf063d3268aee2d56bfe753.mat
I'm not sure if you can access this website. If you can see the image below, then you can access it.
You had a couple of issues that I was able to resolve:
1) You had numerical propagation turned on for all your celestial bodies. That's much slower than the two-body analytic solutions and was causing the propagator to take forever. I had to change things back to analytic two body propagation.
2) I scaled the constraints by their current value (Optimization menu -> Scale Constraints -> Current Value).
3) I scaled the objective function by 10.
Once I did all of these, the optimizer brought the solution in easier and quickly. Here's the file. Please let me know if you have any questions with anything I did! Happy orbiting!
https://drive.google.com/file/d/1t9DxXgP46S7R3gHbPCPeSpExvdcHQHFy/view?usp=sharing
-
On 1/31/2024 at 9:02 PM, EwTE_H said:
Here https://livefile.xesimg.com/programme/python_assets/c79165d17bf063d3268aee2d56bfe753.mat
I'm not sure if you can access this website. If you can see the image below, then you can access it.
I have access but when I try to load your MAT file I'm told that it's not a binary MAT file. Can you try again?
EDIT: Nevermind, seems like it works now.
-
5 hours ago, EwTE_H said:
Can MFMS simulate n-body gravity? I loaded bodies_StockPrincipia.ini and MFMS gave me a maneuver sequence(K-E-E-M). However, when I opened the result file in LVD it violated every constraint. The optimizer didn't help. It just couldn't lower the constraint violation.
btw If MFMS uses stock-like patched conics then how can I get the maneuver sequence?
It can, but I'd still highly recommend using LVD. Did you try scaling the constraints as the text box message that pops up describes? Can you share your MFMS output with me so I can take a look at it and see what's going on?
In other news, work on the new optional sky box functionality for LVD is progressing nicely. No more grey axes boxes if you don't want them!
-
I've been working on a new display for Launch Vehicle Designer (LVD). This is a new ground track display that can show the following on a 2D map of a celestial body. This display is located on a new tab in the Display area of the main LVD GUI.
- Vehicle trajectory
- Ground object location and trajectory
- Celestial body location and trajectory
- Terminator line between lit and shadow (day and night) portions of the celestial body
- Terrain contours (may be turned off and off in View Settings)
Here's a great image of Duna that illustrates these new functionalities.
You'll also notice that the tool tips on this display are much richer and show more useful data. This is true for the 3D orbit display as well!
This will show up in the next v1.6.10 pre-release. Let me know what you think! Happy orbiting.
-
Hi everyone,
I'm happy to announce that after a fairly lengthy hiatus, I've built KSPTOT v1.6.10 pre-release 6. This new build moves KSPTOT to a new MATLAB version, MATLAB R2023b. You will need to download and install the MATLAB Compiler Runtime version for R2023b to run this build. Please do so before attempting to run KSPTOT v1.6.10 PR6.
Here's the change log:
- New MATLAB Version: R2023b
- LVD: Update NOMAD solver to v4.4 and removed v4.3.1.
- LVD: Enabled DiscoMADS option for NOMAD solver.
- MA/LVD: Updated the optimization variable plot mouseover label background color to be not transparent.
- LVD: Run Case Matrix workers now record diary as they run to a "running" log file.
- LVD: Updated tab order for initial state and set kinematic state UIs.
- LVD: New example: lvdExample_MunarFreeReturn
- MA/LVD: Many bug fixes.
Please let me know if you find any issues. Thank you and happy orbiting!
-
On 1/14/2024 at 4:52 AM, 7e21 said:
Yup. I've tried reinstalling too and very consistently get it 100% of the time. This is only with n-body coast + UT but it happens every time for me with that combo, but I'm mostly trying using LVD anyway.
Got it. I found the issue and resolved it for the next release.
-
Just a heads up to everyone, the next pre-release will probably update MATLAB to R2023b (the latest version). It'll also have an updated version of the NOMAD optimizer (v4.4) in LVD, along with some other changes.
-
On 1/9/2024 at 2:54 AM, 7e21 said:
As a side note, when I try to use nbody coast to UT in mission architect I can not open the edit window for that event again at all (like for the lunar transfer tutorial on the wiki). Here is what the log says:
Error using matlab.ui.control.UIControl/set The UIContextMenu's parent figure is not the same as the parent figure of this object. Error in ma_InsertNBodyCoastGUI>coastTypeCombo_Callback (line 688) Error in ma_InsertNBodyCoastGUI>ma_InsertNBodyCoastGUI_OpeningFcn (line 79) Error in gui_mainfcn (line 220) Error in ma_InsertNBodyCoastGUI (line 40) Error in ma_MainGUI>scriptListbox_Callback (line 294) Error in gui_mainfcn (line 95) Error in ma_MainGUI (line 42) Error in matlab.graphics.internal.figfile.FigFile/read>@(hObject,eventdata)ma_MainGUI('scriptListbox_Callback',hObject,eventdata,guidata(hObject)) Error while evaluating UIControl Callback.I actually couldn't reproduce this error. Does it still happen for you?
Quotet certainly helps and the advance to segment is a usefulthing I missed, but I'm still a little confused on how to deal with time disrepancies with how much things rely on the duration since the last state in LVD. Where there is a couple more things like coast to UT, revs before coasting to true anomaly etc in Mission Architect.
Propagating to certain "events", like true anomaly or whatnot, doesn't work as well in the integrator based system that LVD uses. This is why everything is time based. Set up constraints to the target you want at the end of an event and then optimize the propagation time of that event to meet those constraints. This should work out pretty well most of the time.
-
On 12/25/2023 at 12:16 PM, 7e21 said:
Hello! I've recently started messing with KSP TOT and specifically the LVD.
I think I am at a point where I mostly understand how to plan basic missions and gravity assists. But I am not quite sure on how to actually execute the missions.1: I plan an eve gravity assist from Kerbin Orbit with 800km SMA at say 150000s UT.
2: Now I launch a probe to that orbit but obviously the orbit isn't going to be perfect so I import the orbit from the active vessel and re-optimize. So far so good although quite a lot of toggling as I mostly want to change the time first I assume?
3: I execute my departure burn, but it's not going to be perfect, and this is where I am starting to be lost. Now on my way on the first coast event, how do I add a correction burn and re optimize my mission from where I currently am? Is there any way I can again import my spacecraft state tell it where in the mission i am, add a correction burn to fix my course, continue to my eve departure burn and repeat that until mission end?
Am I thinking about this totally wrong or am I missing something?Hey, sorry about taking forever to get back to you. Between Christmas and New Year's Day I've been pretty busy recently.
You've basically got the idea on point 3 about updating the state. Once you're in between major bits of the mission and you're just coasting along, you can remove the previous segments and update the initial state to match your current state in the game via importing it (right click in the Initial State dialog box). At that point you can add maneuvers for corrections and stuff like that.
I think there's an "advance to segment" option in the right click menu of the script listbox that helps make some of this easier!
Does this help any?
-
On 10/23/2023 at 12:57 PM, akyyy said:
Maybe someone could make a Twitch or youtube full lenght video of a gravity assist?
I don't have the skills for that, but if anyone would like to give it a go, they are welcome to let me know and I'll help with info as I can.
-
On 4/26/2023 at 7:16 AM, Sauge said:
Hey I'm trying to follow one of the included tutorials for Mission Architect but the Architects is not working for me, when I setup my initial state I get this warning....which I can't read since I can't resize the window for some reason
https://imgur.com/a/JdhRPSPHey there, that warning just means that the mission needs to be run first. It's not really an error and it doesn't imply that anything is wrong. That said, I'd encourage you to use Launch Vehicle Designer (LVD) instead, since it's a far more fully featured and robust mission planning tool that works for both launch vehicles and spacecraft. I haven't really maintained MA in a number of years now aside from a bug fix here and there.
-
Hi everyone,
Tonight I've built KSPTOT v1.6.10 pre-release 5. There are a few new features I want to discuss in this release before I get to the change log. These are all Launch Vehicle Designer (LVD) related.
First, body (meaning "spacecraft") angular rates are now available to be used as Graphical Analysis tasks and as constraints as well.
Second, there are a variety of new actions you can apply to events. These are:
-
Set Plugin Variable Value: Sets the value of a "plugin variable" to a particular constant number.
- Example: Value = 5
-
Add Value to Plugin Variable Value: Sets the value of a "plugin variable" to itself plus or minus a given constant number.
- Example: Value = Value + 5
-
Set Plugin Variable Value to Quantity: Sets the value of a plugin variable to the value of a given graphical analysis task quantity.
- Example: Value = Altitude [Kerbin Inertial Frame]
-
Set Next Event: Sets the next event to be executed to a given event.
- I'll get to why this is relevant more in a bit.
Third, there is a new Conditional Action. A conditional action is an action that lets you set the conditions by which a group of actions may or may not be executed. The action is structured in the same way that an "if-else if-else" structure written in a programming language. Let's walk through the UI.
The upper section, the "IF" section, is evaluated first. If the condition on the left is true, then the actions on the right are evaluated.
The middle section, if "ELSE IF" section, is evaluated next if the IF condition returned true. You can have as many ELSE IF conditions as you'd like, each with their own conditions and associated actions.
If none of the previous conditions returns true, then the "ELSE" actions are executed, if any.
Conditions are edited through their own UI. Most conditions will be based on Graphical Analysis task quantities, but there are also AND and OR booleans that let you chain conditions together in a very flexible way.
The best part about this is that you can use this, combined with the new "Set Next Event" action, to create conditional, branching trajectories. I've done this is the example UIs that I've been showing here. Note that the "quantity comparison" conditional can compare one quantity against a fixed, constant number like I'm doing here or against the value of a different quantity at that same point in time. This is particularly useful for comparing the values of Plugin Variables, which are really now great places to store numeric data that you might want to compare against as you propagate your trajectory.
Anyway, that's the main gist of things! Here's the full change log.
- LVD: Added new polling and search methods for patternsearch().
- LVD: Warn/alert area is now a UI table.
- The KSPTOT main GUI has been replaced with something that is much more visually indicative of what functionality KSPTOT has to offer.
- KSPTOT UI now supports theming of the colors of all UIs. Access from the new main GUI: View -> Edit Themes and View -> Set Theme.
- New icons for KSPTOT and various KSPTOT tools and applications. There is also a new splash screen shown when loading KSPTOT.
- LVD: GUI areas for initial/final state is now a text area so it can scroll.
- LVD: Added body angular velocity Graphical Analysis tasks.
- LVD: Added body angular rate constraints.
- LVD: Added conditional action.
- LVD: Added Longitude termination condition
- LVD: Added Add value to plugin var value action and other similar actions.
- LVD: Added Set Next Event action.
- LVD: Actions are now sorted alphabetically.
- Fixed bug with Select Departure/Arrival Date button in Compute Departure stuff.
That should be it! Let me know if you have any questions or find any bugs! Happy orbiting!
-
Set Plugin Variable Value: Sets the value of a "plugin variable" to a particular constant number.
-
10 hours ago, RaceToTheMun said:
Corrupted
Hey, it worked for me. Which version of KSPTOT on are you on?
-
50 minutes ago, RaceToTheMun said:
Open up this MAT file in LVD. It should be the optimized trajectory you're looking for!
https://drive.google.com/file/d/14co-L5ztASEKLreLAskrx9YM7Pf1bY2g/view?usp=sharing
-
6 minutes ago, RaceToTheMun said:
Ah well, it took a while but it finished. I have to click the stop button but I choose the left green option iirc but the manuever still won't intersect eve
Can you share a screenshot or the LVD MAT file?
I think I misunderstood at first. I was thinking you were asking how long it took to generate the MFMS data file for LVD. The actual optimization of the trajectory in LVD could take some time, yes.
-
48 minutes ago, RaceToTheMun said:
How long do I wait for this? Do I have to pause it manually or wait for something to pop out
It should be immediate. If you don't see anything, check your KSPTOT.log file for an error message.
-
Hey everyone,
There's another visual change coming to KSPTOT in the next version. It's bothered me for a while now that the "main GUI" was always the porkchop plotter, since that isn't really the focus of the application anymore and hasn't been in years. It stuck around in that role mostly for historical reasons. I've decided to rework the user interface that you're presented with when you start KSPTOT. Going forward, you'll see this instead.
Like the other windows in KSPTOT, this one is theme-able as well. I've moved the old "Tools" menu into this "Android home page" inspired layout and given each tool its own button and icon. The old porkchop plot generator UI is still there under "Porkchop Plot Generator". I think this will A) make it far more obvious that KSPTOT offers more than just porkchop plotting, and B) also make those tools a bit easier to access.
By the way, the icon for the Porkchop Plot Generator is temporary. Unfortunately Midjourney stopped offering trials as I was generating icons, so I don't have that one yet. If anyone has a Midjourney subscription and wants to help me generate an icon for it, please get in touch!
Speaking of icons, spoiler alert, I did pick a new KSPTOT icon and its there in the picture. A new splash screen is coming too that I'm sort of excited to share with you in the future.
Let me know if you have any questions or comments.
5 minutes ago, Nazalassa said:Nope.
I'm honestly not sure then. You have no error messages, no logs, and apparently have permissions for everything. I've never heard of this happening before. Any chance the downloaded files are corrupted? Can you try to download any of the other MCR packages and install them, just to see if they work? Pick a couple from this list and see what happens: https://www.mathworks.com/products/compiler/matlab-runtime.html
-
2 hours ago, RaceToTheMun said:
I copied the Manuever information but it doesn't intersect with Eve
The flyby is Ke-Ev-Mo btwYou're probably running into the "zero SOI radius" assumption. It's a little complicated, but if you want to use MFMS results for actually running your manevuers, here's what you should do.
- Generate the Flyby Maneuver Sequence.
- Tap the "Save Results to File for LVD" button and save the file to some location.
- Open Launch Vehicle Designer (LVD).
- In LVD, File -> New Mission Plan From MFMS Output. Select the file you saved in Step (2).
- Optimization -> Optimize Mission.
See if that gets you a delta-v maneuver that intersects Eve.
[WIN/MAC/LINUX] KSP Trajectory Optimization Tool v1.6.9 [New MATLAB Version!]
in KSP1 Tools and Applications
Posted
I've recreated the bug and I believe I've got it figured out. I reuploaded the PR7 ZIP file. Please go ahead and redownload that and let me know if you have any issues going forward. Thanks!