[old thread] Trajectories : atmospheric predictions

[Svm420]

I get this spammed in my log several times a second when entering Mun SOI on a free return trajectory only while in map view. The spam stops if I click the trajectories button to disable it. and return when re enabled. Using latest trajectories and FAR. Let me know if you need more info. Thanks!

Trajectories: WARNING: FAR/NEAR totalForce is NAN (altitude=3000, airVelocity=3000, angleOfAttack=0

[cantab]

Any chance of an update getting Trajectories to work with airbrakes? Using the latest FAR ("Haack") it doesn't seem to take them into account. I have a lander for Serran (New Horizons planet) that relies on them to maintain stability during descent.

[Tarheel1999]

Any chance of an update getting Trajectories to work with airbrakes? Using the latest FAR ("Haack") it doesn't seem to take them into account. I have a lander for Serran (New Horizons planet) that relies on them to maintain stability during descent.

What would the assumption be? Airbrakes deployed as soon as you encounter the atmosphere?

Out of curiosity, have you tried deploying the airbrakes in space to see if it already takes them into account? Or pressing "update" after you have deployed them? Or maybe this is what you mean when you stay it doesn't work with airbrakes?

[cantab]

I'd expect it to use the craft's current state. In my case the plan was to deploy the airbrakes upon or before hitting the atmosphere and leave them out until hitting the ground. (I use the "handbrake", clicking the altimeter icon, rather than the B key for this)/ I don't remember noticing any difference in the trajectory when deploying or retracting , though I didn't try the "Update" function. I'll try and remember when I fly the next mission.

[Youen]

Any chance of an update getting Trajectories to work with airbrakes? Using the latest FAR ("Haack") it doesn't seem to take them into account. I have a lander for Serran (New Horizons planet) that relies on them to maintain stability during descent.

That's definitely possible, but someone needs to take the time to implement it. Also, it would be best to account for all mods that have implemented airbrakes, and the stock ones, so that players don't get bad surprises once they think it just works (at this time, it just doesn't work ; unless FAR includes it as regular drag in its calculations, but I don't think it does). By the way, it's already in the wishlist of the OP.

- - - Updated - - -

I get this spammed in my log several times a second when entering Mun SOI on a free return trajectory only while in map view. The spam stops if I click the trajectories button to disable it. and return when re enabled. Using latest trajectories and FAR. Let me know if you need more info. Thanks!

Trajectories: WARNING: FAR/NEAR totalForce is NAN (altitude=3000, airVelocity=3000, angleOfAttack=0

Thanks for reporting the issue, I'll take a look when I find some time.

EDIT: I've reproduced the issue with the latest version of FAR (don't know if it happened with the previous one). It's not a big issue since everything continues to work fine, but still bad for performances to trigger exceptions every frame, so I'll see what I can do.

EDIT: bug fixed (in next release)

Edited September 5, 2015 by Youen

[Der Kosmos]

Would it be possible to show a landing prediction in the flight view instead of just the map? I ask because while the map view can get you to within a few KM, you basicly have to eyeball it for the final approach. If you want to build a base and don't have wheels, you often end up having to waste a lot of fuel hovering and fine tuning.

It would be nice if there was some sort of visual indicator showing exactly where you will land.

[Kerbas_ad_astra]

One RSS issue: the prediction's been fine*, but it looks like KSC's location is hard-coded. It ends up somewhere in South America, when RSS moves it to Florida, and there are several mods which can move it to other places.

* Edit: After a lunar reentry test, I should qualify this with some caveats, as the simulation is quite sensitive at long distances/times and high speeds. Once you're down to LEO speeds, it's pretty good (10s of KM or better), but don't expect to stick a landing at KSC from a lunar return. For best results, keep a set of RCS thrusters on the capsule itself for any last-minute corrections (dropping the service module can change landing predictions by quite a bit).

Edited September 6, 2015 by Kerbas_ad_astra

[cdugas4]

Does any of the versions in the release log work with .90?

EDIT: Nevermind, I found it on the KerbalStuff repository.

[RidingTheFlow]

Would it be possible to show a landing prediction in the flight view instead of just the map? I ask because while the map view can get you to within a few KM, you basicly have to eyeball it for the final approach. If you want to build a base and don't have wheels, you often end up having to waste a lot of fuel hovering and fine tuning.

It would be nice if there was some sort of visual indicator showing exactly where you will land.

If you read a wishlist in OP you will see that this feature already there.

But yes, I'd agree it would be immensely useful, especially for precise landings "near flag" (eyeballing via map view is just not suited for the job).

[Youen]

One RSS issue: the prediction's been fine*, but it looks like KSC's location is hard-coded. It ends up somewhere in South America, when RSS moves it to Florida, and there are several mods which can move it to other places.

Indeed. As a workaround, you can use maneuver nodes to place the red cross where you want, and then set the target to the current impact. Or you can find out the longitude/latitude of the new space center, and enter them in the corresponding fields. And if you tell me what the location is, I should be able to hard-code it in future versions, to replace KSC location when RSS is detected.

[Svm420]

Indeed. As a workaround, you can use maneuver nodes to place the red cross where you want, and then set the target to the current impact. Or you can find out the longitude/latitude of the new space center, and enter them in the corresponding fields. And if you tell me what the location is, I should be able to hard-code it in future versions, to replace KSC location when RSS is detected.

Hard coding would be less desirable than a nice easy to sub config file. Preferably able to be changed via MM. That way no further compatibility issues with those suing KSCSwitcher.

[Youen]

Hard coding would be less desirable than a nice easy to sub config file. Preferably able to be changed via MM. That way no further compatibility issues with those suing KSCSwitcher.

Makes sense, but how do you provide a global value that can be changed with MM? Is there another way than creating a KSPField on a PartModule (though that wouldn't be a too bad idea as I already have a PartModule that stores per-vessel trajectories settings)?

[mhoram]

@Youen:

Yes there is.

have a look at this configfile:

https://github.com/mhoram-kerbin/TrackingStationAmbientlight/blob/master/GameData/TrackingStationAmbientlight/TrackingStationAmbientlight.cfg

and the according sourcecode:

https://github.com/mhoram-kerbin/TrackingStationAmbientlight/blob/master/Source/TrackingStationAmbientlight/TrackingStationAmbientlight.cs#L44

[RidingTheFlow]

For landing near targets it would be helpful to:

- be able to set target cross from KSP targeted craft/flag/debris (assuming it landed of course)

- display numeric distance between predicted impact and target. This will help to achieve precise arrival on air-less bodies, when doing it by aligning crosses on map view is too fiddly

Could this be added to wishlist?

[Spheniscine]

One possible idea to check off that elusive staging feature-request, not sure if it's any good, just throwing it out there:

Have an interface for quicksave and quickload specific to Trajectories. The idea is that you Trajectories-quicksave, then activate the staging you want, then Trajectories-quickload. Past the quickload, Trajectories will still remember the characteristics of the post-stage craft, and can make predictions based on that.

[Gryphon]

One possible idea to check off that elusive staging feature-request, not sure if it's any good, just throwing it out there:

Have an interface for quicksave and quickload specific to Trajectories. The idea is that you Trajectories-quicksave, then activate the staging you want, then Trajectories-quickload. Past the quickload, Trajectories will still remember the characteristics of the post-stage craft, and can make predictions based on that.

It's an interesting idea - saving a reentry profile as part of a craft -but I am not sure it would work, at least with FAR. IIRC, Trajectories just uses the information which FAR produces for the active vessel.

Edited September 13, 2015 by Gryphon

[Spheniscine]

The idea is that upon Trajectories-quickload, Trajectories then stores that data somehow in its memory, independent from the savefile, thus allowing it to be used after the quickload. The process I imagine is something like this:

User Trajectories-quicksaves

User then activates stage

User activates Trajectories-quickload:

• Trajectories saves reentry profile data, THEN quickloads while still using the data it saved.
  

Edited September 16, 2015 by Spheniscine

[nablabla]

hi guys,

I am still impressed by your work, just how accurate it is. It makes so much sense to have it!!

Anyways here is something: First I thought it is super inaccurate, but the problem is:

Even if the orbit is a not an orbit but an escape trajectory, Trajectory computes the trajectory after a full or multiple orbits (which is not going to happen).

This confuses, because it looks like my craft falls down or is captured, but in reality it just passing.

[Mitchz95]

^ That has happened to me as well.

[Spheniscine]

I have experienced the problem too. There are three workarounds:

1. In such a situation there are usually some telltale "hairs" that appear near atmosphere exit, where Trajectories predict escape trajectories but fail to render it correctly.

2. Reduce the number of patches that Trajectories predicts to the minimum, then slowly increase it. If no orbit appears in the patch after atmosphere exit (or if the orbit is extremely high and would push you out of SOI), you're not gonna be captured.

3. Try using a maneuver node to lower your orbit. If the orbit "jumps" after a certain point as you're decreasing (or you notice those "hairs"), then you know you won't get captured.

Edited September 21, 2015 by Spheniscine

[cttw]

Is there a way to make this mod account for airbrakes? I've been doing a lot of laythe landings and I found that Trajectories is pretty accurate if I don't deploy airbrakes, but if I do deploy them they are not taken into account. I've fidgeted with the prograde/retrograde setting with the airbrakes deployed but it still didn't seem to take them into account. Airbrakes is how I reentry things that are not round.

Many thanks for this awesome plugin.

[nablabla]

I already had the idea, unfortunately you can not set it to less then 3 "patches", I am not shure what patches are, I thougt its the numbner of orbit circles it computes.

I should detect soi changes and then stop computing future orbits.

[Youen]

Yes, it's a known bug, I haven't got time to tackle it yet. And it is indeed really misleading (I failed an aerobrake because of it too, before I understood what was going on). Hopefully I'll be able to fix it in the next release. But I can't tell when that'll be. Somehow, it escaped being noted on github until now.

- - - Updated - - -

Is there a way to make this mod account for airbrakes? I've been doing a lot of laythe landings and I found that Trajectories is pretty accurate if I don't deploy airbrakes, but if I do deploy them they are not taken into account. I've fidgeted with the prograde/retrograde setting with the airbrakes deployed but it still didn't seem to take them into account. Airbrakes is how I reentry things that are not round.

Many thanks for this awesome plugin.

This has already been suggested, and looks like a good idea, but I'm not sure I'll have time to implement it. Anyone with a bit of programming skill is welcome to take a look and send me a pull request.

Edited September 24, 2015 by Youen

[nablabla]

no panic

in the end it will all be there

[Nothalogh]

So, using FAR and a blunt body reentry, I find my trajectory is wildly off.

My actual landing site is in the mountains west of KSC when I aimed for the ocean east of KSC