V/S Autopilot request

[desertman]

Anybody interested in making a simple autopilot that uses pitch (elevators, thrusters or both)to hold a given verticle speed while ascending (or decending into) in the atmosphere? This would be nice for Delta glider style ships during ascent and decent.

Anybody have an idea how hard this would be? I would happy to be part of the effort and help out though I dont have much programing experience.

Thanks

Ted

 

[Kveldulf]

I think the DGIV has a V/S hold autopilot, and I think the XR1 and XR5 do as well.

Unless you were looking at more of a uniform autopilot, like an MFD?

 

[tgep]

What about using the Attitude MFD ? You can find it at the Hanger.

 

[ryan]

Alot of the MFDs are quiet complicated, orbiter needs one MFD that can do anything and dead-on simple to use. Now all the MFD pilots are scattered all over the place, one MFD to get into orbit another MFD to de-orbit you.

 

[Woo482]

I've got some think for you Autohover MFD > http://orbiter-forum.com/showthread.php?t=2283

 

[desertman]

woo482, do you know if that autopilot uses hover engines to control v/s?

 

[Woo482]

I think It does but if not this does http://orbiter-forum.com/showthread.php?t=1526 < I use that one with the delta glider IV

 

[Linguofreak]

As far as I know, all VS autopilots for orbiter use hover engines to control VS. If you look at the OP for this thread, however, you'll notice that desertman is looking for a VS autopilot that trims a winged craft (with control surfaces or RCS) such that it ascends/descends at the desired VS.

This is not quite as trivial, however, for a spacecraft as for an aircraft. For one, the better-modeled winged spacecraft in Orbiter (XR1, etc.) tend to re-enter well past their stall angle, so there are potentially two AOA's that could give a given vertical speed. For another, during re-entries from hyperbolic trajectories, the spacecraft will actually be flying upside down to keep itself within atmosphere.

 

[desertman]

linguofreak, I am more interested in using the the autopilot during ascent. Pitch/AOA hold on the winged aircraft require constant attention and in the real word they would probably have an autopilot that controls v/s with pitch and possibly engine gimbal.

The DGEX for instance can only climb at a given v/s. If you try to climb excessively then it looses forward airspeed and can not longer maintain altitude. Constantly adjusting pitch trim for 25 to 30 minutes takes away from the simulation. My opinion.

Correct me if I'm wrong but the same code that the DGXR1 uses to hold AOA could be adapted to v/s hold.

 

[Hashy]

An "autopilot MFD" of sorts that I found which is pretty much up this alley is "Redshift". However, it is incomplete, buggy, and not in developement.. Pretty much I can use it to get to the station & moon.. but beyond that, and outside of that, it should crash your Orbiter client nice and good.. =)

Probably the reason for a lack of autopilot MFD's like this, is because of the nature of Orbiter.. I might be more inclined to wanna warp my way to Mars, and enter the atmosphere and land as normal, but many other Orbiteers want the realism of the transit, between Earth & Mars for instance.. Even if they gotta wait two years and spiral around the inner solar system a half dozen times ;D

Would love a good complete Autopilot like Redshift was supposed to be, but I just don't think theres enough of a demand at this time for making Orbiter more like a space game then a sim.

Still, never ask, never receive

 

[computerex]

The thing is there are so many different types of vessels, with all having different attitude parameters that it is hard creating an autopilot to accommodate them all.

 

[spcefrk]

The thing is there are so many different types of vessels, with all having different attitude parameters that it is hard creating an autopilot to accommodate them all.

Then maybe it's high time we divided all Orbiter craft into a couple of types of aerodynamic spacecraft (all pure spacecraft can be their own set as they don't need atmospheric autopilots). A Universal Autopilot MFD would be pretty nice. 1)Set craft type, 2)Set Autopilot (altitude, V/S, etc), 3)Activate!

Just split them into:

Capsules -- no control surfaces, RCS thrusters, Main engines
Lifting Bodies -- no wings but control surfaces, RCS, Mains and/or Hovers
Wing-Bodies -- wings, control surfaces, RCS, Mains and/or Hovers
Aircraft/Rotorcraft -- wings, control surfaces, Mains and/or Hovers

That ought to cover everything. Give each autopilot four different variations on the same algorithm (or a "Not Compatible" warning for improbable vehicle types) and voila it's now just "easier said than done" :rofl:

 

[DevCo]

Then maybe it's high time we divided all Orbiter craft into a couple of types of aerodynamic spacecraft (all pure spacecraft can be their own set as they don't need atmospheric autopilots). A Universal Autopilot MFD would be pretty nice. 1)Set craft type, 2)Set Autopilot (altitude, V/S, etc), 3)Activate!

Just split them into:

Capsules -- no control surfaces, RCS thrusters, Main engines
Lifting Bodies -- no wings but control surfaces, RCS, Mains and/or Hovers
Wing-Bodies -- wings, control surfaces, RCS, Mains and/or Hovers
Aircraft/Rotorcraft -- wings, control surfaces, Mains and/or Hovers

That ought to cover everything. Give each autopilot four different variations on the same algorithm (or a "Not Compatible" warning for improbable vehicle types) and voila it's now just "easier said than done" :rofl:

I've been wishing for the exact same thing for quite some time now, and really it's not as difficult as everyone is thinking it will be, you can make some very simple code that will work for any craft by using a simple learning algorithum Along with what you said above to take different control schemes into account.

Basicly you start with a AoA/Bank hold AP that is very fine, lets say down to single degrees (this is really the one I want, as my one big gripe with the DGIV's PROG104SPEC## AP is that it's not granular enough for smooth re-entry profiles directly onto a target, you have to adjust in HUGE 5deg chunks).

From there you add in an algorithum that monitors the vert speed as compared to a value set in the AP, plus a threshold, and as it drifts, it trims the appropriate control surfaces to maintain the vert speed, as it drifts farther into the threshold it will logarithmicly apply more trim, as if it is bound to the set vert speed by elastic.

As long as the AP is aware of what control surfaces/thrusters to use to effect AoA and Bank, you should be able to specify a vert speed and have the AP maintain that as best it can, if your re-entry profile is off, you will obviously still perish, but not before the AP uses up all your RCS fuel in vain.

*BTW, if anyone DOES know of a mod or ship that allows me to hold AoA and Bank (and adjust them with the keypad,etc) in very fine incriments please let me know (I thought attitude MFD would, but I cant figure out how to use it for anything but pointing at soemthing)

 

[pete.dakota]

It's extremely possible (and more fun) to launch and re-enter manually in winged vehicles. Why all this demand for auto-pilots? I'll never get it...

 

[Cairan]

A few years ago, I had played with coding an atmospheric autopilot for DG-type vessels... It could handle DG-stock, DG-III, DG-IV and DG-Ex... however it went down the drain with my buoyancy add-on's C++ code when my hard-drive crashed without copies of that stuff...

From the old forums:

http://orbit.m6.net/forum/Default.aspx?g=posts&t=4254

 

[reverend]

I've always wanted a V/S autopilot for launching the DG-EX. Something that can hold 20 m/s up to 64km then 0 m/s until orbital velocity. I've done it a hundred times by manually adjusting the trim, it's just exhausting to watch it constantly for an hour. I don't believe there has ever been a spacecraft ever built without some sort of autopilot. a proper simulation is not complete without autopilots.

 

[Cairan]

I've always wanted a V/S autopilot for launching the DG-EX. Something that can hold 20 m/s up to 64km then 0 m/s until orbital velocity.

Yeah that was my motivation back then because I got tired of sheparding the big flying cow all the way to orbit...

Seeing there's actually interest nowadays for it, I might try to dust-off my memory of how I had coded it back then.

 

[Zatnikitelman]

Not only is it realistic, but it is also a good learning experience. I learned how to get to the ISS by watching RedShift do it for me. Granted, the first thing I docked to was my own space station I was building, but still...
It's like what my Java teacher said about why Object-Oriented coding was so great: writing code the first time is fun, but after that it gets boring.
I was flying the stock DG on my laptop the other day and in trying to get into Orbit, I got an ApA of 146Km, then tried to hold 0 V/S to gain the horizontal velocity. It WAS fairly annoying especially when compounded by a Laptop's slightly different control interface.

 

[Linguofreak]

It's extremely possible (and more fun) to launch and re-enter manually in winged vehicles. Why all this demand for auto-pilots? I'll never get it...

Because sometimes the fun is in doing something more complex and you want to be able to take your hands off of one thing so you can concentrate on something else. Sometimes (as a variation on the above), you want to macromanage rather than micromanage. Sometimes the fun is in setting up a scenario a bit more precisely than you can do with your own reflexes, and then flying it by hand once you get to a certain point. Sometimes the fun is more in putting together a little "story" than accomplishing some feat of piloting skill. Sometimes the fun is in sitting back and enjoying the view as a passenger while the spacecraft flies itself.

 

[Cairan]

Sometimes the fun is in setting up a scenario a bit more precisely than you can do with your own reflexes, and then flying it by hand once you get to a certain point.

Two words: Formation flying.