Robotic Assistance for Sailplanes - RC Groups
Thread Tools
Feb 01, 2012, 10:52 PM
Registered User
aeronaut999's Avatar

Robotic Assistance for Sailplanes

I may be posting in the wrong department, I usually fly slope, but I thought I might start out here and see what comes to the surface.

For assistance in flying at very long distances and in other challenging circumstances (such as flying while hiking cross-country), I was wondering if any of the following ideas were already available or might be attained in the near future, at an affordable price and compact size / light weight ( I want to install in a Bird Of Time and call it the BotBot....)

* Push a button on transmitter to activate a "heading hold" function (maybe based on heading derived from piezo gyro, or maybe an actual direction-of-travel-hold derived from GPS data?)

* Push a button on transmitter to activate a "position hold" function (ideally with in-flight-adjustable limit to aft stick travel so the robotics won't pull the glider up into a stall if winds are light.) (Position hold is X Y only, altitude is free to vary.)

* Push a button on transmitter to activate "fly toward pre-programmed location" function (again X Y only, altitude is unconstrained).

* Bank angle (or alternatively turn rate) limiter function- so you don't accidentally steer sailplane into a very steep bank /spiral dive when flying near the limit of visual range

* As an aside I'll note that I suspect that turn rate (yaw rate) data is underutilized in RC autopilots; while flying hang gliders blind I've found that things go much better when you consult turn rate (yaw rate) data as well as GPS heading data, rather than GPS heading data alone...

* Thanks for your feedback of any kind. I am new to this field. I am sincerely seeking something that could be economically added to an RC sailplane for a robotic assist to ordinary visual flying at distances approaching the limit of visual perception.

Last edited by aeronaut999; Feb 02, 2012 at 12:49 AM.
Sign up now
to remove ads between posts
Feb 02, 2012, 10:30 AM
Registered User
If you wish to follow FAA regulations and not put RC flying into any more jeopardy than it already is, the first question to be answered is: When does RC glider become UAV? When considering this, try to think like a FAA bureaucrat and forget anything you know about RC flying. The big issue is the "see and avoid" principle used to maintain separation from other air traffic.

Im not saying that what you propose wouldn't be nice to have, but legally it is a can of worms that imho is better left unopened.
Feb 02, 2012, 12:39 PM
AMA 5285 LSF 8104
Aquila Guy's Avatar
There has been a recent interesting discussion on another thread here in the XC forum about using some form of autopilot (see SpeedPilot). Plus there was an old thread about the AFOLT experiment which was a sailplane with an autopilot.

It boils down to a legality question. AMA and FAI (if I remember correctly) do not allow use of automated controls, the model must be flown at all times by the pilot and remain in visual contact at all times. Plus with the magnafying glass of the FAA focused on this hobby, trying to keep the seperation of the hobby from UAV's is a problem. I believe most experimentation now is in the telemetry system developement.

Feb 08, 2012, 01:53 AM
Registered User
aeronaut999's Avatar
Ouch that hurts-- thinking like an FAA bureaucrat that is --

Well even a simple "heading hold" function would be extremely helpful-- and well within the confines of what is being very routinely done in the RC world (don't many of the giant scale guys use some kind of yaw-damping gyro?)

I think I'll start with that

Any suggestions as to the best (affordable but well-functioning) setups for "heading hold" at the touch of a button?

PS my home site is unregulated so AMA rules are irrelevant...and with "heading hold" only, the model still must be flown at all times by the pilot, obviously, if I lay down for a nap it may drift forwards (or backwards!) into some other county....


Feb 09, 2012, 02:19 PM
Registered User
This has just given me a though. I'm asking it here as there are far more knowledgable, smarter people then me who could answer it! I suspect that one of the things that is hardest for us as RC pilots to control is slip in turns. We don't have the cotton thread taped to the canopy to be sure that our turns are perfectly co-ordinated!

How would a gyro on a rudder work in turns? Could it be fooled into working as something that maintains zero slip in turns (i.e. zero yaw) rather than a heading hold in a DIRECTION?

Just wondering out loud really.

Feb 09, 2012, 06:21 PM
Pat Chewning
Ummmm..... don't we already have these? They are called FREE FLIGHT gliders.

i.e. Gliders that are not under pilot control.
Feb 09, 2012, 11:12 PM
Build to Fly? FLY to BUILD!
Legot's Avatar
I'm pretty sure you can do this with ardupilot.
It shouldn't be too hard to adjust for no throttle.

lots of info at
Feb 11, 2012, 10:50 AM
Registered User
Legality aside, the real issue in autonomous cross-country soaring is recognizing and exploiting lift.

Some years ago, NASA purchased a typical model glider and installed the necessary processing and sensors. I can recall that it was winch-launched, and flown over the dry lake bed at Edwards AFB. . It was able to stay aloft, autonomously, for at least an hour (if I recall properly) without human input of any kind. I am sure there are reports and information available to the public.

It was at least ten years ago, and possibly 15-20 years ago, at roughly the same time that the Aerosonde folks flew the Atlantic with autonomous flight vehicles.

Yours, Greg..
Feb 18, 2012, 11:55 AM
Registered User
aeronaut999's Avatar

gyros, yaw strings

Originally Posted by Zimodile
This has just given me a though. I'm asking it here as there are far more knowledgable, smarter people then me who could answer it! I suspect that one of the things that is hardest for us as RC pilots to control is slip in turns. We don't have the cotton thread taped to the canopy to be sure that our turns are perfectly co-ordinated!

How would a gyro on a rudder work in turns? Could it be fooled into working as something that maintains zero slip in turns (i.e. zero yaw) rather than a heading hold in a DIRECTION?

Just wondering out loud really.

A gyro would not be of benefit in controlling slip in turns, any more than a turn rate gyro can be used to control slip in a full-scale aircraft. If someone were really serious about this, they could make a little wind-vane to sense and correct slip via an automated rudder input.

Question-- where should the vane be located? Not necessarily at the nose. Back by the vertical fin would be arguably just as good. Keeping that point streamlined to the flow would allow a little bit of (slipping) cross-flow over the more forward parts of the model, so that dihedral generates some beneficial rolling-out torque. Since all parts of the model cannot be simultaneously streamlined to the (curving) flow, why not streamline the vertical fin to the flow? Or even allow just a bit of slip at that point, so you are using the whole vertical fin, not just the rudder, to generate the yawing-in torque that is always required to some extent. This would minimize the rudder deflection and may create less drag than trying to have zero slip at the vertical fin...

Guess I'm getting off into a discussion of airflow curvature / curving relative wind here-- anyway a gyro is not going to do anything to minimize slip or center the yaw string....

PS hang on a minute-- rather than dangling another appendage into the flow, slip could be detected by an electronic inclinometer-- the electronic equivalent of a slip-skid ball-- this is simply a tilt meter-- whenever the apparent G-load acts "straight down" in the aircraft's own reference frame, the net aerodynamic sideforce is zero which implies that on the average, the aircraft is meeting the flow "head-on" not sideways. This is slightly complicated by rudder deflection but it's close enough that most pilots never give it a second thought.

The device would eliminate slip simply moving the rudder in whatever direction the "ball" is sensed to be deflected.

Again there are arguably some benefits to allowing a slight amount of slip in a sailplane-- so the point where the line of the fuselage is tangent to to the curving flow lies not at the CG but well aft, near the vertical fin or even behind the vertical fin-- but this could be programmed in, just allow a slight slip. But then the device has to know which way you are turning, so it knows whether it is allowing slip or skid-- hmmm... so now it has to work in conjunction with a turn rate indicator... hmmm... simpler for the device to be programmed to just keep the "ball" completely centered at all times.

It would be most interesting to see whether an aircraft equipped with such a device would tend to out-perform a sailplane without such a device-- would probably depend a lot on the rudder mix the pilot was using etc-- again there is sometimes some benefit in allowing a bit of slip in a thermalling sailplane and this will be especially true for small-scale model sailplanes....

PPS However, such a device might open some cans of worms aerodynamically. It might work too well. For example, if we are truly eliminating slip, then there is no point whatsoever in having dihedral, as the dihedral will never feel any sideways flow and so will never generate any roll torque. A model with flat wings will fly exactly the same as a model with generous dihedral.

Since we don't see this in actual practice, this is a very strong argument that we ARE allowing a non-trivial amount of slip in our thermalling turns. It would be very interesting to verify this simply by videoing a yaw string mounted on a little vertical wire by the nose or by the CG, viewed by a little mini video camera on the vertical fin. In the interest of experimentation, the sailplane could be flown in a steady thermal turn in smooth air with the rudder held in several different positions. A data logger could collect sink rate data and the camera would observe sideslip. Post-flight, the different (known) rudder positions could be correlated with the slip angle and the sink rate. Questions of interest would be-- can we document that some sideslip as measured by the yaw string at the CG does in fact give a better sink rate than no sideslip at all? Meanwhile an on-board inclinometer could be also be recording the forces, so that we can get an "apparent tilt angle" (i.e. ball deflection) for the ideal amount sideslip.

If best results were obtained by a non-zero amount of sideslip, then our automatic turn-coordinator device would need to be programmed to work in conjunction with a turn direction/ rate gyro and programmed to allow some amount of sideslip in the direction of turn...

Well here's another approach that might allow us to dispense with the turn direction/ rate gyro and just use the inclinometer. What if the device were simply programmed to always apply slightly less rudder than is needed to fully center the ball? It will never in any circumstance apply rudder opposite to the "ball deflection", rather it will simply always apply a little less rudder than is needed to fully center the "ball". Theoretically, this could allow a little slip, or a little skid, whichever the model wants to do. But aerodynamically, we know that in a constant-banked turn, the model always wants to slip, never skid. During rolling motions (bank angle changes) this small amount of uncoordination will have a fairly negligible effect, as it will be quite small. Aerodynamically, we know that the ball always tends to swing toward the descending wing during rolling motion, due to adverse yaw (technically we could call this a slip when the bank is increasing and a skid when the bank is decreasing?) If the pilot is worried about this, he can add a bit of rudder input during the rolling motions to augment the input from the device and bring the ball all the way to the center, but this probably isn't needed -- by keeping the ball very close to center (but not quite centered), the device is probably doing a better job than the human pilot could do from the ground without benefit of the yaw string.

Yet another approach-- simply use a windvane or inclinometer to transmit slip information to the human pilot on the ground and let him deal with it!

An inclinometer might not work as well as a windvane because high-performance RC gliders are so sleek that relatively little aerodynamic sideforce (what the inclinometer senses) is generated during a sideslip.... regardless of whether the info is being used by a human pilot or a robot helper.... but if the device were sensitive enough it still might work....

Last edited by aeronaut999; Feb 19, 2012 at 12:18 PM.
Feb 18, 2012, 10:38 PM
Registered User
aeronaut999's Avatar

Yaw string video experiment

Well, I mounted a yaw string on a post near (just aft of) the CG of my 2-meter Spider, and mounted a video camera on the fin, got a beautiful 40-minute sunset flight but when I landed the camera was off- have no idea how much footage I got Bummer!

But I don't have any doubt about what the results were-- the yaw string surely streamed a little toward the outside of a constant-banked turn (showing some sideslip). I think this is going to be true of all aileron-controlled RC sailplanes with any normal rudder mix and any normal flying style (rudder usage). After all most RC pilots feel-- I believe-- that a wing geometry with lots of dihedral contributes a rolling-out torque in a constant-bank turn, causing the glider to need less outside aileron input (or in extreme cases, more inside aileron) input than would otherwise be needed. For example if the model requires zero aileron input in a constant-banked turn, and you then remove some dihedral from the wing, it will then tend to roll into the turn and will require some outside aileron input. Is this not the experience of most RC glider pilots?

This can only happen if the model is being allowed to slip a bit, so that the dihedral "feels" a sideways flow and thus creates a "downwind" roll torque-- toward the outside wing or high wing. With no slip, the dihedral creates no roll torque, and has no effect on the overall balance of roll torques acting on the model.

Allowing a bit of slip makes sense from a performance standpoint as well as a handling standpoint. Assuming that the model has some dihedral, isn't it more efficient to create some sideways flow over the wing and allow the dihedral to create some rolling-out torque, than to keep the model completely coordinated in yaw (slip) and rely solely upon deflecting the ailerons to create the rolling-out torque? (After all, some rolling-out torque is always needed to offset the rolling-in tendency that we would always see in a flat-winged model, or a model with dihedral that was not being allowed to sideslip.) (This rolling-in torque arises from the fact that the outboard wingtip is travelling faster, and covering more distance, than the inboard wingtip.)

Likewise consider the balance of yaw torques-- the outboard wingtip moves faster, and covers more distance, than the inboard wingtip, and so the outboard wingtip experiences more drag than the inboard wingtip. This creates a yawing-out torque, which tends to make the nose point some degrees toward the outside or high side of the turn, relative to the direction of the flight path at any given moment. Does it really make sense to try to eliminate this sideslip completely? Regardless of what the slip angle ends up to be, something must be creating the yawing-in torque that counterbalances the yawing-out torque from the tip drag of the outboard wingtip, bringing the net yaw torque to zero,so that the glider ends up with some constant yaw rotation velocity (zero yaw rotational acceleration.) What is the best way to create this yaw torque? If the fuselage is slender and streamlined (doesn't create lots of drag in a sideslip), then isn't it more efficient to allow the vertical fin to meet the air at a slight sideslip angle, so that its "airfoil" flies at a non-zero angle-of-attack and generates some yawing-in torque, rather than to rely solely on deflecting rudder to create the yawing-in torque, as would be the case if the turn were completely "coordinated" (zero slip as measured at some particular point--say for the sake of this example, as measured at the vertical fin)?

I guess this fin-based argument for allowing sideslip vanishes if the vertical fin is all-moving (no separate rudder)-- then you can have your cake and eat it too-- use the whole fin efficiently not just the rudder, but also keep the fuse streamlined with the airflow-- but it still might be more efficient to allow a bit of slip so that the wing's dihedral creates some rolling-out torque, so that less (or zero) outside aileron deflection is needed to maintain a constant bank angle.

Of course if the rudder is simply left centered in a constant-bank turn, then the vertical fin MUST be meeting the air at non-zero slip angle and creating a yawing-in torque. This is the only way that the glider can counterbalance the yawing-out torque from the increased tip drag of the outboard wingtip, bringing the net yaw torque to zero.

I suppose an exception might be a glider that needed lots of outside aileron input in a constant-banked turn, and was rigged with zero differential aileron travel-- then maybe the adverse yaw from the aileron deflection could create enough yawing-in torque to make the nose point toward the low side or inside of the turn, so that the whole fuselage (including the tail) was feeling a skidding flow (flow toward the low wingtip) rather than a slipping flow (flow toward the high wingtip)? But that would be a non-typical situation I think. Maybe we would see this when flying inverted-- we typically need lots of outside aileron to hold the bank angle when inverted, and if the ailerons had differential throw when upright, they'll have differential throw in the wrong direction when inverted! The only problem with this idea is, the skidding flow would interact with the inverted dihedral (anhedral) to create a rolling-out torque, removing the need for the outside aileron input. The fact that we do typically have to hold lots of outside aileron when inverted suggests that the wing is feeling a neutral (non-slipping) flow, or a flow toward the high wingtip (sideslip), presumably due mainly to the fact that the outside wingtip is travelling further, and creating more drag, than the inside wingtip, yawing the nose to point toward the outside or high side of the turn.

Finally consider that since the flight path is curved, the relative wind is also curved, but the fuselage is not curved. Just as the span of the glider is non-trivial compared to the turn radius (creating the difference in airspeed between the two tips), so too is the length of the fuselage non-trivial compared to the turn circumference. Or to put it another way, there is some significant curvature in the flight path and relative wind even along the length of the fuselage. Therefore the fuselage cannot be streamlined to the curving relative along its entire length. Let's say for the sake of argument that we decided that we did in fact want the vertical fin to be completely streamlined to the flow. Then every point forward of the fin must feel some slip (airflow toward the inside of the turn), unless we somehow make the fuselage curve like a banana to remain parallel to the curving flow along its whole length. Again this argument makes the most sense in the context of a glider with a slender streamlined fuselage-- in something shaped more like "Le Fish" or a Schweizer 2-33, we might want a point near the CG, where the fuselage has a broad flat side, to be streamlined to the flow.

Another way to look at the curving relative wind is to recognize that the glider is rotating as well as translating linearly. The rotation creates the difference in relative wind direction along the fuselage length. If the glider were only rotating, the difference in relative wind direction between the nose and tail would be obvious. We still have that same rotation even in a normal turn, where the rotation rate equals once per 360 degrees of turn.

Well I'll have another go in a few days. I predict that the yaw string mounted just behind the CG will deflect toward the outside of a constant-banked turn. If I add a second yaw string at the nose, it will be interesting to see if it deflects noticeably more than a yaw string mounted further back. It will also be interesting to see if the yaw strings deflect toward the outside or high side of the turn even during an inverted constant-banked turn....

Last edited by aeronaut999; Feb 19, 2012 at 02:07 PM.
Feb 19, 2012, 01:15 PM
Registered User
aeronaut999's Avatar


I did get some footage of the yaw string. Rather than posting slope-soaring videos to this forum, I'm going to resume this branch of this discussion here in the "sailplane" forum-- Curving relative wind, how much slip is ideal in turns, yaw strings videos .

Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Discussion RC Heli First Timer - Waiting for Z008 To Arrive and Seeking Assistance Meggy Coaxial Helicopters 10 Nov 21, 2013 09:20 AM
Question Quad almost ready for flight.. need assistance BobsQuad Aerial Photography 5 Sep 25, 2011 09:08 PM
Discussion Assistance with the motors and Escs for payload - Tricopter Novello Multirotor Drone Talk 2 Jul 25, 2011 05:18 PM
Discussion Requesting assistance with flying training for a mental block Sackie Sport Planes 3 May 01, 2011 09:38 PM
Discussion Robotic simulator surface robotic project possible glue-ee treehog Robotics 0 Oct 02, 2006 03:17 PM