View Full Version : Flying Model Airplane

03-23-2009, 12:52 AM
I am not sure if this has been tried before, but I would like to fit a model airplane with the necessary equipment to enable it to pilot itself.· The eventual result I would like to acheive being an airplane that, after being loaded with predetermined destination coordinates, would be able to fly from it current position to its destination without any intervention on my part.· The automated flight would include the take off from one airstrip, the flight, and the landing.

My intention is none other than to find out whether this would be possible using the propeller (multiple propellers) chip as the CPU.

I have seen the propeller chip used in the navigation of·robotic vehicles and boats.· I have successfully used a stamp to navigate to predetermined GPS coordinates.· I think its time to introduce microcontrollers to·three degrees of movement.

My request for the forum is·to tell me·how you might accomplish this?· Any ideas are welcome.http://forums.parallax.com/images/smilies/yeah.gif

03-23-2009, 01:11 AM
Cool! A UAV project! I want to play.
Gyros for stability, accelerometers for attitude control and a Compass and GPS for free flight navigation some A to D's for air data, engine sensing and fuel management.
Take off would not be that hard. Landing might be entirely another problem. I'd be tempted to put either an IR or US ping sensor on the bottom of the plane for tighter altitude control near the ground. Your going to have some issues with lining up on the runway for landing so you'll need a runway wider than 2x your GPS EPE or you'll need some type of localizer beacon at the end of the landing strip.
I think the prop would be a great choice for that project as it can handle a bunch of simultaneous processes; altutude control, navigation, pitch, roll and yaw, turn coordination, mapping, power and servo control.
I bet the prop is up to it but it's allot of code.

Signature space for rent!
Send $1 to CannibalRobotics.com.

03-23-2009, 01:41 AM
Check out 'Propellor with Propellors' in completed projects,

Signature space for rent!
Send $1 to CannibalRobotics.com.

Jimmy W.
03-23-2009, 01:41 AM
Pop a camera pointing down to find the landing strip center and keep itself centered on it while on final approach.

03-23-2009, 01:48 AM

Signature space for rent!
Send $1 to CannibalRobotics.com.

Thomas Fletcher
03-23-2009, 02:02 AM
Great site

diydrones.ning.com/ (http://diydrones.ning.com/)

Ken Gracey
03-23-2009, 02:18 AM
This product is based on the Propeller, too:


Ken Gracey

03-23-2009, 07:57 AM
I was actually thinking about using IR for the fine altitude tuning on the landing.· I agree that the landing will be the most difficult part.·

My plan was to first acheive propellar controlled flight where I do the take off and landing but the prop does the in flight part.· Once I was able to fine tune the in flight code I would then work on the take off.· Then, saving the best for last, I would work on the landing.·

There are·a couple of retired runways around here where the hobbyist model airplane pilots hang out.· I eventually would like to be able to have my airplane take off from one and land on another while I follow in a car.· I get so excited about this, I just can't wait to see my dream fulfilled! http://forums.parallax.com/images/smilies/hop.gif· (I am a simple man)

I have the plane that I think will be a great donor·for the project.· I only lack money, the seemingly ever-present obstacle to all my projects.

Thank you for your input this far.· Its fun when everybody has a hand in the project.· Thank you for the websites also, I will look at them now.

03-23-2009, 08:32 AM
here is a queston for you, what is the basic algorithm for going from one gps coordinate to another, i have a project plan for this summer doing something similar, (last project before college thing)

Tom C
03-23-2009, 08:40 AM

To the best of my knowledge, none of the model autopilots I have looked at are able to take off and land on their own.

The autopilots designed and built on DIY Drones require the operator for take off and landing and only go into the autopilot mode for level flight.

If you can pull off take off and landing with a MCU that will fit in a small model plane and be of very low cost, the military will want to talk to you.

Just a thought.


If you are going to send·a Robot·to save the world, you·better make sure it likes it the way it is!

Post Edited (Tom C) : 3/23/2009 1:45:37 AM GMT

Bill Drummond
03-23-2009, 12:44 PM
How about starting with MS Flight simulator?
I have a hack somewhere that sends GPS sentences to the serial port.
Then use the propeller to emulate the yoke an pedals.

Erik Friesen
03-23-2009, 08:01 PM
Have you flown model aircraft before? If not, you will want to learn. You will need a larger model, 7-15 lb, to do what you want. I would suggest using a switch on the transmitter to switch between controls. That way you can debug without wrecking your aircraft because you didn't get something right. Model aircraft are typically only about 1-2 seconds from being a small pile of rubble. Don't forget that they can kill people.

03-23-2009, 08:27 PM
Erik Friesen said...
Model aircraft are typically only about 1-2 seconds from being a small pile of rubble.

One of my all time favorite quotes from a *very* experienced ex-RAF chopper pilot was the average new model helicopter pilot has a MTTS of less than 30 seconds. (Mean Time To Superglue).

Cardinal Fang! Fetch the comfy chair.

03-23-2009, 08:46 PM
To land you will need some sort of on the ground guidence to set the glide slop. Maybe you could use a laser light.

03-23-2009, 09:04 PM
FearTurtles said...
To land you will need some sort of on the ground guidence to set the glide slop. Maybe you could use a laser light.
That's an important point. There's a lot of different ways to fly to a spot on the ground. Most Almost all of them result in a crash. Landing is a process that begins long before you're over the runway, and if you're not set up correctly in advance there's usually nothing you can do to recover and land. You need to be in the right place, at the right altitude, and the right airspeed, and then at the next right place and right altitude and right airspeed, and then the next, and the next, and so on*.

Ever land an airplane? I suggest that in addition to flying a model plane, you get up for a flying lesson or two. It'll give you a much clearer sense of your task. Fly out of a relatively small airport, so you get a sense of what it takes to locate the runway.

* Also going the right direction, at the right throttle setting, with the right flap settings. Hopefully you'll be landing in a field so you can avoid worrying about crosswind landings, which add a couple more variables. I strongly suggest starting that way: landing on a runway is going to be a couple notches more difficult.

Post Edited (sylvie369) : 3/23/2009 2:09:52 PM GMT

03-24-2009, 02:26 AM
Fun Project!- This has had me motivated for several years. Seriously check out the www.DIYDrones.com for a great group of Amateur UAV engineers and developers.

Specifically, there are a number of different projects (and different processors) that we have been building on over the years.


03-24-2009, 07:36 AM

I've been checking out diydrones.com.· It looks like there are quite a few "autopilot" setups that have already been accomplished.· Has anybody tried to fly in anything but a straight and level line?· I aspire to have a bit more complicated flight path.

@FearTurtles, Sylvie369,

I've actually never been inside a small aircraft let alone land one.· I understand the principles of flight for its always been something that fascinates me.· I have flown several model airplanes and have developed·basic flying skills for them.· However, I did find out that the local small airport will let you take an introductory flight in a·small aircraft for about $100.·

The landing concept that I have been considering is a portable laser beam that I can put at the beginning of the airstrip, just as the commercial airliners fix on before they land.· Of course, I would need to adjust·this to a certain angle each time I move it to a new location.


I heard a comparable statement: "There are only two kinds of model airplanes in the·world;·those·who have crashed and those who have yet to crash."

@Erik Friesen

As previously stated, I have flown model airplanes enough to fly ok.· However, I know that I could use a lot·more practice.· I do already have the model that·I would like to convert, I·will·post the dimensions and a picture later on.


Thank you for your recommendations and comments.· You're really giving me a bunch of ideas to use.· I·was pretty sure that the military had already developed a comparable UAV for·infantry recon missions.· I thought that·I saw it in a Popular Science.· However, if Tom C is correct then I·aspire to·accomplish something that·has not been done before.· Imagine a plane put together from parallax parts and a little hobbyist know-how that the military could not achive with there advanced technologies.·

03-24-2009, 08:51 AM
Sadly landing a real small aircraft is pretty nothing like landing a model plane IMO. There are things you can do with a model plane that would be a very expensive and painful day for you in a real aircraft. Such as landing a taildragger in a 3 point landing completely stalled from 5+ wing widths high. But it works great in a model plane, it doesn't exactly look the greatest but....

Hell flying a model aircraft isn't very similar either. Originally I bought a model plane being a pilot and going hey I can fly, Ill be able to fly this thing. It lasted about 10 seconds before it was in many many pieces. I once went out and did an experiment, after flying my 3 channel plane around. I took my real plane up, and tried to turn it purely using the rudder. It doesn't work, aside from being thrown around the cockpit like a rag doll from the side loads. I think it all really boils down to inertia. Even most big models really are not dynamically similar to real small aircraft. Also there is the relationship of Reynolds number similarity, a very complex idea I still do not under stand fully, but short summation because air is a rarefied fluid its effects are not linear with size. Look at how they test aircraft in wind tunnels. It may be a scale model, but the air actually being blown down the tunnel is much higher in pressure (4-5 atm I think for jet speed subsonic flow), so that the model acts the same as the full size plane.

My thought for landing on a point, put a powerful enough motor to allow it to prop hang similarly to a helicopter and teach it to set it self down on that point. Now that would be cool to see. Or take advantage of the model flight characters set it up in a slow stall until it touches the ground.

Since I didn't see this or I missed it (if so sorry). Thought about sensors and GPS. A plane is a pretty fast responding system as for its standard controls. So really I think you will need two underling systems. One that actually flies the plane, probably driven off gyros or accelerometers, and another that tracks it position and goal to fly towards. The parallax GPS unit updates once per second. I would say you would need a refresh rate of 30-40 times a second on your actual flight control sensors so purely using GPS I don't think will work. Also consider how much information can change in 1 second in a model aircraft and plan your sensors around that. Plan for the worst, a gust of wind takes it and turns it over? How would it recover? Would it know its height above the ground? Gps give you altitude, but I live @ 5,000 ft, so AGL is means a lot more. Will it be able to correctly self correct for the wind? If there is any cross wind the heading the plane is pointing is actually different to the heading the plane is traveling. Gps will give you your traveling heading, and a compass would tell you the planes heading. Just some things to consider when laying out your design.

So the 2c summery, treat a model like a model, it can do so much more than a standard aircraft, and use it to your advantage.

I owe everyone here a bunch, So thanks again for answering my dumb questions.
Projects. RG500 ECU system. PropCopter. Prop CanSat. Prop Paste Gun.
Suzuki RG500 in a RGV 250 frame.
Bimota V-Due (Running on the fuel injection system)
Aprilia RS250

Post Edited (TJHJ) : 3/24/2009 1:57:08 AM GMT

03-24-2009, 10:08 AM

Hey now thats an interesting concept for the landing.· I will definately have to consider this in my final plan.· It may require a large weight to thrust ratio due to the onboard equipment and batteries, fuel...· This plane may end up weighing in heavy.·

The way I see it this airplane is not going to be able to rely on GPS alone for its location data.· This produces and interesting question.· How do you navigate in a real airplane? not by GPS alone I suspect?· Could the GPS be used to make general observations until the craft drew near to its target and then using a video camera and an imaging algorithm to locate the airstrip? Afterall, a pilot in a real airplane uses more of his/her senses once s/he is near the runway.

03-25-2009, 12:39 AM
The best way to think about this is to think about an IFR approach - 0 visablity.

In order to line up for the approach there are a number of different options, but the quick summation is for·a percision approach* when you get to 200 ft above the ground you must be able to see the run way in order to land.

Now as for setting up the approach, there are many different types but Ill stick to the 2 most common.

RNAV- or VOR/ILS technology. This was the orginal way this was percision approaches were·done, basically you have a station on the ground that braodcasts a signal that allows you to know your distance, bearing, and speed relative to that station. A series of these were used to allow you to intercept the "Glidslope" / ILS. This is similar technology except that it only broadcasts one heading, you fly until you pick up the signal, then it provides 2d gidance towards the runway. Additional information is used to determine your distance and missed approach point( The point if you cant see the runway you fly up and away). Think of an ILS like a beam being projected from the start of the runway up to the sky. The pilot it trying to fly the plane along that beam. The key is you must always be able to see the runway in order to land, a common misunderstanding. Now see is also a loose definition, as see can mean make out the lights on the runway. Its the pilots choice.

GPS- With the new WASS enabled units gps is now able to give an accurate enough vertical position in order to be able to create a "glidslope", It creates the effect of the same older VOR/ILS technology displaying to the pilot in the same way, but it all being calcualted by the GPS unit. If you want to see what one looks like, check out the Garmin GNS430 or GNS530.

Now if flying VFR(Visual flight rules). I personally use my gps as a referance position to get me to the·airport·but all the setup/ alignment / speed and things that actually come into play landing the plane are done by visual referance, and feeling. I know that when I am even with the 1000' markers on downwind, I should be at 100mph, gear down, and flaps 15 deg. I turn base when the end of the runway passes behind the wing of the plane. On base I want to be at 85mph, 30 deg flaps. Then I turn final so I line up stright. And i want to be at 70mph, 30 deg flaps, and make damn sure the gear is down. I then fly to the edge of the runway and cut power, about 100 feet up. I fly downwards to the runway until about 10 feet, at which I flare the plane to slow the vertical descent, and bleed off the last of the flying speed. Touching down smoothly if all went well.

A thing to remember about planes, throttle controls your altitude, pitch controls your airspeed. It seems wrong at first but its true, a plane can fly nose high very slow, and nose flat it will fly flaster all at the same altitude.Add power in either state and the plane will climb, or visa verse.

Hope this gives you some insight.


I owe everyone here a bunch, So thanks again for answering my dumb questions.
Projects. RG500 ECU system. PropCopter. Prop CanSat. Prop Paste Gun.
Suzuki RG500 in a RGV 250 frame.
Bimota V-Due (Running on the fuel injection system)
Aprilia RS250

03-25-2009, 01:30 AM
Don't trust GPS units on altitude!

Trust me on that...

I hike in the mountains here in Norway, and always carry a GPS. As I often have to stop for a breather on my way up, I use to check the GPS quite often, and one thing I have noticed is that during the 30seconds breather, my altitude may be adjusted as much as 10meters(33') upwards.
And yes, that has happened even when WAAS was enabled and one of the 'W' satellites was available.
Also, it's not unusual to see that the reported altitude is 10meters or more 'off' from what even modern maps lists.

Altitude is the least reliable function out of a GPS. The system was NOT originally designed to deliver that data.

Don't visit my new website...

03-25-2009, 04:38 AM
@Calaway - The Ardupilot is pretty far along in terms of complexity, and for $25 can't be beat (It's way better than mine...) The AttoPilot is looking really really solid based on comments and vids I've seen (and is based around a Propeller). I just need $800 to buy one as it was just released last week. For a very robust open source projects, take a look at the paparazzi project to get your feet very wet. The Paparazzi allows for inlight waypoint updates, hold around a point, etc. Not sure if it lets you do loops...

These are hard tech solutions and there is no consumer friendly version out there just yet. There has been a LOT of thought both amateur and professional on flight control systems, so learn what you can and start playing small.

@Gadgetman - Agreed on the GPS Altitude. Its not all that reliable but better than nothing!


03-25-2009, 06:41 AM
GPS altitude can be accurate but is often thrown off by earth topology. I have a very good GPS(Garmin 60CSX) it can be extremely acurate as low as 2m on flat ground but on the side of the mountain the accuracy always drops drastically.

My GPS uses an on board barometer to make GPS altitude far more accurate though.

Need to make your prop design easier or secure? Get a PropMod (http://propmodule.com) has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.

03-26-2009, 08:43 AM
What a great fuzzy logic problem. At the bottom of the problem its just three axis: X and Y are taken care of by having a landing strip wider than 2x your Gps EPE plus a comfort margin... My RC expirence tells me another 2x. Then it's all Z. Perfect problem for 8 Different processors crawling over the same data set.

Ps: every ex RC pilot, including me has said either; it's not the tool time, it's the part$. OR, it's not the part$, it's the tool time between "landings".

Signature space for rent!
Send $1 to CannibalRobotics.com.

03-26-2009, 10:00 AM
The one military mini-UAV I saw used a "controlled crash" for landing. It got close to the ground in the right place, then went to a high angle of attack where it was moving slowly and descending at a controlled rate. Have you looked at using a slow flying model as the first test bed? Maybe even a slow flying trainer? (does such a beast even exist?)

My 2 cents,

Lunch cures all problems! have you had lunch?

03-26-2009, 10:33 AM
Lawson said...
The one military mini-UAV I saw used a "controlled crash" for landing. It got close to the ground in the right place, then went to a high angle of attack where it was moving slowly and descending at a controlled rate. Have you looked at using a slow flying model as the first test bed? Maybe even a slow flying trainer? (does such a beast even exist?)

My 2 cents,

The Telemasters at Hobby Lobby might be a good test bed. Have good weight capability and·a good·slow fly/landing capabilities.




Tom C
03-26-2009, 06:23 PM

I believe that you are refering to the "Raven".

There is a discussion going on over at DIY Drones as to duplicating the Raven design for a dyi drone.



If you are going to send·a Robot·to save the world, you·better make sure it likes it the way it is!

03-27-2009, 01:50 AM
Yeah, that's the 'Raven' from Aerovironment (www.avinc.com/uas_product_details.asp?Prodid=1 (http://www.avinc.com/uas_product_details.asp?Prodid=1))
It's predecessor, the Pointer (www.avinc.com/ADC_Project_Details.asp?Prodid=43 (http://www.avinc.com/ADC_Project_Details.asp?Prodid=43), did the same thing.

It's a particularly harsh (but simple) way of recovering a vehicle. Rate of descent is actually quite high, and impact is quite severe, so they're designed with bits that pop off, and they may end up replacing bits after some landings.
There's really not much control either.
Google 'Deep stall' (Some airliners had a nasty habit of doing this. not good).
But once again, it's simple.

Now this www.insitu.com/documents/videos/AUVSI%202007%20No%20Nets%20Booth%20Launch%20Recove ry%20Videos.wmv (http://www.insitu.com/documents/videos/AUVSI%202007%20No%20Nets%20Booth%20Launch%20Recove ry%20Videos.wmv) is an elegant recovery method....!