Glider with GPS guided recovery

The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

delta22

Well-Known Member
Joined
Jan 21, 2009
Messages
1,015
Reaction score
7
Fellow CMASS flier Dave Lindbergh has been doing advanced electronics work and field tests in the pursuit of GPS guided rocket recovery.

Details at: https://nerdfever.com/?page_id=695

So far Dave's flights have used a steerable parachute controlled by a servo, logic unit and GPS.

Over the last several years, there have been some successes in the pursuit of this very ambitious project. However, the complex and fast moving dynamics associated with steering a parachute continue to be a very tough challenge.

We discussed the possibility of replacing the steerable parachute with glide control surfaces on the rocket.

Dave invited me to collaborate with him on this idea and I immediately agreed.
 
The rocket is a delta wing designed for fast gliding.

The flight profile we want to use is:

1) straight up liftoff from normal rail

2) near apogee, control unit (with input from GPS and altimeter) adjusts elevons to glide position and guides rocket on a fast glide back towards the pad

3) at 100-200 ft altitude over the pad, elevons are raised to flare rocket and a conventional chute is deployed

This rocket will be based on a LOC 3.1" diameter body tube and nose cone, and fins/wings will be 1/4" balsa with reinforcement at the edges.

I built a 32% scale of this for testing.

Attached are RockSim files of both these projects.

View attachment GPS3.rkt

View attachment GPS3small.rkt
 
Several pictures of the small test prototype.

Built with LOC 1" body tube and 1/8" balsa fins/wings.

Elevons are straight for liftoff and lifted for glide recovery.

Parachute deploys conventionally by separating nose cone from body tube. However, a Kevlar line on the outside of the rocket connects chute to BACK of rocket body. This is to prevent the control surfaces from striking the ground during recovery. Line can be seen in last picture.

IMG_0305.jpg

IMG_0307.jpg

IMG_0308.jpg

IMG_0310.jpg

IMG_0313.jpg
 
This design is intended for straight up liftoff from standard rail and a FAST glide recovery (think drogue replacement), followed by chute recovery.

The rocket needs to glide fast enough to easily cut through the wind and guide itself back to the pad.

As a result, planned wing loading is at the heavy end of the normal range for an RC plane.

Attached are a couple screen shots of calculated values for our proposed full scale glider, from this web page: https://adamone.rchomepage.com/cg_calc.htm

Design started by examining the very cool Delta Star rocket glider found here: https://www.randrmodelaircraft.com/Delta Star.htm Then design was modified for build simplicity and a fast glide.

aero1.jpg

wl1.jpg
 
Wing Volume Loading takes into account the fact that volume and weight scale with a cube factor and area scales with a square factor.

It allows for meaningful comparison a wing loading regardless of scale. It accurately scales from small models to full sized airplanes.

WVL is defined as (weight of model in ounces) divided by (wing area in square feet raised to the 1.5 power)

Some typical values:
sailplane 4-6
aerobatic 8-10
sport 10-11
heavy warbird 11-13

Plugging in values for this project:

350 sq in wing area / 144 sq in / sq ft = 2.43 sq ft
2.43 to the power 1.5 = 3.79
48 oz rocket weight / 3.79 = 12.7 WVL

Another definition of WVL uses the area and thickness of the wing instead of raising area to the 1.5 power. This method is not used here because these wings are flat.
 
Any and all input is welcome.

Questions:

> I made the model with about 5 degrees of dihedral, is this a good value?

> Elevon control surfaces are each 10" wide (width of wing) and 3" long, is this a reasonable size?
 
For a fast gliding airplane like this, those elevons are HUGE. Full width is good, but they don't need to be nearly so wide/thick/deep whatever the word is. The width will make it very twitchy as the trailing edge is so far from the hinge.

Compare to the delta star / gamma star designs. The full throw (up for instance, from neutral) on my gamma star is about 1/4 inch at the trailing edge. While gliding slowly a full up elevator will stall the plane immediately. At landing you can work to slowly apply elevator, ending up in full stall and a "plop" into the grass.

Will a vertical flight result in enough lateral distance for the GPS to get away from the RTL coordinates? The google maps I've seen produced by GPS autopilot tracks show very large circles around a point. The error in the GPS requires you to have a fairly broad buffer zone around a point or the autopilot will end up chasing the point all over the place.

N
 
Sounds to me like you've thought this thing through. I'm not super tech savy so I won't address your gps recovery. That said, I think your on the right track. Your dihedral is probably a good value. Lets call it like it is --it's a good round number to work off of-esp with a delta. I've done a few delta's and while fast they are stable and forgiving for the most part. A couple things to watch out for ---it sounds like you know this--watch your CG. It looks like your carrying a bit of payload in the nose, remember your flying the wing not the control surfaces, they are just there for adjustment.The concern is not the size of the control surfaces but thier deflection. You want to fly the nose up --not twist it up or muscle it up. Accelerated stalls are no fun and its part of the territory with high wing loads, perhaps some kind of progressive deflection would help.Remember its a delta so speed is life--don't bleed it.I think you've got a great thing going here just don't rush it. By the way I know the whole CP-CG challenge and the delta is the way to go if you don't mind the speed---but then there's always the forward swept swing wing:rolleyes: Keep this posted I'm interested
 
Last edited:
Hornetdriver makes a good point re control surface size. It's true, it IS the amount of deflection that matters. Not really measured in inches at the trailing edge, but in the total area that is deflected. A large and deep control surface will present a lot of area with very little trailing edge movement. Servos and the simple linkages used are limited in their resolution so it may not be possible to move a large area a super small amount.

Find a hill and do some toss glides with your scale model to see how it flies. This will give you and idea of where your CG is/should be. A properly trimmed RC model will fly fairly well with the radio off.

N
 
Thanks for the input gentlemen.

I did a glide test three times with the small model by tossing it fast from a second floor window, much to my wife's amusement. It dropped 15" and traveled 30" horizontally each time maintaining a fairly steady attitude. The NC stuck firmly in the soft ground each time but the rocket is not damaged.

Thought I had secured the nose weight adequately, but later found it was about 3/4" too far forward. The elevons (1" wide in this 1/3 scale model) were deflected 1/2" up.

I realize in this position they are attempting to fight the nose up. This flight profile is halfway between a glide and a controlled fall.

Will repeat this test adjusting weight position.

This is the logic I used proposing large elevons, please let me know if some assumptions may not be accurate.

The Delta Star has elevons that are about 1" wide and 15" long on each wing. It weighs about 24 oz and is 36" long.

The 3" GPS rocket is proposed to have wings with a 10" span, weigh about 48 oz and be 48" long.

Starting with the Delta's 1" elevon width
x 3/2 (smaller GPS wing span)
x 4/3 (longer GPS rocket)
= 2"
plus some more to adjust for weight and high wing loading.
 
Completed a series of "glide" tests.

Rocket preparation:

1) Tweaked RockSim model to exactly match the pictured small glider.

2) Marked calculated Cp and desired Cg for a stability margin of 1.0 on rocket.

3) Adjusted internal weight in small glider to match desired Cg and wing volume loading expected for larger 3.1" model.

Test preparation:

Repeatedly threw small glider from second floor window as previously described. All flights dropped 15ft to lawn and traveled roughly 30ft forward.

Due to adjustments noted above and careful verification of actual Cg before each "glide", Cg was about an inch further back than the glide tests described in an earlier post.

As seen in pictures the elevons were held in position with tape.

What was varied in the test series was the width of the elevons. Between each round of glide tests, I trimmed 1/8" off the width of the elevons.

Test results:

1) Elevons 1" wide (in this small model, scales to 3.1" wide in larger GPS model). Threw rocket 3x, each time the rocket flared the nose up, then landed on its belly.

2) Elevons 7/8" wide, 2x throws, very slight flare.

3) Elevons 3/4" wide, 2x throws, steady attitude throughout flight.

4) Elevons 5/8" wide, 2x throws, steady attitude throughout flight.

5) Elevons 1/2" wide, 3x throws, rocket dropped its nose as flight progressed.

This result suggests that for this small glider set up as noted, elevon widths in the 5/8" to 3/4" range worked best.

Scaling this up by 3.1x for the GPS glider, this suggests elevon widths of about 1.9" to 2.3".

This value is in line with the rough calculation noted in the previous post and moves in the direction suggested by Wingarcher.

Picture is of glider after final trim and last glide test. I am pleased that it is still in good shape after plunging it's nose into the lawn well over a dozen times.

IMG_2167.jpg
 
Last edited:
Very cool build! What kind of electronics are you using? What kind of GPS module is the final build going to use?
 
Interesting project. I have an ambition to do something similar, but with forward-stowed wings that would swing out. I know a guy who cuts RC wing foam cores, and I'm working on some prototypes of RC aircraft electronics.

Are you thinking about doing anything to shift the CG back for glide mode? If you have the CG far enough forward for good rocket stability, you'll need an awful lot of elevon deflection (and associated drag) to keep the pitch controllable.
 
Very cool build! What kind of electronics are you using? What kind of GPS module is the final build going to use?

I'm doing the easy part, building the rocket glider, so I am not familiar with the details of the electronics.

Dave has a lot of information posted at: https://nerdfever.com/?page_id=695

I know he is planning to use the same electronics he has been using in the guided parachute flights.
 
Interesting project. I have an ambition to do something similar, but with forward-stowed wings that would swing out. I know a guy who cuts RC wing foam cores, and I'm working on some prototypes of RC aircraft electronics.

Are you thinking about doing anything to shift the CG back for glide mode? If you have the CG far enough forward for good rocket stability, you'll need an awful lot of elevon deflection (and associated drag) to keep the pitch controllable.

I started with a R&R Aircraft Delta rocket glider as a starting reference point.

As it is an excellent flier with a delta wing and no Cg or Cp shift when transitioning from rocket boost to recovery glide, I'm hoping to accomplish something similar.

For this project, a fast moderately steep recovery is desirable.
 
snip

As it is an excellent flier with a delta wing and no Cg or Cp shift when transitioning from rocket boost to recovery glide,

I have a gamma star (smaller version of the delta star) and while I have not flown it I know you are going to get a pretty big forward shift in Cg between the start of boost and burn out. No idea how this affects flight qualities yet, after I fly her I will give a report.

You guys might want to look at the JPADS program for inspiration!

https://faculty.nps.edu/oayakime/AD...ADS Programs Overview and NATO Activities.pdf
 
Last edited by a moderator:
Very cool build! What kind of electronics are you using? What kind of GPS module is the final build going to use?
__________________
Justin

I'm using the GlobalTop PA6B GPS. Here's a link:

https://www.gtop-tech.com/jsf/modul...puid=3fd4de25b0dd3bf98c210cb99566759047d54dec

I'm getting a blog posting ready for NerdFever.com with the latest status on the project. I had some good flights in July and learned a lot. It turns out my main problem has been tangling/twisting of the parachute control lines - I'm going to have to redesign the attachment of the payload bay to the parachute to fix that.

But that problem won't affect the rocket glider that Boris is building - the idea of the rocket glider is to test out the GPS steering system without the complications of the parachute deployment and tangling issues.

--Dave
 
Last edited:
Love the designs, I seen something like it before.

Go's up nice and glides like a cement brick, was very hard to control.
RIP!

Black_Wolf_Glider_rear.jpg

Black_Wolf_Glider_side.jpg
 
Last edited:
At the 8/13/11 CMASS launch in Acton, MA the small prototype glider flew twice on C11-3 motors.

Both flights went up fast and arrow straight.

I had intended to put the CG 1 caliber (in this case 1.0") in front of the CP. Due to the use of a lot of wadding, which pushed the parachutes and the CG slightly more forward, the actual CG was 1.25 caliber in front of the CP.

This model is 1/3.1 scale of the planned GPS guided rocket. It did not have any active control, this was just a test lift off of stability.

The small model's control surfaces were fixed in a flat position. Used "elevon" sizes that will scale to 2" wide in the 3.1x larger rocket.

IMG_5730.jpg
 
In addition to some other preparations for the arrival of hurricane Irene last week, I put my larger rocket projects on top of tables.

Normally they sit just off the floor of my basement, but the risk of loosing electrical power for days at the same time as heavy rainfall could have resulted in serious basement flooding.

It takes two sump pumps and a dehumidifier to keep my 24' x 26' basement dry.

These plus other rockets not shown here represent well over 500 hours of work.

Fortunately we did not have heavy rain here in southern Massachusetts and our house only lost power for one night.

IMG_5889.jpg
 
Last edited:
Built up the wings/fins with 1/4" x 6" sheets of balsa glued double wide. Bordered fins with 1/4" x 1/4" basswood strips.

Ordered 1/8" thick bulkhead and 38mm to 3" centering rings from LOC. Got thinner parts to reduce weight, their standard pieces would have been 1/4" thick. Top of motor tube has bulkhead bonded to CR.

Then used trig to figure out where along the wings to place 38mm and 54mm tubes to get desired 5 deg dihedral.

All bonding and fillets using Elmers wood glue.

IMG_5892.jpg

IMG_5895.jpg

IMG_5897.jpg

IMG_5904.jpg

IMG_5908.jpg
 
Something I have been wondering about.


Note - I have restored a number of things I had deleted from this message.

---------------
How will the GPS guidance avoid over-control? I could imagine this may have been addressed somewhat in the parachute version. But a storable parachute does not respond the same as a winged vehicle using elevons.

What I wonder about is basically this. When the model goes into a glide, it is not pointing anywhere near the landing area. So, the GPS realizes it needs to turn the say “left”. Does it give a LOT of left roll command to it? Because if it gives a lot, it may cause the model to roll. I do not mean bank a few degrees and do a left turn like an airliner, I mean roll left (or spiral-dive left).

That issue will be one of those things that is in one sense a mechanically “simple” thing to adjust, but may prove difficult to find the right ballpark.

I would suspect that it will need a lot less elevon roll control to do it than some might think it would need. I mean for example no more than 2 degrees of differential, possibly even no more than 1 degree if the model is built evenly (no warps, both wings the same mass), so that it would glide straight ahead with both elevons at the same angle (no differential).

If the system is smart enough, it could be programmed so that when there is less than 15 degrees of error, it will give less roll throw than when the error is over 15 degrees of error. And when there is less than 5 degrees of error, VERY fine tuned roll control. Otherwise, if it was “Bang left” and “bang right” type of turn control, then (assuming it had not gone out of control from over-control to begin with), from say a hard right turn once it finally pointed at the landing spot it would overshoot, make a “bang left” correction, quickly bank left, overshoot again, “bang right”, and so forth, zig-zagging. So, the turning control needs to be somewhat more proportional to the angle of error in that last 15 degrees or so. While also making sure that beyond 15 degrees of error, there is not an excessive amount of turn control used.

I will mention that the steering control might work better if it used rudder. That would help to solve some of the over-control problems that would occur with elevons used for ailerons. The system would still have to take into account how much error there was (in those last 15 degrees), but the response would not be so severe as long as the rudder was not commanded to extreme angles. When I first did my R/C Shuttle orbiter I used mixed elevons for aileron control, which was hard for me to fly. I later changed them to elevons that moved only for pitch, and added rudder control for turning, which made the flying a lot smoother.

>>>>>
3) at 100-200 ft altitude over the pad, elevons are raised to flare rocket and a conventional chute is deployed
<<<<<

OK, a couple of things come to mind here.

One, what if due to wind the glider cannot make it back to the pad? It seems this has no R/C override for the chute, so if it is 100% automated, what will it do when it is say at 100 feet altitude and 500 feet from the pad? Try to keep gliding to the pad, where it almost surely will hit the ground in a fast glide, well short of the pad (bad), or pop the chute (good).

It is possibly implied, be not stated for a fact, that the guidance system knows the model&#8217;s altitude second by second. So, does it know its altitude or not? If it does, then it should be programmed to pop the chute at no less than 100 feet regardless of where it is

If not, then this will need an R/C override for deploying the chute. Actually that even assumes the system works 100% correctly and the only problem is the wind. It would be best all around if there was an R/C override to pop the chute if anything else went wrong, like going out of control.

The other scenario that occurs to me, if the system DOES know the altitude, is what if the glider gets over the pad way early, at say 500 feet. Will it pop the chute then, or go into a &#8220;holding circle&#8221;? OF course possible issues with circles have already been brought up. And the model might not be able to complete a 360 circle when descending from 500 feet to 200 feet.

---------------------

One last thing. And this is a big reason why I did not reply to this thread a long time ago. I&#8217;m going to copy a phrase and then repeat it with one word changed:

&#8220;.... a fast glide back towards the pad&#8221;

&#8220;... a fast glide back towards the TARGET&#8221;

This is one of those global problems our hobby has to be concerned with. Not that you guys are planning to do anything bad, but that in developing something neat like this, the technology could be used by someone who did want to do something bad. This is certainly not the first time anyone has talked about using GPS to make a glider land at a certain spot, but then in the past it has been talk with no action.

What I am strongly urging is that the technical details of this project NOT be made public to anyone.

I will give you an example of something long ago on rec.models.rockets, where the project definitely crossed the line. But nobody else picked up on it before I did. Someone wanted to be able launch their rocket high up, arc over, and come down over a certain spot. He was going to have it go up thousands of feet, and fall back down most of that distance, where by R/C he would deploy the chute a few hundred feet up in the air (Take note that it would be coming in at hundreds of miles per hour nose first until that R/C deploy).

Well, maybe that does not set off any warning bells yet. But his system for getting the rocket to know where to point sure did. He wanted to be able to point a LASER at the &#8220;landing area&#8221; for the guidance system to steer towards. So, pretty much like laser-guided bombs use. Now, that was too much of a stretch. Because if he &#8220;forgot&#8221; to fire the onboard R/C deployment system (if it existed), the rocket would hit EXACTLY where the laser was pointed! Well, if the system had been perfected, anyway.

And perhaps he himself was not up to anything bad. But, there was too much risk of trying to help someone work up such a thing when that technology would be used be someone else as a near-perfect terrorist weapon system. Fortunately for once, the rest of the rec.models.rockets group members had sense enough to not have anything to do with it.


I&#8217;m not putting this in quite that same category. But it&#8217;s definitely in the gray zone (not by your intent, but by its potential misuse if perfected and technical details made public).

- George Gassaway
 
Last edited:
In addition to some other preparations for the arrival of hurricane Irene last week, I put my larger rocket projects on top of tables.

Normally they sit just off the floor of my basement, but the risk of loosing electrical power for days at the same time as heavy rainfall could have resulted in serious basement flooding.

It takes two sump pumps and a dehumidifier to keep my 24' x 26' basement dry.

These plus other rockets not shown here represent well over 500 hours of work.

Fortunately we did not have heavy rain here in southern Massachusetts and our house only lost power for one night.

Which manufacturer is you Saturn V from??? I really want one, LOVE the build so far! :wink:
 
[POW]Eagle159;232310 said:
Which manufacturer is you Saturn V from??? I really want one, LOVE the build so far! :wink:

Thanks. The Saturn V is a beautifully detailed kit from Sirius Rocketry. I highly recommend it, great customer service from Sirius too. After taking a core sample in its third flight, David Miller sold me the replacement parts needed to get it back in the air.

The kit is not cheap, but if you want a very finely detailed Saturn V that will fly on H and up power, I think this is the best available.

Sirius Saturn V kit: https://www.siriusrocketry.com/Saturn51.htm

My build and flights: https://www.bpasa.com/SaturnV.htm

Going up a J520 Skid and 4x G71 Redlines.

7aaa-1000.jpg
 
Something I have been wondering about.
...
- George Gassaway

Dave, who is handling the electronics, has just posted additional information on his website: https://nerdfever.com/?page_id=695

There is an altimeter onboard to fire an ejection charge and deploy the main chute. This is designed to occur whether the GPS control system is working well or not.

Our efforts at controlled glide replace the standard drogue phase of the flight. We will combat wind by building the glider to fly fast, so as to cut through the wind.

There is an RC override capability that we will use if necessary.

As far as concerns about the fact that there is an element of flight control, I will comment only once, as forum discussions that get into policy and politics get into food fights very quickly:

> Many RC planes of all sizes are widely available, offering much more proven control than what we are attempting here.
> Bad guys have already demonstrated that they prefer cars and trucks as they can reliably deliver large payloads.
> Nothing we do here poses any threat to anything. It is an attempt to make rocket recovery a shorter walk. Nothing more.
 
Epoxied Kevlar strap to motor mount. Strap will position rocket so that when it comes down under chute, the control surfaces will be away from the ground. Chute will still be deployed from the nose cone, and the Kevlar will be positioned on the outside of the rocket body during flight.

Epoxied (3" x 10-24) threaded rod next to motor tube for motor retention. Rod protrudes 1.25" from back of rocket.

Slotted body tube for fins and cut 5" long opening for electronics. Body tube is a full 34" section of LOC 3.1" tubing.

Epoxied motor mount/fins assembly into body tube. Then started on tiny epoxy fillets.

IMG_5915.jpg

IMG_5917.jpg
 
> Many RC planes of all sizes are widely available, offering much more proven control than what we are attempting here.
> Bad guys have already demonstrated that they prefer cars and trucks as they can reliably deliver large payloads.

That does not mean that putting a home-made GPS-based weapon delivery system on-line for public access is something that no terrorist would ever want to try to get their hands on. There are some scary scenarios where a rocket could be more useful for certain targets. Especially if the objective was not to kill many (or any) people, but to cause &#8220;terror&#8221; just by the fact that a guided rocket hit xxxxxxx, or yyyyyyyy, or zzzzzzzzz. And no, I&#8217;m not going to state specific scenarios public, but if you think I&#8217;m just BS-ing and promise to keep it totally confidential, that nobody else will ever see it, I&#8217;ll post you one or two in e-mail (send me your private e-mail as I won&#8217;t post it in a forum private message).

> Nothing we do here poses any threat to anything. It is an attempt to make rocket recovery a shorter walk. Nothing more.
Never claimed the intent of what you guys are doing is anything more than what you have said.

But there is quite a potential for someone else to take it from there.

And yes I know how discussion of this potential can get messy, but that does not mean it is a smart thing to just ignore it and pretend it&#8217;s not a problem.

Edit - The further explanations and safeguards Dave has mentioned in this regard has made me change my mind on this project. I have sent him a private e-mail discussing this, some other possible safeguards, and the prototype glider steering control issues.

- George Gassaway
 
Last edited:
That does not mean that putting a home-made GPS-based weapon delivery system on-line for public access is something that no terrorist would ever want to try to get their hands on.

Hi George,

I don't often post here, but I want you to know that I share your concerns, and have been taking steps to do this project in what I think is a responsible manner.

All of us involved in any hobby (rockets, R/C airplanes, model boats, autonomous vehicles, etc.) that has a potential for misuse by "bad guys" have an obligation to think about these issues.

Back in December 2010 I posted the source code for the program that runs on the microcontroller. I thought hard before posting it, and this is what I ended up with (the following is a cut-and-paste from https://nerdfever.com/?p=1116, posted 12/3/2010):

The code I’m posting today, like the earlier “Rev3&#8243; code, does not include the navigation code. But it does include everything else - logging altimeter, parachute deployment, GPS, servo control, etc. Because of the “abuse potential” of the nav code (think of navigating things to places where they ought not to be), I don’t intend to make the nav source code public. Once it’s working well, I might be interested in working with reputable vendors to sell hardware that includes this function, but if I do, it’ll have some protections in place against abuse. The main protection I’m considering is to limit the target location the system will aim for – it has to be a place the unit has physically been since it was powered up (you won’t be able to program in some other location). That way, if you’re not allowed to go somewhere with a handful of blinking rocket electronics, you can’t land the rocket there either. I’d like feedback on this idea. (Yes, anything can be hacked if you put enough effort into it, but my goal is to make it harder to hack the system than for “bad guys” to build their own – there are, after all, books on the subject…)

I hope this makes you feel at least a little more comfortable with what Boris and I are doing. FWIW, nobody other than myself has access to the navigation code (not even Boris). I have no intention of making that public in source code form or without the kind of protections I described above.

If you can think of serious scenarios enabled by this project that wouldn't be prevented by that kind of protection, I would be very grateful if you emailed me (privately - dave "at" nerdfever.com) about them, so I can think about how to thwart them.

Nothing is 100% secure. Bad guys can go to college and learn how to build things. They can also read books on the subject or attach bombs to R/C airplanes bought at a hobby store. All these threats are real, but they are not a result of what we're doing.

I'm trying to do the right thing here.

Best regards,

--Dave
 
For those following the GPS-guided rocket recovery project, I've just finished a bunch of new posts catching up my blog on the project:

https://nerdfever.com

Cheers,

--Dave
 
I have enjoyed this thread... Nerdfever is a cool site... keep it up...
this type of build has been on my bucket list for a long time, but is above my - ability.
I'll get there one day..

Its not like your Bruce Simpson.. Don't let the techno-illiterate thwart or say your enabling terror. If a university can post DIY drone sourcecode, no one says a thing.. A hobiest does it, and the other hobbiest jump bones...
 
Back
Top