Hi,
I found several developers who are working on a Quadrocopter (or tried to build one). I’m not sure if there is still anybody working on it?!?
Are there any results? Is one flying?
Regards,
Michael
the big problem with doing a quad copter based on NETMF is that NETMF is not a realtime OS, which means you can’t really guarantee that it will execute an instruction when you tell it it (it might be a millisecond or two later). Eventually, this kind of thing would cause noise to creep into your stability controls.
That said, you can still do it, but you might need to have a chip running native code flying the aircraft, taking input from a NETMF board.
Of course, it could happen that an interrupt will be handled a millisecond later. Do you have done any tests?
No, but what I will say is that I have been using my FEZ as the driver for a high speed (~30MPH) robot and I have yet to see timing become an issue. Although the FEZ drives the car, I’m controlling it, so it’s not like I have run control loops on it or anything, though.
NETMF is not real time OS but you have control over everything so you can get timing to be accurate enough for most applications. You will also be using hardware controlled PWM so that is always working no matter what your system is doing.
Well, I’m searching for anybody who is interested in such a project to see if it is really possible to fly a Quadrocopter with .NET MF. For my part I’m missing all the electronic knowledge… but could offer a MSDN subscription if there is anybody interested. 
I have two thoughts:
[ol]My brain/nervous system is not a hard real time system either but I suspect I could manually fly one of those little quadrocopters (with some practice) even though my reaction time is widely variable and would be on the order of 200ms.
Most of the small helicopters and quadrocopters I have seen can barely lift their own weight let alone a payload of electronics. As you progress to larger machines with larger motors and battery packs you have a bit of spare capacity. [/ol]
I bought one of the AR.Drone from parrot.com. It is running a Linux and is controlled via WiFi from an iPhone.
http://ardrone.parrot.com/parrot-ar-drone/de/technologies
I’m currently only thinking of a GPS enabled Quadrocopter which should get a request to fly to x/y/z position.
And yes, one of the end result should be that it can have a payload to take a photo or video.
The AR.Drone is using ultrasound and the bottom camera to hold its position if there is no new command.
But the first test would be to see if a .NET MF device is able to response to the sensors fast enough.
Making minor course corrections to maintain a stable flight will be the real test. I say this actually having built a quad before. The Rhino i own will be used as for one but i have competition robots to build first before i goto work on the quad.
This is a very interesting project and I am sure GHI will support whoever makes it.
@ bstag: perfect!!
Maybe there are others that did some research already.
@ Jeff, the toy quads can’t lift, but the bigger ones certainly can. See: MikroKopter - HexaKopter on Vimeo
Also, you can control a quadcopter because you are a human. You can learn how the thing is going to react. Obviously, a computer will have a much harder time with it.
@ Chris: mikrokopter.de, this is a really nice one, I had a look on their Web page several times, but 1.500 EUR for it plus some extras like GPS… 
The AR.Drone can lift 100-150gr, which is enough for a HD camera.
Yeah, professional quadcopters like that are very expensive. They are also highly stable in flight and can lift a large payload 
But, for cheaper, see: ArduIMU quadcopter part III - Blogs - diydrones
The one thing I see in common with all these projects is that the drone is remotely controlled. Does anyone have any info about autonomous drones?
Most of those quadcopters are autonomous, in that they have heading hold, altitude hold and location hold autopilots. The Arduino based quad I linked above is also autonomous, it’ll run until something comes close to it, then it will go the opposite direction.
Nah, I mean autonomous. as in it does some task without a remote control. Like how industrial robots do some task without a remote control. Or the RC car navigating to different way points on it’s own.
Did you see the quad I linked? It just flies around until something gets in it’s way. Technically, that [italic]is [/italic] autonomous behavior. Nobody’s telling it where to fly.
I think if it is working with .NET MF (analyze sensor data, control motors) it doesn’t matter if you control your quadrocopter at your own or react as autonomous process.
Does anybody have some free time to start with first small tests? Where could we start?
Michael
A flying bot is next on my list (ie, 2011) after the driving bot is done.