Replies to This Discussion

Permalink Reply by bGatti on September 22, 2008 at 7:37am

Quads are controlled about their plane by changing power to each blade, and around their axis by moving power between the pair of counter-rotating blades.

You'll need an gyroscope for each (3) dimension, and a cpu to process the information.

It would seem that some studying would be in order for what you are "designing".

Good luck

• ▶ Reply

Permalink Reply by rochak on September 22, 2008 at 7:48am

thanks
what i had thought was
i wud use an IMU to record roll pitch and yaw along 3 axis
accn along 3 axis
use some sort of transfer function programmed on the cpu to convert the error b/w desired roll pitch yaw and accns along 3 axis to change in rpm of each motor

the transfer function requires an elaborate mathematical modeling and understanding of modern control theory
dats where i am stuck up!!

• ▶ Reply

Permalink Reply by Stefano on September 25, 2008 at 12:20am

Some times ago I found this on the net...

It's in French, BUT math is international language and all equations are developed here.

Hopefully this helps

• ▶ Reply

Permalink Reply by Peter Seddon on March 2, 2009 at 10:47am

I would suggest a visit to www.mikrokopter.de where you will find all the info on hardware and open source software you require.
regards Peter

• ▶ Reply

Permalink Reply by Noland on April 1, 2011 at 12:05pm

Hi, forgive me if I'm blind but I'm having trouble finding the source code on the mikrokopter website. I can find the hex files and change logs but not the source. Could you point me in the right direction please?

• ▶ Reply

Permalink Reply by Squalish on April 1, 2011 at 1:29pm

Mikrokopter's codebase is not OSS, it is just (mostly) publicly available code, subject to restrictions.  The navigation code is completely closed-source.

 

http://svn.mikrokopter.de/

 

 

• ▶ Reply

Permalink Reply by Noland on April 1, 2011 at 1:33pm

I see. Thanks.

 

guess I'll have to write something myself.  Anyone know of any good tutorials on control system theory for copters?

• ▶ Reply

Permalink Reply by Squalish on April 1, 2011 at 1:47pm

That's just Mikrokopter.  There are a dozen others.  This site is centered on Arducopter, an Arduino-based autopilot.  The current version, Arducopter NG, is going to be replaced by Arducopter Mega in a few weeks, so we can use the same hardware & code as the Ardupilot system.

Read up a good bit on the forums and blogs here, so you can get to the really productive questions, and you can probably be building by the time the ACM 1.0 is released.

• ▶ Reply

Permalink Reply by Roy Brewer on April 2, 2011 at 6:13am

May I humbly offer my Quadrotor Physics and Control Theory thread:

 

http://www.rcgroups.com/forums/showthread.php?t=1284741

 

- Roy

• ▶ Reply

Permalink Reply by Noland on April 2, 2011 at 3:47pm

Wow. I read the first paragraph and I'm expecting great things now. Very exciting, thank you!! looks like I have lots of reading material now.

But I'm impatient, can you tell me quick how the story ends? Did you end up getting a decent control algo developed? Is there hope for an electronics engineer with no control background?

• ▶ Reply

Permalink Reply by Roy Brewer on April 2, 2011 at 5:15pm

I will let you know how the story ends when I get there!

 

There's a lot of folks with neither an engineering nor a controls background who are making good progress with quads. So there's more than just hope :-)

• ▶ Reply

Permalink Reply by quix-fz on March 2, 2009 at 12:16pm

hey
i recently got mine flying, it's very stable, check out my page.

i would recommend, that you first get your imu running, at a fast enough rate. at the minimum least 200Hz (better 500hz or more). that means, this often read out the imu-data and integrate the gyro-angles, experiment with mixing in accelerator data. try out filters on the data and look what they do.

you'll also need to interface with brushless controllers. so find some that support such a high update rate (suppersimples from hobbycity are great and cheap, or build the ones from mikrokopter.de). for the beginning it's enough when you buy/build just one, to get it running and see how it works. in case you give up that quad-rotor idea, you haven't thrown away too much money.

also, find i way to interface with a rc-receiver. pulseIn in arduino wont work (well), because it waits for the ppm pulses from the receiver, and this will slow down your imu computation and stabilisation. you'll need to use some kind of input capturing of your mcu. you can also browse rcgroups.com and the mikrokopter.de forum, they should also contain lots of info on that part (also for 2.4ghz receivers)

once you have all that running, build a frame and buy the rest of the esc and motors. the actual stabilisation is (relatively) very easy.

• ▶ Reply