Last week CanberraUAV gave a presentation and flight demonstration to the Canberra branch of the Royal Aeronautical Society (RAes).

This covered the history of CanberraUAV in competing in the UAV Outback Challenges in 2012 and 2014, the group's purpose and past/current R&D activities. This also included an introduction to the world of open-source UAV development by Ardupilot developer (and CanberraUAV member) Andrew Tridgell.

At the end there was a flight demo of a quadplane, showing off it's autonomous flight modes.

Views: 669


Developer
Comment by John Arne Birkeland on March 12, 2016 at 11:26am

Always nice and interesting to hear Tridge speak.

Comment by John Dennings on March 12, 2016 at 2:05pm

Thanks for sharing. (I am such a sucker for those).

Incredible accomplishments ...

Comment by JB on March 12, 2016 at 10:47pm

Nice presentation.

I was wondering with the quadplane demo where it misses it's landing WP do you know why this happened? Was it caused by the wind? Also doesn't it normally hold position before decent? Thx.


Developer
Comment by Andrew Tridgell on March 13, 2016 at 1:17am

@JB, yes, that was actually our first test of a quadplane with significant side-wind on landing. We've tested this in the simulator before and it correctly weaher-vanes around so that the nose is into the wind. In this case for a real flight it didn't, and it didn't manage to hold position against the side-wind. I suspect the yaw rate controller was just too agressive.

Wind on landing is a real weak point of quadplanes. I think if we get the tuning right then we can handle it, but it certainly didn't work right "out of the box" for this aircraft.

Comment by JB on March 13, 2016 at 2:17am

Thanks Tridge.

I could see that there was a fair bit of wind with the normal airplane approach. It looks like that it cannot maintain heading into the wind as soon as it switches to hover. Is the yaw control inducing roll as well or is it only trying to maintain level flight? It rocks about a bit. It's pitching up a fair bit too in the video. To avoid it becoming unstable maybe the roll/pitch angles need to be limited in hover, as high angles make the airfoil interference worse?

As previously the wind feathering seems to be fairly important to maintain position. I'm wondering if there's another way we can get accurate wind headings, even when stationary, so that this can be fed into commands to make pos. hold accurate, let alone not make the quad components work overtime to do so as the affect of wind on the airfoils are not taken in consideration when in quad mode.

Currently once in quad hover mode I believe the pusher motor is switched off completely. Maybe a solution would be to use the pusher prop to maintain position against the wind instead, and let the quad only deal with level attitude and altitude control instead. Is it possible to use previously acquired wind heading from forwards flight to give a heading range in which the PXH tries to maintain direction? At least that way it won't turn out of the wind, which in turn exposes more wing area to wind, unless of course the wind direction changes suddenly. It seems to me that we need to find a solution to give reliable wind heading before a accurate control is possible.

Possibly a good test would be to try to land the quadplane with pusher motor on, into the wind, at wind velocities over airfoil stall in simulation. In simulation does it model both the airplane airfoils and quadplane aerodynamics? 

Comment

You need to be a member of DIY Drones to add comments!

Join DIY Drones

© 2017   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service