Its been a while so lets try again. Had a hangout with Chris and we briefly spoke about the next T3 round. Chris will work on a top prize ;-)

The mission is simple, get airborne climb to 20m 

Fly a cube with 50m sides, pausing for one minute at each corner, so your flight time cannot be anything less than 8 minutes..... Bonus points if you can stay longer at each corner....

The neatest cube KML wins.

If you can do this with a 3D aircraft I feel you would have a very strong chance of winning!

I will close the competition on April the 14th. I will be looking at where you are flying very closely. Please don't try this in public places or within 500m of any building / road. 

Views: 13578


T3
Comment by Steve Westerfeld on April 12, 2013 at 5:24pm

The AQ uses the MP3H6115A pressure sensor with a little open cell foam covering it and out of direct sunlight.

 The code is all open source except for a couple of programs that are provided that does the curve fitting for the calibration.  

The hardware is closed but everything hardware wise is openly discussed it's just the schematics aren't available to help prevent cloning.

I'm disappointed that more AQ flyers didn't post but I know of one plot that is excellent and he worked pretty hard to do it sitting on a snow mobile.  I hope it will show up soon.

Comment by Rob_Lefebvre on April 12, 2013 at 5:29pm

If I understood the thread at the AQ forum correctly, the trace shown is using the reported baro reading, not the GPS altitude.  And therefore the track is what the system believed it's height to be, not what it really was? 


Developer
Comment by Randy on April 12, 2013 at 5:33pm

Steve, oops - I jumped to the conclusion that it was the bmp when i saw the single hole in the top.


Developer
Comment by Randy on April 12, 2013 at 6:38pm

I've posted a note over on the AeroQuad forum, I think it would be a lot of fun if some AeroQuad guys were able to make it to the Sparkfun AVC.  Some of the most experienced DIYDrones developers will be there and we will be giving it our best shot.

Comment by Bill Nesbitt on April 13, 2013 at 4:21am

R_Lefebvre:  The altitude the GPS reports is what the GPS module thinks the altitude is, not the actual altitude. Just as the pressure sensor derived altitude is not the true altitude.

Unless / until you have a trusted 3rd party observation of the true altitude, any system has to take this sometimes conflicting information, fuse it with all other available sensor information (like inertial) and come up with its own estimates.

With AQ, you have the choice of following the GPS or pressure dictated absolute altitude for navigation.  Most fly withe the latter as it is smoother as it does not jump around as much.  Steve's flight was using this standard as well.

If you plot the estimated flight altitude against the pure GPS reported altitude, you will see slight deviations between the two.  The truth is that there is no way to know which is more correct or true unless you have an outside observer.  In practice, we find that the pressure derived altitude tends to be more correct more often.  I would maintain that the deviations seen between the two are due to the GPS's altitude fluctuation and in fact the craft was flying a better estimate of altitude.

If Steve had chosen to fly the cube with following the GPS altitude, you'd see an even more consistent, straight horizontal lines if you plotted the GPS altitude as the machine would be targeting exactly what it is being judged by.  AQ does not have any control problem keeping an altitude it is told to.  The problem (as it should be) is estimating what the "true" altitude actually is.

A way to get the required 3rd party observation to truly judge such competitions is to collect raw GPS data that is provided by some GPS modules, like our LEA-6T.  You can then use it in connection with ground station collected data to post process using RTK methods to get centimeter or better accuracy as to what exactly happened during the flight.

Comment by Bill Nesbitt on April 13, 2013 at 4:25am

Randy: AutoQuad, not AeroQuad.  Two different beasts.

Comment by Rob_Lefebvre on April 13, 2013 at 6:08am

Bill, OK, then I misunderstood that part.  I think maybe somebody at least asked the question about baro Alt recording?  That's what made me think that.

Anyway, doesn't matter that much as the AQ system clearly outperformed ours.  Congrats on that.

I don't think there is any need to get into 3rd party observation.  This is supposed to be just for fun.  Also, the results you posted is probably far and away a higher performance than any real-world mission requirements.  Even the Arducopter results, if this was really a building inspection, would have been quite good enough for the mission.

I really wish we could also get some professional systems to compare to.  Like Draganflyer, or AV Scout, etc.  See how they stack up.  I've seen very little video evidence of their performance.  I did see one from the Scout, and it was highly unimpressive.

Comment by Bill Nesbitt on April 13, 2013 at 6:40am

I own a Dranganfly.  Its performance is below everything I've seen here.

I agree that such observation is not needed in this case if all of the contestants have shown themselves.  But what happens if there are other AQ entries, or some other evenly matched platform?

I disagree that what we have is "good enough."  Maybe good enough if you wanted a building inspection done from 50 meters away :)  There is always room for improvement.  For me, that is the fun part.

Comment by Rob_Lefebvre on April 13, 2013 at 7:21am

Bill, I honestly have no idea how Gary is going to be able to pick which AQ entry is the winner.  Will need to use Matlab or something. ;)  Really, it's almost pointless. ;)  I guess a couple AQ guys will get a shiny new APM as a prize! :D

The only thing I am considering doing at this point, is a long duration mission with my heli, but *with a payload*.  Something the quads aren't doing as all have been pretty much stripped down flying batteries.

But I'm just running out of time.  Probably do first test hops today if weather permits.

There's no doubt that I'm going to run this mission in the heli, even if it's after the contest is closed, just to show it can be done.

Arducopter IS unique in that regard, other than DJI Ace One Waypoint.  And *maybe* OpenPilot?  Am I missing anybody else?


T3
Comment by Aerhead on April 13, 2013 at 9:08am

Hi

I have an Autoquad Cube challenge to post.  I'm a farmer and as soon I saw the first quad I knew that I wanted to use that type of aircraft as a tool for farming.  For the last seven years I've built multicopters using many of the opensource autopilots. The Autoquad is the first autopilot that is getting close to the precision that I'll need to do the tasks I have in mind for multicopters.

Everything to do with multicopters has been a challenge for me because of little background in computers or electronics.  Without the help of the people on many forums there's no way I could have been able to do this challenge.  Steve (Kisssys) has help me with my octo  many times. 

Here's a link to the challenge video of the AQ Octo.

A zip file containing the KML file is available here.

Larry

Comment

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

Join DIY Drones

Groups

Season Two of the Trust Time Trial (T3) Contest 
A list of all T3 contests is here. The current round, the Vertical Horizontal one, is here

© 2019   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service