ArduPilot, PX4 dominate AUVSI drone competition

When we got started ten years ago, the annual AUVSI student drone competition was dominated by commercial autopilots, such as Piccolo. Now it's almost entirely open source autopilots, led by ArduPilot (14 of top 20) and Dronecode/PX4 (3 of top twenty). I'm super proud of this having co-founded ArduPilot and now leading Dronecode. Only one commercial autopilot in top twenty -- next year they will be gone entirely!

From sUAS News

Views: 911

Comment by Rana on June 26, 2017 at 10:23pm

Its really great to hear all this !

Comment by Andy Fabian on June 27, 2017 at 5:52pm

I'm on Virginia Commonwealth University's team.  We run our own flight control system, both hardware and software.   Chris, could you please fix this on your posting?  We're quite proud of our FCS.  (Nothing against ArduPilot either, it's great and I've run it for years personally.)


Developer
Comment by Tom Pittenger on June 28, 2017 at 12:00am

A few of the universities have contacted Gary about their autopilot status. Seems Bringham Young is using ROSflight instead of PX4 too. However, they are using ArduPilot for their estimator system because ROSflight appears to still use DCM.


3D Robotics
Comment by Chris Anderson on June 28, 2017 at 2:37am

Gary updated his post with those corrections, so I've incorporated them above


Developer
Comment by Tom Pittenger on June 28, 2017 at 3:45am

@Andy Opps, I mixed things up there. I was actually referring to VCU's Aries regarding using the estimator. VCU is listed as using your own autopilot but your paper says you're using a pixhawk hardware with ArduCopter's sensor code and EKF. Can you comment on that?

Comment by Andy Fabian on June 28, 2017 at 6:26am

Sure.  I think you pretty much described it though - we've swapped out actual GPS and Magnetometer signals for the equivalent data coming from the Pixhawk, which benefits from their inertial navigation and Kalman filter algorithms.  We did this by modifying the ArduPilot source to output a special data stream at 50 hz over one of the serial ports, and then modifying our autopilot to receive and use that data.  This is a stopgap as we work on adding an in-house Kalman filter to our autopilot, but (as you might have noticed), those aren't exactly trivial to develop and test.


Developer
Comment by Tom Pittenger on June 28, 2017 at 8:47am

@Andy I see that the suasnews website has updated your column

Comment by Andy Fabian on June 28, 2017 at 10:34am

@Tim Thanks for the heads-up.  I've just sent the following message to Gary of SUAS news:

Gary,

I see that our listing again includes ArduPilot.  Can you please clarify in the listing that our Flight Controller is VCU Aries, and our position sensor is ArduPilot?  A good title might be "VCU Aries (using ArduPilot for position data)".  Although, since none of the other listings go into this level of detail (there's no "ArduPilot/uBlox", for example), I think that leaving our autopilot listed as simply "VCU Aries" is most consistent.  
What I want to avoid is giving the impression that Aries is just a variant of ArduPilot, which is significantly downplaying all of the work that myself and others have done to create an autopilot from scratch.

Developer
Comment by Tom Pittenger on June 28, 2017 at 4:20pm

@Andy,

I was the one who contacted Gary regarding the autopilot title showing both. It felt more accurate because it was an autopilot you used. You used two. uBlox is not an autopilot, the column lists autopilots with the exception of "Pixhawk mini" which should say PX4.

Comment by Andy Fabian on June 28, 2017 at 7:36pm

@Tom,

We did not use two autopilots, by any rational definition of "autopilot".  We used Aries to fly the airplane, and Arducopter to provide position data.  I think that's pretty clear from our technical paper and from my previous posts.  I assume your definition of autopilot would involve piloting an aircraft, right?  If not, then I think we can just stop here, but if so let me clarify how incapable our pixhawk was of piloting the airplane:

1. It was not attached to any of the servos or motors

2. It was not linked to our GCS

3. If powered off, our aircraft would continue to fly with full stability and remote control.

4. It was running ArduCopter, not ArduPlane.

ArduCopter is generally used as an autopilot.  I'm not disputing that because it's obvious and I'm not trying to diminish any of the brilliant work that's gone into it.  But we used it as a position sensor.

Our school developed, from scratch, a full hardware and software autopilot, took it to the AUVSI competition, and performed well with it.  We're absolutely appreciative of the quality of ArduCopter's position estimation algorithms, and chose to use them as a position sensor, because they're better than raw hardware, and cheaper than other similar alternatives.  We mention this is our technical paper, our flight-readiness video, and in a poster we bring the competition - it's not like we're trying to hide it.  

But ArduCopter didn't fly the plane, Aries did.  In a table where the columns are "Team" and "Autopilot", the only correct values for our team are "VCU" and "Aries".

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

© 2017   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service