At the recent LCA'2014 conference in Perth I gave a couple of talks about research projects I'm working on with ardupilot. These show a couple of the things that we are working towards with APM.

The first is a talk on differential GPS which is part of a collaboration between Ben Nizette at ANU and myself.



The second talk is about the research project to run ardupilot directly on embedded Linux boards, which opens up some really interesting possibilities for APM:



These projects are really just a small part of what we are working on, but I thought a few people might be interested in them.

Cheers, Tridge



Views: 13231

Comment by Andrew Tridgell on February 3, 2014 at 2:36pm

the gumstix line of boards are indeed very nice, and small! However for an initial port with the same set of autopilot features as the Pixhawk the BeagleBoneBlack with a cape seems an ideal target. I suspect that once we have that flying that people will port ardupilot to other boards (I have it running on a RaspberryPi for example).

One of the things we plan on doing with this Linux port is auto-detecting sensors where possible, and making it easier for people to control the startup to use the right UARTs and other devices for their boards.

There are heaps of great embedded Linux boards available, and I'm sure that if the BBB port is a success that others will follow.

Cheers, Tridge

Comment by Dan Wilson on February 4, 2014 at 3:28am

Awesome work Tridge, particularly on the DGPS, I'm really looking forward to seeing the result! I've also been playing around with a BBB based autopilot - you can see a brief glimpse of the cape I'm using in this video, you may find the Simulink part interesting too. As you highlighted, one of the challenges one faces with a Linux based autopilot is the hard real-time requirements necessary for hardware interfacing. I ended up putting a microcontroller on the cape to facilitate this.

Comment by Adam Kroll on February 4, 2014 at 7:29am

Hi Tridge,

You mention at the 40 minute mark of the dgps talk that the geoscience australia gps receiver is at 25m under sea level.  This is because GPS height uses an ellipsoid reference model of the earth, not mean sea level, due to problems with using the mean sea level as a reference.  This page explains it well

So if in the future we do manage to develop an autopilot that has good gps accuracy for autolanding, don't use absolute gps heights, its best to use a relative height from your take off position.

Comment by Adam Kroll on February 4, 2014 at 7:35am

Forgot to say that you're a genius Tridge, how you get so much done in so little time amazes me.

Comment by Muhammad Al-Rawi on February 4, 2014 at 3:18pm

Brilliant presentations, Tridge. Very informative and captivating. 

Comment by David Pawlak on February 6, 2014 at 1:49pm

Bought a BeagleBone right away. Should be arriving tomorrow.

Exciting! Just waiting for the cape. Even feel as though I could spit into the wind!!

Comment by David Pawlak on February 6, 2014 at 1:50pm


Comment by Julian Josephs on February 6, 2014 at 11:16pm

I look forwarding to more information about the beaglebone fire cape.

Comment by Andrew Tridgell on February 7, 2014 at 12:07am


yep, we worked out the problem was the datum later. The difference between WGS-84 and the datum that GA uses is quite small in Canberra, whereas it is quite large in Perth. This was our first test in perth so we got fooled.

Now we need to implement the geoid that GA uses in python so we can apply the correction in our code :-)

Cheers, Tridge

Comment by David Pawlak on February 7, 2014 at 3:52am

Tridge, do you use the Debian 7 (small flash) for this.  What about Ubuntu 13-10, too big?


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

Join DIY Drones

© 2020   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service