Andrew Tridgell ("Tridge") is a well-known programmer in the open source world (Samba, etc) and one of the leaders of the APM dev team. His UAV team, CanberraUAV, will be competing in this year's Outback Challenge with an APM-powered Telemaster, and at the Australian Linux conference he gave a spectacular lecture about the team's strategy and the technological challenges in the competition.
This is a must watch for anyone interested in more advanced UAV functions. Tridge and his team have added a Panda board to APM to do image processing. This is a great opportunity to watch a world-class technologist explain the strategy they're using to try to win this competition.
Comments
Cool talk, thanks for that.
Question regarding the geofencing crashes you mentioned: what sort of conditions caused it to fail and what can be done to improve that?
You're right, my mistake sorry.
I watched the video and thoroughly enjoyed how the implementation of APM in such a contest can ponder up so many new ideas. I wonder about the dialogue, and in addition the APM 3.0 road-map that mentioned moving to an "ARM" based system. Can anyone elaborate on that?
Best of luck Tridge! You better get some great video from the competition!
@ Chris Gough
Hey Chris this quote " Also, very ambitious to attempt complete automation the first year out, can you highlight some reasons for this approach as opposed to mastering the minimum completion requirements first?"
Was posted by David on page 1, not by me.
As a fellow CanberraUAV member I can answer a few of these questions (for myself, not for Tridge):
@Brian Adrian
It's a competition, competitions are for winning... I don't understand your question ;-)
Yes it is ambitious but then we are a determined and organised team. Minimum completion requirements are ambitious too, especially for a small team with just a couple of people (we are big enough that people get to specialise).
@UAVOZ
tuning is just a minor detail. If we had crashed/broken The Big Plane that many times we would have run out of time and money already. We try to find simple tests for each concept, then keep validating and improving the sucessfull elements as they are integrated into increasingly sophistocated combinations.
If we tried to work out all the details up front using theory, we would still be debugging on the bench and not flying. That would not be as much fun. We have many flights already, and learned many lessons from them, but have only flown The Big Plane a few times, and even then only with conventional RC (autopilots as passengers).
@Chris Anderson
A minor correction - the telemaster is probably not our competition airframe (possible fallback though, I suppose). It's just a very handy test workhorse because it's big enough to carry the long range radio gear and vision stuff at the same time (just). We still haven't validated the datalink or vision system next to the thumping great petrol motor! safe to assume that will throw up some issues, like everything we'll find them and fix them one step at a time (even if we have to sacrifice a washing machine).
@ Andrew, I wonder why your team use many platform to test the Autopilot? So you need to tune the autopilot for different airframe every time you change them?
@Squalish, we did consider using high-angle photos, but decided that our current strategy of covering the whole area from 100m was simpler and much easier to implement with fast computer vision algorithms.
The altitude code in APM already combines the barometer with the GPS home location, with the barometer calibrated at startup. For the sorts of altitudes we fly at the local gradient corrections would be very small I think. Mixing GPS in during flight is problematic as GPS altitude varies so much - it is quite common in a 15 minute flight for the GPS to say that the airstrip is 30m higher on landing than it was on takeoff, whereas we tend to get barometer based altitude within a few meters.
Thanks Andrew! And thank you for doing the presentation and keeping open source alive.
@Brian, only one team flies at a time. Each team gets 1 hour plus some prep time. The competitions runs over multiple days. It is also quite likely that some teams will not make it past the D2 and later deliverables, if last year is anything to go by then we'll probably be down to a dozen or so teams by the day the competition starts.
It is open to everyone, and there are 6 US teams at the moment. See http://www.uavoutbackchallenge.com.au/2011/index.cfm?contentID=30 for a list of the teams.
Thanks for posting this! It’s the first time I’ve heard about it, I think Joe maybe a bag of bones by now. I do have a few questions, he mentioned that 53 teams are in the competition, do they all start at the same time? Seems like there be a problem with air traffic control with 53 UAV flying. What a great competition, is it open to anyone and everyone? And one last question, anyone from the US entered? Anyone interested from the US?
Again I’d like to say what a great competition! This is how people make a change in the world, all working together with no hidden objectives. Just having fun and collaborating to achieve a common goal.
-
1
-
2
-
3
of 3 Next