UAV finds (and nearly saves) Outback Joe
In what is believed to be a world-first for a non-military drone, an Unmanned Airborne Vehicle (UAV) was able to locate a dummy of a missing bushwalker without human intervention during the A$50,000 Outback Challenge in Kingaroy, Queensland.
15 October 2012
Glen Paul: G’day, and welcome to CSIROpod, I’m Glen Paul. The Annual Outback Rescue Challenge was recently held, and again placed developers of unmanned airborne vehicles, or UAV’s, into competition to find a missing bushwalker.
As in previous years, the outback is near Kingaroy Airport in Queensland, and the missing bushwalker is a mannequin, but the unmanned airborne vehicles are very real, and becoming more sophisticated with each competition. This year that led to a world first for a non-military drone.
Joining me on the line to discuss the Outback Rescue Challenge is its co-founder, CSIRO’s Doctor Jonathan Roberts. Jonathan was there a clearly defined winner, and what was this world first?
Dr Roberts: Yeah, there was – there was. We had four teams that passed the final safety checks and actually took off and went to go and find the mannequin, who we call ‘Outback Joe’, and there was a clear first place team, but as in previous years, unfortunately no team actually got the grand prize. So for this competition we have $50,000 on offer for the team that can launch their aircraft into the search arena, find Outback Joe, drop a water bottle to Outback Joe, and recover their aircraft back at the airport.
This year, though, we did have a first. We had a team called Canberra UAV, who are a group of amateurs, and hobbyists, and open source software developers, they had a very impressive setup, and they actually managed to automatically find Outback Joe, so they launched their aircraft into the search area, searched for about 35, 40 minutes, and then the aircraft amazingly reported back that it had found Outback Joe, and some little pictures of Outback Joe appeared on the screen at the ground station, it was all very exciting, and they’d found him.
Unfortunately, and the reason they didn’t win the $50,000 was because sort of about ten minutes into the flight their water bottle actually detached from the aircraft in an unintended time, and in an uncontrolled manner, so they didn’t actually have a water bottle to drop, and then when it came off the aircraft it slightly damaged the aircraft as well, so there were complicating factors. So they were close, but it was not to be this year.
Glen Paul: Hmm, but at least there was still this world’s first of where the aircraft was able to find him without human involvement. How did the plane itself achieve that?
Dr Roberts: So, the aircraft has got onboard a camera, colour camera, looking down. We tell the teams beforehand, and this is detailed in the rules, that roughly what the bushwalker will be wearing, so they actually use that information to help find them. So the bushwalker wears blue jeans, sort of brown work boots, and one of those highly reflective worker’s shirts, and we use a bright yellow one.
And so they use that information then to program a little computer onboard the aircraft, where the video stream is going from the camera that’s looking down. And the team that won, they actually were looking for blue things, so they were looking for the jeans, so as they were flying around, and they’d done some analysis, and they determined that blue things aren’t that common in the natural environment, so they had a little algorithm onboard looking for blue things. They know roughly how big Outback Joe is, because it’s a human mannequin, they know how high they’re flying, so they can figure out roughly how big an object they’re looking for in the scene, so they look for blue things of a certain size, and all the candidates, where they find a blue thing, will then look for a yellow or an orange thing right next to it, which is the shirt. So they find the jeans, they then look for the shirt to match. And that’s how they did it, and it worked incredibly well.
Glen Paul: Hmm. Amazing. And what about the status then of Outback Joe, he didn’t get the water, so did he perish under the guidelines?
Dr Roberts: Well, that’s right, so he’s still out there, so this is… this was the fifth year we’ve run this competition out at Kingaroy, which is in south-east Queensland, and unfortunately according to his Twitter feed, because he is on Twitter, he does have good 3G coverage out there even though he’s lost, he has been reporting that he’s still thirsty and would like to be rescued next time.
Glen Paul: Fair enough. We’re able to avoid the memorial service. (Chuckles).
Dr Roberts: That’s right (chuckles).
Glen Paul: Let’s go back then to the very first contest. How sophisticated are the UAV’s becoming now? Would a top UAV from back then still be competitive today without an upgrade?
Dr Roberts: No, no. Back in the first two years of this competition not one team managed to launch their aircraft from the airport. During the year we have numerous teams register at the beginning, and then there’s a number of checkpoints they have to go through over the duration of the competition – they have to submit safety plans, risk assessments, documentation about demonstrating whether they’ve flown well, and all those sorts of things.
And this year, for example, we had 70 teams enter 18 months before the competition, but only four teams made it through to actually launch their aircraft after all the final checks. The team that came first, they came back with a slightly damaged aircraft, but nonetheless they landed successfully, and we had a team from Canada, called Forward Robotics, who managed to do about 65% of the searching, but they had a problem with their camera. There was… for whatever reason the camera system wasn’t working, so there was nothing for them to look through to actually find Outback Joe. They also successfully returned to the airport. We had two other aircraft that unfortunately crashed on the range.
Glen Paul: So they’re coming along in leaps and bounds. Obviously the challenge demonstrates the potential good that can come from UAV’s in areas such as search and rescue, but of late there has been a focus on their potential misuse, such as being used for intrusive purposes. Do you think as the technology develops it will become a real issue, say with paparazzi spying on celebrities from the air?
Dr Roberts: Oh, of course, you know, any new technology can be used for good things or bad things, and there’s no doubt it has to be controlled. People have to understand what this technology can and can’t do, and we have to really develop a culture around how this sort of technology is used correctly.
Of course there are also regulations, so this industry is actually regulated by CASA, the Civil Aviation Safety Authority, and all people using these sorts of devices for commercial purposes have to actually operate correctly and under those regulations. Everybody really has to know what they’re doing, and act appropriately.
Glen Paul: So is the trend then for bigger or smaller UAV’s?
Dr Roberts: Oh, the trend is for smaller. A lot of the technology developments are being led by the toy industry, so you can now buy – mainly for indoor use I should say – but you can buy these very small flying machines for as little as $50 now, to fly around your house indoors. But the fact that they’re being made smaller, the electronics is becoming smaller and cheaper, that can actually be applied to slightly larger ones outside.
Also the technology in your mobile phone, a lot of the sensors in there, the little accelerometers and gyros that everybody has in their phones, that make their phones magically change the orientation of their screen when they tilt them around, those sorts of sensors can also be used in unmanned aircraft, and the fact that they’re mass produced now for phones means that they’re cheap and available for smaller UAV’s.
Glen Paul: Well let’s take a peep into the future then, big or small, with computer power doubling every couple of years, what will UAV’s be capable of doing for poor Outback Joe say in ten years time?
Dr Roberts: Well I’m really hoping the next time we run this competition, and we’ll run it again, that we actually complete this mission and manage to, as well as automatically find Outback Joe, deliver that water bottle, and have the aircraft return with no human input. So that really would be fantastic – take off by itself, find Outback Joe by itself, drop the water bottle, and come back, no-one doing anything in between – that would be marvellous.
The next planned challenge would then go to the next level, where an aircraft has got to perform a mission and then potentially land at an unprepared airport, maybe a farmer’s bush strip, or a large paddock somewhere out in the outback, where the aircraft’s got to circle around, make some sort of assessment of the area where it’s got to land, and actually plan and decide for itself how to land, and where to land, and which way the wind’s going, and all those sorts of things that a pilot would normally be able to do, and actually land itself.
We would love to then be able to show how a farmer could put in a sample, you know there might some medical emergency and they want a blood sample, put in a little vile of blood into the UAV and have it take off by itself and go back to base. That would really be a fantastic thing to show, maybe in five or six year’s time.
Glen Paul: Righteo. So if someone wants to start preparing now for next year’s competition, where do they start; where do they go to for information about it?
Dr Roberts: So we have our website, which is www.uavoutbackchallenge.com.au . If you simply type in uavoutbackchallege or uavchallenge into any search engine, you’ll find all sorts of references to it, and you won’t have a lot of problem finding things. We’ve also got a Facebook page, a Twitter feed – do a quick search, you’ll find our sites.
Glen Paul: Excellent. And it certainly is an exciting future. It’ll be very interesting to see where it all ends up. Thank you very much for discussing it with me today, Jonathan.
Dr Roberts: OK. Thanks a lot, Glen.
Glen Paul: Doctor Jonathan Roberts. And to find out more about the research, or to follow us on other social media, visit www.csiro.au.