Robots in the skies: how Outback Joe was found and rescued

By Jonathan Roberts, Queensland University of Technology

Lost and thirsty in the Australian bush, Outback Joe waited eight years. And finally, in September this year, he was found – by a flying robot.

Outback Joe is not a real person, but for a week each September over six of the past eight years, he lay in a field waiting to be found. A stuffed dummy in a high-visibility shirt and Akubra hat, Outback Joe was the target of the Unmanned Aerial Vehicle (UAV) Challenge Outback Rescue, an international robotics competition.

Outback Joe has a Facebook page. He’s on Twitter. He likes to go for walks in the country, but sometimes he gets lost. Joe could be you.

Outback Joe is lost and has collapsed in a peanut field. He needs to be found, and desperately needs water.
Stefan Hrabar/CSIRO/UAV Challenge, Author provided

Stories in the media about people getting lost are unfortunately common. Some are lost adventurers, but most are not. Typically, it’s the very young who wander off, the elderly who get confused, or the inexperienced, who don’t appreciate just how easily they can lose track of where they are.

Searching for the lost is a race against time. We have all seen in the media great successes when the lost are found, but also the tragedies when time runs out.

For many years, searching from the air has been a valuable tool, and is required in certain situations. But the opportunity to undertake an aerial search is sometimes limited due to a lack of availability of aircraft or pilots, or practical restrictions such as poor weather.

As technological capabilities continue to develop, search and rescue organisations around the world are keen to use robotic aircraft to help find lost people. Known as remotely piloted aircraft in official circles (and more commonly as drones), these aircraft are usually small, cheap to buy – especially compared to traditional aircraft – and somewhat expendable, as there are no people onboard.

And that’s where we came in.

Lost and found (eight years later)

A group of organisations in Australia launched the UAV Challenge Search and Rescue Competition in 2006, offering a A$50,000 prize for the first team in the world to find and rescue our lost bushwalker, Outback Joe. We specifically developed the competition to guide teams to create low-cost solutions that, one day, could be used for real search and rescue missions.

The Challenge required each team to launch their robotic aircraft from Kingaroy Airport, in rural Queensland, and head out to a large search area approximately four kilometres away.

A robot aircraft returning from a successful mission to rescue Outback Joe.
Stefan Hrabar/CSIRO/UAV Challenge, Author provided

The aircraft had to cover the search area, locate Joe and drop him a 500 millilitre water bottle, ensuring it landed within 100 metres of his position. The aircraft typically flew for up to an hour, and covered between 50 and 100 kilometres of ground during the search.

The UAV Challenge was a true challenge – it wasn’t something that would be completed overnight. In fact, it was a rescue mission that kept 350 teams and more than 2,000 team members busy for eight years.

Some teams came close as time went by, but it wasn’t until 2014 that a team of passionate enthusiasts from Canberra finally won the grand prize. Three other teams also completed the Challenge successfully this year.

The secret to the winning entry was their accuracy of the water bottle drop. The aircraft automatically flew a special pattern around Outback Joe’s location to estimate the wind direction. The on-board computer then calculated the best flight path over Joe taking into account the wind. The result was spectacular, with the water bottle landing just 2.6 metres from Joe.

The moment just before CanberraUAV dropped a water bottle on a parachute to Outback Joe. Outback Joe is in the blue rectangle.
CanberraUAV, Author provided

So did the UAV Challenge do its job and drive down the cost of producing highly capable robot aircraft for search and rescue and inspire innovation in the field? Happily, I think the answer is yes.

The teams that successfully completed the mission did so using specifically designed electronics and new autopilot technologies (which, incidentally, were inspired by the UAV Challenge itself).

One radio communications device developed by a Brisbane-based company has now sold several thousand units worldwide. Another company, based in Adelaide, developed a safety system for teams that has been commercially successful around the world.

The open source software movement embraced the UAV Challenge, seeing three of the most widely used low-cost autopilot developers enter the event. In 2014 alone, software for two of these auto-pilots has been downloaded more than 170,000 times by users across the globe. Flying robots with an automatic search ability could assist Australian State Emergency Services to make rescues more efficient.

The dream of search and rescue organisations is that in the near future they will be able to have small fleets of robot aircraft, each costing no more than a few hundred dollars, stored and waiting for the day they can be used to save a life.

We know this will be possible, in time. The UAV Challenge has demonstrated that – and Outback Joe lives to tell the tale of that time he got lost in the bush.

The Conversation

Jonathan Roberts is a co-founder of the UAV Challenge Outback Rescue.

This article was originally published on The Conversation.
Read the original article.

SpaceShipTwo cost a life, so why do we still use human test pilots?

SpaceShipTwo cost a life, so why do we still use human test pilots?

By Jonathan Roberts, Queensland University of Technology

Tragically, the experimental spaceplane SpaceShipTwo crashed in the Mojave desert during a test flight on Friday, killing one pilot and injuring another.

It is not clear what went wrong, and the coming investigation will no doubt find the cause, but a question arises: should human pilots be testing new rocket engine and fuels in flight when we live in an age when automation and remote control could probably do the job?

Remotely piloted aircraft, or RPAs as the International Civil Aviation Organisation (ICAO) refers to them, are regularly used by the military across the world. They remove the risk to pilots because the pilots remain firmly on the ground while their aircraft fly over potentially hostile ground.

We are now also seeing RPAs being used by civilian organisations and businesses, such as police services, fire departments and media. You might know these new aircraft as drones, but the term “drone” is misleading as pilots on the ground are still in control, even if sometimes they have a more supervisory rather than hands-on role.

The aviation industry has a history of fatal accidents during the testing phase of new developments. This is not because our aviation pioneers were careless; it is because flying in new ways has always been inherently more risky than developing new ways of doing things in most other industries, and still is.

Wilbur Wright after landing a glider in 1901.


When things go catastrophically wrong in an experimental aircraft, the aircraft rarely survives, so the crew of that aircraft have a high chance of death or serious injury. A test pilot is a special kind of person that is prepared to risk their life, more than the average pilot would, in order to help progress the field of aviation.

If brave and brilliant test pilots in the 20th century had not flown in precarious machines, it is clear that we would not have progressed aviation to the point now, where it is one of the safest forms of transport you can strap yourself into.

Covering new ground

SpaceShipTwo was developed by US aerospace company Scaled Composites for future use by the world’s first spaceline Virign Galactic, and on the weekend a flight test was conducted to assess the flight worthiness of a new rocket engine fuel.

This aviation industry is highly regulated and the new fuel and engine had passed ground certification and were cleared for flight tests.

Virgin Galactic has sold several hundred tickets for future flights to the edge of the atmosphere and there are proposals that SpaceShipTwo will be used to send research experiments briefly into space – all worthy endeavours.

But when we have advanced RPAs, why are we still testing new rocket motor and fuel combinations, with human pilots strapped just metres away from them flying at 45,000ft?

Combining spaceplanes with remote control and automation is not new.

Just a few weeks ago, a US Air Force robotic spaceplane came back to Earth after nearly two years in orbit. Its flight was either totally automatic or partially remote controlled; but we have no idea which, as it was a top-secret program.

In 1988, the former Soviet Union’s space program successfully flew their robotic space shuttle Buran, which performed two orbits of the Earth and landed back on a runway, just like the human piloted NASA space shuttles.

The reasons we still sometimes use human test pilots are complex. The SpaceShipTwo vehicle was inspired from SpaceShipOne. That vehicle won the US$10 million Ansari X-Prize in 2004 when Mojave Aerospace Ventures showed that it was possible to launch a non-government developed or funded spaccraft into space with a pilot onboard, twice within two weeks.

SpaceShipTwo’s core business will be to take passengers into sub-orbital space, so of course having human pilots onboard makes sense as eventually there will be human passengers. Would you be prepared to board a commercial jet now if you knew their were no human pilots onboard? Knowing that the pilot’s own safety is twinned with that of the passengers means we feel safer with experts at the helm.

But the SpaceShipTwo disaster was not a commercial flight – it was the first flight-test of a particular engine and fuel combination. Could the engine and fuel have been tested in a rocket first? Could SpaceShipTwo have been designed with an autopilot capability?

Regardless, the aviation industry needs to reflect and consider whether such high-risk flights are acceptable when we have a much safer alternative – one that could actually accelerate the speed of testing, and bring new technology into regular use faster if we did not have people in the new aircraft in those very early, risky days.

Does the passion for flying cloud our judgement? If I was as talented and brave as these SpaecShipTwo test pilots, I would put myself into the vehicle and go for it – but should I be allowed to?

The Conversation

Jonathan Roberts was previously the Deputy Director of the Australian Research Centre for Aerospace Automation.

This article was originally published on The Conversation.
Read the original article.