While understandable, it is also a little surprising. Many of us used to dream of owning (and thanks to those $40 shopping mall jobbies, probably now do own) an RC helicopter. These little model-sized flying devices often do little more than rotor up and fly a few metres away, but they’re cheap, fun, and relatively harmless. You can tell when a technology has become commonplace: it’s when you can crash it into a wall and laugh about it, instead of sobbing over a $2500 pile of burning steel and plastic. 

The problem now, though, is that thanks to modular design, miniaturised devices like cameras and GPS, and of course the ever declining prices of sophisticated electronics and flight tech, it’s a very short step from the little helicopter to the slightly larger, but significantly more sophisticated distance-travelling, photo-taking, highly manoeuvrable UAV.

A Question of Science

For people like Duncan Campbell, associate professor in Robotics and Aerospace Systems at Queensland University of Technology and a member of the Australian Research Centre for Aerospace Automation (ARCAA), this kind of convergence and consumer-level flow-on presents unique challenges for safety and regulation of all kinds of technologies, but especially for UAVs. New tech is being developed by ARCAA to try and minimise potential problems with UAV use, and to improve the prospects of more widespread civilian use of the aircraft. 

“There are two identified key technology challenges,” he said, “one is called the See and Avoid technology, and really that’s picking up on what a human pilot would do. A private pilot is basically using eyes as the primary method for visual detection for possible or oncoming collision scenarios. That’s identified as being at least the minimum that we need to at least mimic in a UAV.

"A private pilot is basically using eyes as the primary method for visual detection. That’s the minimum that we need to at least mimic in a UAV.""The second one is around what’s called the forced landing problem, or emergency landing. That is, if the aircraft has a failure, say an engine failure, then how can it get itself down safely in order to avoid injuring people or damaging property, and ideally to protect itself? Again, coming back to a vision-based system is key there, having a camera looking roughly down, and - doing as a human pilot is encouraged to do - having a good situational awareness of what possible landing sites exist.” 

According to Campbell, See and Avoid (also known as Detect and Avoid) is already at the commercialisation stage, while vision based landing is only just about to make it off the drawing board. But this technology, while obviously providing safety benefits, also shows that the industry is in a state of transition, which itself causes all sorts of problems in terms of regulation.

The Days Before

So it is that our Civil Aviation Safety Authority (CASA) is currently in the middle of UAS regulation reform and is particularly looking at licensing and certification. Currently, to fly a UAS for anything other than sport and recreation (in other words, if you’re somehow earning money from flying them around), you have to pass the same theory exams and modules as a manned aircraft pilot in order to be certified. 

“The background to what we’re doing was based on a scenario from some ten or twelve years ago. As a result, technology and access has changed remarkably"However, given the variety and complexity of UAS tech, and the growing use not only by industry but by private individuals and smaller businesses, it’s time that the relevant legislation and guidelines, now over ten years old, were brought up to speed. That happens to be the job of Phil Presgrave, the UAS Specialist at CASA and the man leading reform from the government side of things.

“The background to what we’re doing was based on a scenario from some ten or twelve years ago,” he said. “As a result, technology and access has certainly changed remarkably, so all sorts of people have dollar signs in their eyes, and believe they can achieve a commercial result from flying unmanned aircraft, and can offer a superior product for a lot less than what you could get a conventional aircraft for, to do the same thing.

“When the regulation was written, we were imposed with conditions that all relate back to conventional aircraft, like the radio certificate of proficiency, like the aviation license and theory exams, like the requirement for an instrument rating exam. What we’re doing in the new syllabus is drafting those things that apply directly to unmanned aircraft.”

Today's Model is Tomorrow's...

One of the looming problems for the industry is also one of its most fundamental - what should be considered a ‘model’, and thus be exempt from most regulations, and what counts as an ‘aircraft.’? The distinction currently comes down to what you’re using it for - are you going to charge people to do a topographic analysis of a potential mining site, or do you just want to scare relatives with your mad dive-bombing skills? 

"If you’ve got two aircraft side by side, exactly the same, which is the model and which is the remotely piloted aircraft? You don’t know until you use one."Now though, with the tech that’s available, it’s not difficult to see an increasing overlap between a model and an aircraft in terms of capabilities, possible applications, and subsequent risk. It used to be that only corporations could afford a fully-fledged aerial system. Now, it’s possible to grab parts from consumer-level shelves. It’s a problem that everyone - including CASA, the model flight guys and industry - has their eyes on. 

“It does make it difficult,” says Presgrave, “and what I’m trying to develop I guess is a more sophisticated definition of remotely piloted aircraft. If you’ve got two aircraft side by side, exactly the same, which is the model and which is the remotely piloted aircraft? You don’t know until you use one. I think, what’s probably going to have an impact in the future is the onboard systems of the aircraft, their accuracy, their reliability, certainly their cost, all of those sorts of functions will start to come into play. 

“I would think, therefore, that from a model’s perspective, you wouldn’t want to spend thousands of dollars on something that you didn’t need to maintain at a height within twenty centimetres, or tracking distance within twenty centimetres, or whatever the case may in fact be.” 

Mike Close, the head of the Model Aircraft Association Australia, and chairman of the MAAA’s UAV Working Party, is pretty much the go-to guy when it comes to the model aircraft side of things, and was at time of writing in the United States, as an observer of similar discussions between the American Model Association and their Federal Aviation Authority. The Obama administration recently laid out a timeline for integration of unmanned aircraft into general US air space, a move that hasn’t been without its share of controversy. 

“The potential overlap [between models and aircraft] is a concern to both the regulators and the operators,” he said via email from the US. “As the MAAA, we have to ensure safety whilst at the same time try to avoid regulations that are beyond those that can be justified on a risk assessment basis for the specific application. 

“Whilst the flight technology that is built into what we call Self- Guided Model Aircraft, (or SGMA), may have a similar performance to that contained in the bottom end UAV or UAS, it is the limitation of use that we rely on to maintain the differentiation between model aircraft and those used for any other purpose.” 

While use is the main factor in distinguishing between models and aircraft, it doesn’t seem to be the only one. According to CASA’S Advisory Circular 101-3, a model aircraft has to come in at under 150kg in weight, including fuel. But what happens if a UAS used for fun comes in over that weight? While an actual case doesn’t seem to have come up yet in Australia (though according to Mr Close it has in the US and Europe), Presgrave says he has been asked that specific question in his recent discussions with stakeholders, and there doesn’t seem to be a clear-cut answer. 

"It is the limitation of use that we rely on to maintain the differentiation between model aircraft and those used for any other purpose.”“Not being a student of weights and model aircraft and what have you, it certainly does raise a question, what does happen after that 150 kilo bracket?” said Presgrave. “I suspect there are models out there that are larger than that, but again that’s not been my purview and I have not chased it.” 

Seemingly, regulators from ten years ago couldn’t conceive of anyone wanting to use something that weighed that much just for kicks. Probably in the same way people couldn’t imagine others wanting to use the internet for cat videos. Yet here we are. 

Droning On

Still, the existence of this kind of problem is understandable. Laws and regulations, by their nature, aren’t designed to change quickly. Tech will always run ahead of law reform; at least CASA has recognised that and is moving to change things up. 

But for UAVs it’s a little trickier than in many other industries, not least because of the mix in public opinion. As Presgrave is at pains to tell me, when we say ‘drones’, people don’t think of unmanned vehicles that can transport goods, capture amazing aerial photography, search for people lost in bushland, and otherwise do jobs that manned aircraft can’t do. They see Afghanistan, they see explosives, they see spy planes, and they see that camera aimed through their window instead of at the horizon. 

“If that had happened in a civilian sense, then all remotely piloted aircraft would be grounded, there would be no commercial operations whatsoever,”The recent media attention on the US-Australia agreement to house a UAV base on our soil is proof of that. As a result, the word ‘drone’ seems almost to be anathema to industry. The emphasis from various industry and government spokespeople on the acronym UAS seems as much a PR exercise as it is an earnest attempt to name the devices as matter-of-factly as possible. 

For Presgrave, there’s one particular example that shows this distrust of UAVs in vivid detail. A US military C-130 Hercules transport helicopter was involved in a mid-air collision with an RQ-7 Shadow UAV while on duty in Afghanistan late last year. The Hercules made it down safely, but there was still quite the noise made about the incident. It was said by some that that this showed UAVs were too unsafe for use in a civilian context. While an inquiry into the incident cleared the Shadow of wrongdoing, the ripples are still spreading, and show how razor-thin the line industry finds itself walking actually is. 

“If that had happened in a civilian sense, then all remotely piloted aircraft would be grounded, OC operator certificates would be withdrawn, there would be no commercial operations whatsoever,” says Presgrave. 

“That would be the overreaction, even though in the case of the Hercules and the Shadow, it wasn’t the UAV remote pilot’s problem, he was flying as he was required to fly. He was in the right spot, the Hercules pilot chose to cut the corner.”

A Problem Shared

The UAV industry isn't the only one running into this problem. Google, in between dominating web search and causing governments privacy headaches the world over, also seems to find time to design cars that drive themselves. Just because they can. They’ve even driven them on actual roads, clocking up more than 300,000 km in total. 

The cars are street legal as long as someone is sitting inside, ready to override if it looks likely to cause an accident, but according to Google, almost every time the car has been overridden, it was when a human driver in another vehicle did something unexpected and illegal. The car’s only crash was when a human was actually driving. 

“I can vaguely remember when elevators in buildings used to have a person on the little control handle there. There was a similar discussion around automating lifts.”In some respects, the integration of human and machine intelligence in the skies is a more hair-raising prospect than the actual machines themselves. It’s certainly a problem if we ever plan to have unmanned, remote-controlled passenger flights. For Professor Campbell, though, that integration will be necessary, and we simply have to find ways to make it work. 

"I think that’s an issue. I’m not sure the answer is to, you know, chuck the human pilots out, but particularly when you’ve got an aircraft full of people sitting behind you, I think the issues are a bit deeper than that. It also comes back to what we’re doing with UAVs, which is demonstrating the safety case. 

“I can vaguely remember when elevators in buildings used to have a person on the little control handle there. I think there was a similar discussion around automating lifts, you just have to get that public confidence going.” 

In some ways, air safety is the least of the humble little UAV’s problems. While CASA plans to liaise with the Privacy Commissioner down the track to get the rubber stamp on its new regulations, the truth is that there’s probably always going to be some measure of concern around privacy. Things aren’t necessarily helped when our own government suggests using UAVs to check license plates. 

Even still, UAVs are definitely in Australia’s future. From your mate down the road who does a little aerial photography on the side, to emergency services using UAVs to track the spread of fire, to the transnational corporations packing these things to the brim with cutting edge hardware, it’s not hard to see these devices coming into greater and greater use. UAVs are here.

Just don’t mention the war.

This article originally appeared in the May 2012 issue of Australian Popular Science.