DIY-DRONES UAV

[dabong1]

At last year’s Paris Air Show, some of the hottest aircraft were the autonomous unmanned helicopters—a few of them small enough to carry in one hand—that would allow military buyers to put a camera in the sky anywhere, anytime. Manufactured by major defense contractors, and ranging in design from a single-bladed camcopter to four-bladed multicopters, these drones were being sold as the future of warfare at prices in the tens to hundreds of thousands of dollars.
In May, at a different trade show, similar aircraft were once again the most buzzed-about items on display. But this wasn’t another exhibition of military hardware; instead, it was the Hobby Expo China in Beijing, where Chinese manufacturers demo their newest and coolest toys. Companies like Shenzhen-based DJI Innovations are selling drones with the same capability as the military ones, sometimes for less than $1,000. These Chinese firms, in turn, are competing with even cheaper drones created by amateurs around the world, who share their designs for free in communities online. It’s safe to say that drones are the first technology in history where the toy industry and hobbyists are beating the military-industrial complex at its own game.

Look up into America’s skies today and you might just see one of these drones: small, fully autonomous, and dirt-cheap. On any given weekend, someone’s probably flying a real-life drone not far from your own personal airspace. (They’re the ones looking at their laptops instead of their planes.) These personal drones can do everything that military drones can, aside from blow up stuff. Although they technically aren’t supposed to be used commercially in the US (they also must stay below 400 feet, within visual line of sight, and away from populated areas and airports), the FAA is planning to officially allow commercial use starting in 2015.

What are all these amateurs doing with their drones? Like the early personal computers, the main use at this point is experimentation—simple, geeky fun. But as personal drones become more sophisticated and reliable, practical applications are emerging. The film industry is already full of remotely piloted copters serving as camera platforms, with a longer reach than booms as well as cheaper and safer operations than manned helicopters. Some farmers now use drones for crop management, creating aerial maps to optimize water and fertilizer distribution. And there are countless scientific uses for drones, from watching algal blooms in the ocean to low-altitude measurement of the solar reflectivity of the Amazon rain forest. Others are using the craft for wildlife management, tracking endangered species and quietly mapping out nesting areas that are in need of protection.

To give a sense of the scale of the personal drone movement, DIY Drones—an online community that I founded in 2007 (more on that later)—has 26,000 members, who fly drones that they either assemble themselves or buy premade from dozens of companies that serve the amateur market. All told, there are probably around 1,000 new personal drones that take to the sky every month (3D Robotics, a company I cofounded, is shipping more than 100 ArduPilot Megas a week); that figure rivals the drone sales of the world’s top aerospace companies (in units, of course, not dollars). And the personal drone industry is growing much faster.

Why? The reason is the same as with every other digital technology: a Moore’s-law-style pace where performance regularly doubles while size and price plummet. In fact, the Moore’s law of drone technology is currently accelerating, thanks to the smartphone industry, which relies on the same components—sensors, optics, batteries, and embedded processors—all of them growing smaller and faster each year. Just as the 1970s saw the birth and rise of the personal computer, this decade will see the ascendance of the personal drone. We’re entering the Drone Age.

What exactly do we mean by drone? The definition has changed over the years, but today it refers to aircraft that have the capability of autonomous flight, which means they can follow a mission from point to point (typically guided by GPS, but soon this will also be possible through vision and other sensors). This differentiates them, on the one hand, from radio-controlled aircraft, which need to be manually piloted, and on the other from uncontrolled vehicles like balloons or ballistic rockets. Usually drones—also known as unmanned aerial vehicles (UAV) or unmanned aerial systems (UAS), to include the ground-station components—also carry some sort of payload, which at a bare minimum includes cameras or other sensors as well as some method to transmit data wirelessly back to a base.

That definition fits a $140 million Global Hawk drone, circling over Afghanistan and transmitting video to Air Force intelligence analysts in California. But it also describes the $500 foam plane that my children fly on weekends. Both have sophisticated computer autopilots, high-resolution cameras (we’re partial to GoPros), wireless data connections for video and telemetry, ground stations with heads-up displays and real-time video (my kids were disappointed at a recent tour of the Oshkosh air show to see that today’s military drone pilots have worse ground stations than they do), step-by-step mission scripting, and the capability to play back footage of the mission in full. The main difference between the two drones is that the Global Hawk can fly at 60,000 feet for 32 hours and our craft can fly at 400 feet for 30 minutes. (What we lack in high-altitude optics we make up in proximity: We can easily read license plates from the air.)

The key ingredient in a drone is the autopilot, a technology that first came into use as a flying aid in the 1930s. Initially, all that autopilots did was keep the aircraft level. A combination of a barometric-pressure meter, a compass, and mechanical gyroscopes (motorized flywheels with analog electrical outputs) allowed a pilot to set a heading and altitude and take a nap, knowing that the aircraft would continue to fly straight ahead until told otherwise. Starting with commercial jetliners, pilots could set waypoints and the autopilot would fly an entire route. By the early 1990s, aerospace technology could automate an entire flight, including takeoff and landing (though FAA rules still require that commercial pilots handle takeoffs manually).

Today, all the sensors required to make a functioning autopilot have become radically smaller and radically cheaper. Gyroscopes, which measure rates of rotation; magnetometers, which function as digital compasses; pressure sensors, which measure atmospheric pressure to calculate altitude; accelerometers, to measure the force of gravity—all the capabilities of these technologies are now embedded in tiny chips that you can buy at RadioShack. Indeed, some of the newest sensors combine three-axis accelerometers, gyros, and magnetometers (nine sensors in all), plus a temperature gauge and a processor, into one little package that costs about $17.

Meanwhile, the brain of an autopilot—the “embedded computer,” or single-chip microprocessor, that steers the plane based on input from all the sensors—has undergone an even more impressive transformation, thanks to the rise of the smartphone. Once Apple’s iPhone showed that fluid and fast visual interfaces on touchscreens were what people wanted, the same insatiable demand for computational power that kicked in with the graphical user interface of desktop computers came to phones. But unlike the desktop, these mini supercomputers also needed to use as little power as possible. The result was a shift to the hyperefficient “reduced instruction set computing” architectures—led by British chip designer ARM, which now dominates the single-chip industry—driving the performance gains of our smartphones and tablets. As it turns out, these chips are also perfect for drones: Fast and power-efficient processors mean that they can go beyond simply following a preprogrammed mission and start to think for themselves.

And the smartphone-drone connection goes far beyond the processors. These days, a standard smartphone has a full suite of sophisticated inertial sensors to detect its position, a feature that’s integrated into everything from games to maps and augmented reality. The demand for higher-quality cameras in phones has launched a similar revolution in image-capture chips, which are used in drones. The need for smaller, better GPS in phones has brought the same technology to drones, too, such that GPS performance that cost tens of thousands of dollars in the 1990s can be had for as little as $10 in a thumbnail-sized device. The same goes for wireless radio modules, memory, and batteries.

In short, this new generation of cheap, small drones is essentially a fleet of flying smartphones. More and more, autopilot electronics look just like smartphone electronics, simply running different software. The technical and economic advantages of coattailing on the economies of scale of the trillion-dollar mobile-phone industry are astounding. If you want to understand why the personal-drone revolution is happening now, look no farther than your pocket.

Every industry has its garage-creation myth. Here’s mine, on the start of the personal-drone movement. One sunny Friday afternoon in March 2007, I started planning what I’d hoped would be a deliciously geeky weekend with the kids. In the usual stack of products that had come into the Wired offices that day to be reviewed, there were two that seemed especially promising: a robotics kit and a ready-to-fly radio-control airplane. I settled on a schedule: We would build robots on Saturday and fly planes on Sunday. Awesomeness would surely ensue.

By midmorning on Saturday, things were already going wrong. The kids were happy enough to open the robotics kit (from Lego’s Mindstorms line) and assemble the starter bot, a three-wheeled rover. But once we powered it up, they could barely hide their disappointment. Hollywood, it turns out, has ruined robotics for kids, who now expect laser-armed humanoid machines that also transform into trucks. Back in the real world, after an hour of assembly and programming, the rover could only roll forward and bounce feebly off a wall. Online, we could see that hobbyists were doing amazing things with Mindstorms: robotic Rubik’s Cube solvers, working photocopiers, and more. We wanted to invent something like that, but it was impossible to see how. The kids lost interest after lunch.

OK, but at least we still had the plane. On Sunday, we took it to a park, and I promptly piloted it into a tree. The kids just looked at me, appalled not merely by my lack of ability but also by the yawning gap between the promised coolness of the plane itself and the actual experience of flying it. I threw sticks at the plane to try to dislodge it from the tree as my mortified children pretended not to know me. My geekdad weekend was an utter failure. I was annoyed at myself for getting it so wrong and annoyed at my kids for being so unappreciative. I went for a run to let off some steam.

While on the run, I started thinking more about the impressive range of sensors that Mindstorms had. There were accelerometers (“tilt sensors”), electronic gyroscopes, a compass, and a Bluetooth link that could connect to a wireless GPS sensor. It occurred to me that those were exactly the same sensors you’d need to make an airplane autopilot. We could solve both problems at once: build something cool with Mindstorms that had never been done before and get the robot to fly the plane! It was sure to be a better pilot than me. The moment I got home, I prototyped a Lego autopilot on the dining room table, and my 9-year-old helped write the software. We took some pictures, posted them, and our project was on the front page of Slashdot by that evening. We put it in a plane—the world’s first Lego drone, I think—and took it out a few weekends later. It almost-kinda worked, staying aloft and steering on its own, albeit not always to the places we told it to go.

With a few weeks more of tinkering, I developed a Lego autopilot that had most of the functionality of a professional device, if not the performance. But it became clear that Mindstorms, for all its charms, was too big and expensive to serve as the ideal platform for homemade drones. Looking for a better way, I decided to conduct my search for answers online in public, sharing what I’d done and found. Instead of setting up a blog, I registered DIYDrones.com and established a social network for people who were experimenting with autonomous aircraft.

That distinction—a site created as a community, not a one-man news and information site like a blog—turned out to make all the difference. Like all good social networks, every participant—not just the creator—has access to the full range of authoring tools. Along with the usual commenting, they can compose their own blog posts, start discussions, upload videos and pictures, create profile pages, and send messages. Community members can be made moderators, encouraging good behavior and discouraging bad. Open to anyone who chose to participate, the site was soon full of people trading ideas and reports of their own projects and research.

Initially, members would just post code and design files for their own projects, showing off for each other in a form of nerd braggadocio. But over time we set up more-organized systems of collaboration, including version control systems and file repositories, wikis, mailing lists, and formal team assignments. I was blown away by what people in our community were doing with sensors from mobile phones and chips that cost less than a cup of coffee. Feature by feature, they were matching—or besting—aerospace electronics that had cost tens or hundreds of thousands of dollars just a decade earlier. It felt like the future of aviation: Just as the PC emerged from the Homebrew Computer Club hobbyists and eventually overturned mainframe-based corporate computing in the 1980s, I suddenly saw how the same sort of movement would bring robots to the skies.

For me, it’s become not just a hobby but a second career: I cofounded a company, 3D Robotics, to make the open source hardware that the DIY Drones community was designing, and we’ve already shipped more than 10,000 autopilots and countless other drone parts. By our estimates, 3D Robotics’ customers alone are flying more drones than the total number operated by the US military (7,494 today, according to a recent congressional report). And there are dozens of other companies making drone technology for the “hobby” market, including Hoverfly, DJI Innovations, MikroKopter, Droidworx, and uThere.

But really, I do this because I love exploring the boundaries of what’s possible with the technology. This summer I’m planning to tinker with the ultimate “personal droid” idea: an autonomous plane that can film your own activities at the press of a button. For example, if you’re into extreme sports, such as kite surfing, you’re always looking for better ways to document your exploits. The best vantage point to do this from is the air, about 30 feet above and behind you. That’s a perfect job for a drone. Just imagine if you could touch a button on your iPhone or Android phone, and it would summon a quadcopter to position itself above you, keeping its camera on you as you perform your stunts then flying back to shore when its batteries got low. We’ve already assembled a module that contains a GPS sensor to record your position and a wireless module to communicate with the drone; porting it to a waterproofed smartphone is the next step. Your personal cameradroid awaits!

Drones in pop culture have gotten something of a bad reputation. In novelist Daniel Suarez’s new techno-thriller, Kill Decision, clouds of killer drones are programmed to swarm like weaver ants, and they attack everything from people to entire container ships. The gun-carrying quadcopters arrive by the thousands and hurl themselves at windows and walls until they break through, sacrificing themselves in countless numbers so that others of their kind can advance. They are entirely autonomous and make their own decisions about where to go and what to shoot. They are single-use, like ammunition, and mass-produced, like cheap cell phones—a vision of drones as a disposable commodity modeled after insects, not planes.

There’s no reason to believe that cheap drones will usher in a weaponized hellscape any more than the invention of helicopters did. Even on privacy issues, there are existing laws that cover most of the concerns people have about personal drones. But in its vision of flying robots that are mass-produced like cheap toys and smart enough to think for themselves, Suarez’s fiction is closer to reality than most people think. Indeed, he was inspired not by military drones but by the Parrot AR.Drone, a quadcopter toy you can buy on Amazon and control with your smartphone, complete with dual cameras transmitting real-time video streams to your screen. Although it’s not autonomous—you still have to fly it yourself—that’s simply a matter of design choice. For less than $30, you can buy a little circuit board that can connect it to an autopilot.

People are already dreaming of what companies could do with drones over America, from playful speculations like the “Tacocopter,” a meal-delivery service, to real game changers such as FedEx fleets that use hyperefficient aircraft designed from the start as drones. A switch to unmanned operation would transform the entire concept of air cargo: Aircraft would be free of the design constraints—pressurized cabins, tube-shaped bodies—necessary to accommodate humans, and flocks of such drones could fly in a V-formation like birds to employ efficient aerodynamic drafting.

What we will do with our personal drones? That question is just as unanswerable—but just as tantalizing—as the same question about personal computers back in 1977. When the Apple II came out, the answer was not much more than “Program it!” But over time, as regular people found uses for PCs in their own lives, they came up with better answers: word processors, spreadsheets, videogames, email, and, eventually, the web. Today we know what personal computers are for, but it took the liberation of the technology to show us.

So too for personal drones. Remember, the military created the Internet, but the people colonized it and created the web for their own purposes. The amateur UAV community is hoping to do the same with drones—demilitarize and democratize them so they can find their full potential. There will be good uses and bad ones, but the same is true of any tool, from a crowbar to an ultrasound machine. Ultimately the way society best figures out how to think about a powerful new technology is to set it free and watch where it flies.

Chris Anderson (canderson@wired.com) is editor in chief of Wired. He started DIY Drones and cofounded 3D Robotics.
How I Accidentally Kickstarted the Domestic Drone Boom | Danger Room | Wired.com