But their ultimate value may be in making robots and the algorithms that run them even smarter by providing more data from new perspectives. JL
Chris Anderson reports in Harvard Business Review:
Drone economics are classically disruptive. Drones can accomplish in hours tasks that take people days. They provide detailed visual data for a tiny fraction of the cost of acquiring the same data by other means. And they bring new perspective and capabilities to fill the “missing middle” between satellites and street level, digitizing the planet in high resolution and near–real time at a tiny fraction of the cost of alternatives.
Every morning at the construction site down the street from my office, the day starts with a familiar hum. It’s the sound of the regular drone scan, when a small black quadcopter flies itself over the site in perfect lines, as if on rails. The buzz overhead is now so familiar that workers no longer look up as the aircraft does its work. It’s just part of the job, as unremarkable as the crane that shares the air above the site. In the sheer normalness of this — a flying robot turned into just another piece of construction equipment — lies the real revolution.
“Reality capture” — the process of digitizing the physical world by scanning it inside and out, from the ground and the air — has finally matured into a technology that’s transforming business. You can see it in small ways in Google Maps, where data is captured by satellites, airplanes, and cars, and presented in both 2-D and 3-D. Now that kind of mapping, initially designed for humans, is done at much higher resolution in preparation for the self-driving car, which needs highly detailed 3-D maps of cities in order to efficiently navigate. The methods of creating such models of the real world are related to the technology of “motion capture,” which drives movies and video games today. Normally that requires bringing the production to the scanners — putting people in a large room outfitted for scanning and then creating the scene. But drones flip that, allowing us to bring the scanner to the scene. They’re just regular cameras (and some smart software) precisely revolving around objects to create photo-realistic digital models.
In some ways it’s astonishing that we’re using drones on construction sites and in movies. Ten years ago the technology was still in labs. Five years ago it was merely very expensive. Today you can buy a drone at Walmart that can do real enterprise work, using software in the cloud. Now that it’s so cheap and easy to put cameras in the sky, it’s becoming commercially useful. Beyond construction, drone data is used in agriculture (crop mapping), energy (solar and wind turbine monitoring), insurance (roof scanning), infrastructure (inspection), communications, and countless other industries that touch the physical world. We know that “you can manage only what you can measure,” but usually measuring the real world is hard. Drones make it much easier.Industries have long sought data from above, generally through satellites or planes, but drones are better “sensors in the sky” than both. They gather higher-resolution and more-frequent data than satellites (whose view is obscured by clouds over two-thirds of the planet at any time), and they’re cheaper, easier, and safer than planes. Drones can provide “anytime, anywhere” access to overhead views with an accuracy that rivals laser scanning — and they’re just getting started. In this century’s project to extend the internet to the physical world, drones are the path to the third dimension — up. They are, in short, the “internet of flying things.”You might think of drones as toys or flying cameras for the GoPro set, and that is still the lion’s share of the business. But like the smartphone and other examples of the “commercialization of enterprise” before them, drones are now being outfitted with business-grade software and becoming serious data-collection platforms — hardware as open and extensible as a smartphone, with virtually limitless app potential. As in any app economy, surprising and ingenious uses will emerge that we haven’t even thought of yet; and predictable and powerful apps will improve over time.The peripatetic nature of drones isn’t so different from the nature of Chris Anderson’s career. Both move hither and thither with ease, accomplishing much along their respective paths. Anderson studied computational physics, played in a band called REM (sadly, not the REM), worked at Los Alamos National Laboratory, and eventually landed at two august publications: Nature and Science. While editor in chief during the rise of Wired magazine, he wrote several best-selling and influential books: The Long Tail; Free; and, most recently, Makers: The New Industrial Revolution. Along the way he founded several companies (now “lost to the mists of time,” he says) and won too many influential thinker awards to list here.
Or you might think of drones as delivery vehicles, since that’s the application — consumer delivery — that the media grabs on to most ferociously when seeking click-generating amazing/scary visions of the future. Frankly, delivery is one of the least compelling, most complicated applications for drones (anything that involves autonomously flying in crowded environments is the black-diamond slope of technology and regulation). Most of the industry is focused on the other side of the continuum: on data, not delivery — commercial use over privately owned land, where the usual concerns about privacy, annoyance, and scary robots overhead are minimized.Anderson’s first encounter with drones came after he tried to get his kids interested in robots (which they found slow and boring) and remote-controlled planes (which crashed into trees). He thought, What if the robot could fly? It would be more fun, and it couldn’t be a worse pilot than me. So, as he told one reporter, “I literally googled ‘flying robot’ and got to ‘drone’ and googled ‘drone’ to get to ‘autopilot.'”Anderson believes that we’re in the first minutes of a massive transformation with drones and other ways to extend sensors into the world. The key breakthrough, he argues, comes when you stop thinking about the hardware as the product and start thinking about the data drones can collect as the product. “We are in the process of measuring the planet at unprecedented resolution in both time and space,” he says. “The next step will be to figure out how to use all that data to manage it.”
Drone economics are classically disruptive. Already drones can accomplish in hours tasks that take people days. They can provide deeply detailed visual data for a tiny fraction of the cost of acquiring the same data by other means. They’re becoming crucial in workplace safety, removing people from precarious processes such as cell-tower inspection. And they offer, literally, a new view into business: Their low-overhead perspective is bringing new insights and capabilities to fields and factories alike.
Like any robot, a drone can be autonomous, which means breaking the link between pilot and aircraft. Regulations today require that drones have an “operator” on the ground (even if the operation is just pushing a button on a smartphone and idly watching as the drone does its work). But as drones are getting smarter, regulators are starting to consider flights beyond “visual line of sight” — ones in which onboard sensors and machine vision will more than compensate for the eyes of a human on the ground far away. Once such fully autonomous use is allowed, the historic “one pilot/one aircraft” calculus can become “one operator/many vehicles” or even “no operator/many vehicles.” That’s where the real economic potential of autonomy will kick in: When the marginal cost of scanning the world approaches zero (because robots, not people, are doing the work), we’ll do a lot more of it. Call this the “democratization of earth observation”: a low-cost, high-resolution alternative to satellites. Anytime, anywhere access to the skies.
The drone economy is real, and you need a strategy for exploiting it. Here’s how to think about what’s happening — and what’s going to happen. We’ll start back at the construction site, a work environment in desperate need of what drones can provide.
Capturing Reality for the Cost of a Nice Lunch
The construction industry is the world’s second largest (after agriculture), worth $8 trillion a year. But it’s remarkably inefficient. The typical commercial construction project runs 80% over budget and 20 months behind schedule, according to McKinsey.
On-screen, in the architect’s CAD file, everything looks perfect. But on-site, in the mud and dust, things are different. And the difference between concept and reality is where about $3 trillion of that $8 trillion gets lost, in a cascade of change orders, rework, and schedule slips.
Drones are meant to close that gap. The one buzzing outside my window, taking passes at the site, is capturing images with a high-performance camera mounted on a precision gimbal. It’s taking regular photos (albeit at very high resolution), which are sent to the cloud and, using photogrammetry techniques to derive geometries from visual data, are turned into photo-realistic 2-D and 3-D models. (Google does the same thing in Google Maps, at lower resolution and with data that might be two or three years old. To see this, switch to Google Earth view and click on the “3-D” button.) In the construction site trailer, the drone’s data shows up by mid-morning as an overhead view of the site, which can be zoomed in for detail the size of a U.S. quarter or rotated at any angle, like a video game or virtual reality scene. Superimposed on the scans are the CAD files used to guide the construction — an “as designed” view overlaid on an “as built” view. It’s like an augmented reality lens into what should be versus what is, and the difference between the two can be worth thousands of dollars a day in cost savings on each site — billions across the industry. So the site superintendent monitors progress daily.
Mistakes, changes, and surprises are unavoidable whenever idealized designs meet the real world. But they can be minimized by spotting clashes early enough to fix them, work around them, or at least update the CAD model to reflect changes for future work. There are lots of ways to measure a construction site, ranging from tape measures and clipboards to lasers, high-precision GPS, and even X-rays. But they all cost money and take time, so they’re not used often, at least not over the entire site. With drones, a whole site can be mapped daily, in high detail, for as little as $25 a day.
Drones take off
READ ALL CHARTS
Rising from the Ground to Fill the Missing Middle
The ascent of the drone economy is a steep one. Ten years ago unmanned aerial vehicles were military technology, costing millions of dollars and cloaked in secrecy. But then came the smartphone, bringing with it a suite of component technologies, from sensors and fast processors to cameras, broadband wireless, and GPS. All those chips enabled the remarkable supercomputer in your pocket, but the economies of scale of smartphone production also made them cheap and available for other uses. The first step was to transform adjacent industries, including robotics. I call this proliferation of components “the peace dividend of the smartphone wars.”
Companies including my own came out of this moment. Cheap high-powered components and a maker’s attitude allowed enthusiasts and entrepreneurs to reimagine drones not as coming down from higher in the sky but as rising from the ground. Rather than seeing “airplanes without pilots,” we saw “smartphones with propellers.” Moving at the pace of the smartphone industry, not the aerospace industry, drones went from hackers’ devices to hobbyists’ instruments to toys costing less than $100 at your local big-box store in less than four years — perhaps the fastest transfer of technology from CIA to Costco in history. Five years ago the main commercial objection to the word “drone” was that it had military connotations. Now it’s that people think of the aircraft as playthings. Has any word changed its meaning from “weapon” to “toy” faster?
And it doesn’t end there. Wave one was technology, wave two was toys, and now comes the third and most important wave. Drones are becoming tools. The market for people who want flying selfie cameras may be limited, but the market for data about the physical world is as big as the world itself.
Drones are starting to fill the “missing middle” between satellites and street level, digitizing the planet in high resolution and near–real time at a tiny fraction of the cost of alternatives.
The trajectory of this third wave — drones as tools — is more dramatic than that of the two preceding waves. First drones will populate the skies in increasing numbers as regulations and technology allow safer use. Estimates vary widely; some data predicts that by next year more than 100,000 operators will be managing 200,000 drones that will fill the sky, doing some work or another.
Next, the market for drone apps will explode as more and more people find ingenious uses. Drones will remain primarily data-collection vehicles, but the breadth of apps for them is only just beginning to be discovered. For example, drones have already been used for search and rescue and for wildlife monitoring. They can provide wireless internet access (something Facebook is investing in) and deliver medicine in the developing world. And they can not only map crops but also spray them with pesticides or deposit new seeds and beneficial insects.
Then, drones will gain even greater cost advantages when they don’t just remove the pilot from the cockpit but remove the pilot entirely. The true breakthrough will come with autonomy.
Autonomous, Small, and Countless
Technology to allow drones to fly themselves exists and is improving quickly, going from simple GPS guidance to true visual navigation — the way a human would fly. Take humans out of the loop, and suddenly aircraft look more like the birds that inspired them: autonomous, small, and countless; born for the air and able to navigate it tirelessly and effortlessly. We are as yet tourists in the air, briefly visiting it at great cost. By breaking the link between man and machine, we can occupy the skies. The third dimension is the last frontier on Earth to be properly colonized (yes, both up to the skies and down under the seas, but we’ll leave the latter to our aquatic-drone cousins). Colonize it we will, but as with space and the ocean depths, we’ll use robots, not humans.
Why now? A combination of three trends. First, the price/performance bounty of the smartphone tech we talked about earlier made drones cheap and good. For example, the gyroscopic and other sensors packed into a tiny $3 chip in your phone were just a decade ago mechanical devices costing as much as $100,000 and mounted in enclosures ranging in size from lunch boxes to dorm fridges.Second, the ability to make cheap and good drones put them within the reach of regular consumers (willing to spend up to $1,000) who had a real use case (aerial video and photography). As a result, companies had to make them easy to use — just swipe and fly — to drive adoption. Drones had to become more sophisticated as users became less sophisticated.
Third, once the consumer drone boom unexpectedly put more than a million drones — ranging from small toys to high-end “prosumer” models — into the skies over America in less than four years under a “recreational use” exemption to the FAA’s strict rules about flying things, the regulators had to respond. To steer the market toward safer use without inhibiting it, the agency accelerated rules that would allow drones to be used commercially without the need for pilots’ licenses or special waivers. The new rules took effect in August 2016, essentially kicking off the commercial drone era.
The Rise of Cloud Robotics
To this point we’ve focused mostly on drones themselves — the hardware, its cost and capabilities, and what we can attach to it to get work done. But when setting a drone strategy, it’s important to think less about drones and more about apps. The hardware is primarily an empty vessel to fill with work to be done: taking photographs and video, scanning, moving objects, enabling communication.
And collecting data. More than anything, drones are collection vehicles. Their ability to amass data from a unique, valuable perspective (above, but not too far above) fast and at low cost makes them ideal collectors. Any drone strategy has to go beyond the drone to the data. And that means moving innovation to the cloud.
“Cloud robotics” is just the combination of the last two activities: connecting robots to the cloud so that both get smarter. That includes all robots — not just drones but also driverless cars, manufacturing and warehousing robots, and maybe someday robots in your home. But for now, we’ll focus on drones.
The biggest change in drones (and in robotics — indeed, in electronics broadly) over the past decade is the assumption of connectivity. Unlike earlier generations of robots, which required bespoke communications systems, the robots that have come out of the smartphone industry inherited their “born connected” architecture.
Already it’s hard to remember how things used to work: Amass data, then download it, then analyze it. No more. Data flows from source to device to analysis automatically and invisibly. Increasingly, it does what technology should always do: just work.
The implications of this shift are profound. When devices are designed from the ground up to be connected, three big things change:
1. The devices tend to get better over time, not worse. Unlike in the old stand-alone model, in which products start their march to obsolescence the moment they are made, connected devices get most of their features from their software, not their hardware, and that software can be updated, just like the software on your smartphone. Think of a Tesla, which gets new features automatically on an almost weekly basis. The technical term for such devices is “exotropic,” and they tend to rise in value over time — unlike “entropic” devices, whose value tends to decline. Of course, the hardware has limits, and eventually even connected devices become obsolete. But the point is that rather than follow the traditional long decay slope from the point of purchase, connected devices improve in utility for as long as they can. In the case of drones, new abilities, from improved performance to new autonomous features, just appear overnight via “over the air” upgrades.
2. They have “outboard intelligence.” They’re part of the internet of things — not the silly part, like connected lightbulbs, but the clever part (which, being clever, usually avoids the buzzwordy internet-of-things label). For example, Amazon Echo has enough intelligence in the box to harness immense intelligence in the cloud. It’s not just a sensor for the internet but also a limb by which the internet can project into the physical world. For a drone, this means that it doesn’t have to be programmed to scan a site using a standard path. Instead, it starts by taking a few pictures of the site, and then it uploads them to the cloud so that algorithms there can analyze them in real time and prepare a custom scan path that’s just right for that site, on that day, with that lighting and those shadows. Think of this as the data determining the mission, not the mission determining the data.
3. They make the internet smarter too. Connected devices don’t just get intelligence from the network; they feed data back to it. The current AI renaissance is due less to improved computation and algorithms than to the ability simply to access vastly more data. Much of that data, today and tomorrow, comes from measuring the world — both people and their environments — and connected devices are how the sensors spread. In the case of drones, this means they can not only download up-to-date 3-D maps of their world to help them navigate but also potentially upload data to make those maps better.
Cool Is Not Enough
Where all this really kicks in is the enterprise. There, nobody is using a drone because it’s cool. They’re using it because it does a job better than the alternative. All that matters is the job, and every step that stands between wanting the job done and having it done is friction that inhibits adoption. The perfect enterprise drone is a box with a red button. When you push the button, you get your data. Anything more complicated is a pain point to be eliminated. (And after that, we’ll get rid of the button, too.)
What that means is seamless integration between drones and enterprise software, such that all the data is automatically collected, sent to the cloud, analyzed, and displayed in useful form, ideally in near–real time.
What will this look like? Although it might surprise you, I hope the future of drones is boring. As the CEO of a drone company, I obviously stand to gain from the rise of drones, but I don’t see that happening if we are focused on the excitement of drones. The sign of a successful technology is not that it thrills but that it becomes essential and accepted, fading into the wallpaper of modernity. Electricity was once a magic trick, but now it is assumed. The internet is going the same way. My end goal is for drones to be thought of as just another unsexy industrial tool, like agricultural machinery or generators on construction sites — as obviously useful as they are unremarkable.
My inspiration in this is my grandfather, Fred Hauser, who in the 1930s invented the automatic sprinkler system (his patents decorate our walls). You may not think of a sprinkler system as a robot, but it is: Today’s are connected to the internet, collect data, operate autonomously, and, best of all, just work. Now imagine farm drones doing the same: boxes scattered around the farm with copters inside and solar cells outside, to recharge their batteries. Like the irrigation systems, at some point in the day they wake up, emerge from the boxes, and do their thing — crop mapping, pest spotting, or even fertilizing like bees. When they’re done, they return automatically to their boxes; the lids close, and they sleep until they do it all again the next day. All the farmer needs to know is that the daily crop report on his or her phone is extraordinarily detailed, with multispectral analysis of everything from disease to dampness, measured to the individual leaf and analyzed by machine-learning software to flag issues and make recommendations for the day’s work.
Drones as ubiquitous as sprinklers: We’ve come a long way from weapons, sci-fi movies, and headlines. But in the prosaic applications of advanced technologies lie their real impact. Once we find drones no longer novel enough to be worthy of HBR articles, my work will be done.The Big Idea
1 comments:
rent car usa, green motion miami international airport green motion miami international airport car rental salt lake city airport
Post a Comment