KITTY HAWK, FLYING CARS, AND THE CHALLENGES OF ‘GOING 3D’
TO GET A peek at the future of transportation, you could start by visiting a warehouse building in an industrial stretch of Mountain View, California. Above the door, which has its window papered over, a safety sign catalogs various health and fire hazards that lurk within. A red and black wet suit hangs in a tree nearby. Alex Roetter strolls over from another building, scans his security badge, and leads me inside. Roetter, who is a division president at this aviation startup, Kitty Hawk, steps into a vast room with cement floors. There, arrayed in a U shape, are 13 or so Flyers, an oddball aircraft that few people have seen and even fewer have piloted.
“It's the kind of vehicle that anyone can learn to fly in 15 minutes,” Roetter says. “The computer does all the hard work, so the human is just left to do the things that people are really good at. Look out the window, decide where you want to go, and just point the stick to where you want to go, and land.”
The Flyer is an airborne trimaran in gleaming white. The middle pod, where the pilot sits, resembles a Formula 1 car cockpit. It's flanked by a pair of pontoons, so it can land on water as well as earth, with two beams protruding from their sides. With one seat and 10 propellers, it's 13 feet long, 7.5 feet wide, and 5 feet tall. Thanks to carbon fiber construction, the aircraft weighs in at only 250 pounds. Powered by electricity, it is remarkably quiet. It takes off, lands, and flies pretty much like a helicopter.
Actually, this is the second generation of the Flyer; the first, which Kitty Hawk showed off over San Francisco Bay in June 2017, looked like something a comic book villain might fly, with a seat you straddled. The shift to the protective cockpit, though, pales in importance compared with the transformation in the vehicle's onboard computers. Learning to fly the first version took several days: five hours in a simulator, a day of training, a series of flights while tethered to the ground. The new aircraft lets people with zero aviation experience take off after that 15-minute lesson, thanks to smarter software.
In a helicopter, the pilot works four controls at once, while monitoring how each impacts the others. In the Flyer, the pilot's left hand works a thumbwheel to go up or down. The right handles a joystick to send the aircraft forward, back, left, right, and around. Let go of the controls and the Flyer holds itself level and in place, like a ship at anchor. That's it. The digital 0s and 1s translate the human pilot's basic commands into aeronautic expertise. The computer sets rotor pitch and speed while using GPS, an inertial measurement unit, lidar laser scanning, and radar to ascertain its position in space. It's as easy as flying a quadcopter drone, except you are inside it.
Roetter has taken about half a dozen low-speed flights in the Flyer over water. He's more engineer than salesman—a software engineer, to be specific. The bullet point on his CV that won him this job was developing an advertising software tool at Twitter that generated billions in revenue. But he's also a licensed pilot, and he lights up while talking about the ride: “It's like being a kid.”
The Kitty Hawk model we're looking at is recreational, but what makes it more than a lark for thrill-seekers is that its creators see it as a first step toward something momentous—the flying car. “Our long-term vision is to free the world from traffic,” Roetter says. The idea is to create a vehicle that can soar over congestion, making short, driving-distance, point-to-point hops—all without fossil fuel emissions. That craft is “probably not the Flyer,” he continues, “the same way that the Wright Flyer was not the vehicle that shrunk the Atlantic Ocean.”
Orville and Wilbur Wright's airplane, which took off from the sand dunes of Kitty Hawk, North Carolina, in 1903, was not sold commercially. It was 11 years before the business of passenger aviation got started, and another 25 before Pan Am inaugurated the first transatlantic service. “Sometimes you do things in the beginning because they're the first thing,” Roetter says. “And you learn and grow from there.”
Indeed, while Roetter's team works on the Flyer in the Bay Area, another division of Kitty Hawk is in New Zealand, refining and testing the company's second craft, the Cora. With room for two in its teardrop-shaped cabin, the 12-rotor electric aircraft can take off, fly, and land all by itself.
Kitty Hawk is far from the only enterprise with ambitions to reshape short-distance travel. Some two dozen companies—from mammoths like Boeing and Airbus to Volocopterin Germany and EHang in China, along with a raft of startups—are riding a jet stream of technological advances to create smaller, battery-powered flying vehicles. Lithium-ion batteries have gotten far cheaper and more energy-dense. Distributed propulsion, where a single power source supplies a bunch of small rotors instead of one big one, has enabled designs that are more efficient than helicopters. Boeing and Airbus convinced regulators that lightweight composite materials can withstand the rigors of commercial flight. Consumer drones proved that software could make controlling a complex vehicle as simple as working a joystick.
As Roetter walks through the warehouse, he points to a cracked pontoon on the airframe that was sacrificed for crash testing and flicks a rotor as he explains that on the Flyer, five propellers spin clockwise, five counterclockwise. At one end of the room a worker with a blowtorch applies a vinyl coating to one Flyer; at the other end, a small crane sits ready to hoist a finished craft into an aboveground pool, to verify that it can land on water without flooding. Other Flyers are waiting to have their battery packs installed in the pontoons or for the flight-control computers to be loaded into the hollow space behind the pilot's seat.
The Flyers that Kitty Hawk is building now are mostly heading to Nevada's Lake Las Vegas for flight testing, as the company continues to refine that all-important software. Crash testing is done via remote control, but if someone's in the thing, Roetter won't let them fly more than 10 feet up or faster than 20 mph.
These are early days.
THE 21ST-CENTURY FLYING-CAR industry took off in October 2016, when Uber announced it was working on an air taxi service called Elevate. The ride-hail giant wouldn't design or build vehicles itself. Instead, it would contract with manufacturers and help coordinate efforts among private and public players to work out regulation, build infrastructure, and develop an air-traffic management system. Once that's all settled and companies start delivering aircraft, Uber would pull them into an urban “on-demand aviation” service. Five manufacturers, including Boeing subsidiary Aurora Flight Sciences, helicopter maker Bell, and Brazil's EmbraerX, have signed on to produce aircraft. Uber might buy and operate the vehicles itself or work with a company that owns them. The aircraft might fly themselves or have pilots.
Uber wants to launch a few flying cars for demonstration flights in Los Angeles and Dallas-Fort Worth as soon as next year, with a proper commercial service in both metros by 2023. The idea is to move Uber's ease of use to the sky. You'd pull up an app that would connect you to a vehicle sitting on a helipad. The sweet spot would be routes that might take one or two hours to drive—San Francisco to San Jose, say—but 15 minutes to fly. Morgan Stanley predicts the market for these short-hop electric aircraft could be worth $1.5 trillion by 2040. Airbus wants to produce a “productizable” demo model next year. Boeing CEO Dennis Muilenburg says flying cars will be real in the next five years. Yet for all the technological progress, actually delivering this dream safely to the skies will require a lot more than optimistic talking points.
ON THE DAY I make the long, traffic-clogged drive from Berkeley to Mountain View to meet Sebastian Thrun, Kitty Hawk's CEO, it's supposed to rain. Thrun meets me in the lobby, says hello, and notices my umbrella. He asks me if it's raining. I tell him no, but that the forecast said it might. “Ah,” he says with a grin. “You are a pessimist.”
Slender, with a shaved head and wearing a suit just because he felt like it today, Thrun looks the part of a 21st-century Captain Nemo. At a time of backlash against Big Tech, he bounces along with untroubled enthusiasm, a blue-eyed riposte to Peter Thiel's 2011 quip mocking Silicon Valley: “We wanted flying cars, instead we got 140 characters.” By the time the German computer scientist was 36, he had tenure at Stanford University and was running its AI lab. In the mid-2000s he started working at Google, helping to create Street View and running Ground Truth, the massive effort that underpins Google Maps. Thrun launched the company's self-driving car project as well as Google X, its “moon-shot factory.” So it's no surprise that despite Thrun's lack of aviation experience, his frequent patron, Google's Larry Page, tapped him to run Kitty Hawk. (The company was founded in 2010 by Stanford aerodynamicist Ilan Kroo, who created an early version of what's now the Cora aircraft. Page became the main funder, put Thrun in charge of the effort in 2016, and named it Kitty Hawk.)
Thrun believes his latest venture can “free the world from traffic,” but he doesn't overestimate where things stand now. “We're still in the infancy,” he says, “still at the very beginning.”
The first problems Kitty Hawk, Uber, and their ilk must stare down are technical. Take propulsion: Battery tech is getting cheaper and more competitive with gasoline, but it doesn't provide anywhere near the energy density of jet fuel. For every pound of battery, you get 0.1 to 0.15 horsepower-hours. For liquid fuel, it's 7.3 horsepower-hours, a 50-fold advantage.
Electric motors are more efficient than combustion engines, but not by enough to close that gap. And the work of building a machine that defies gravity means weight matters more than any other consideration, like the price of materials. “In cars, the ranking is cost, volume, weight,” says Richard Anderson, director of the Eagle Flight Research Center at Embry-Riddle Aeronautical University. “In airplanes, it's weight, weight, weight, volume, cost.”
Reducing noise—a necessity for making flying cars ubiquitous—exacerbates the problem. You can quiet a flying car by slowing the rotors and changing their angle to produce more lift, but that requires more torque and more power. Basically, the less noisy an aircraft is, the more energy it demands.
Physics and chemistry can be brought to bear; battery power is getting better. The bigger obstacle (for flying-car fans, anyway) involves regulation. Self-driving-cardevelopers have been able to charge ahead because they've found a hole in the regulatory regime; most states don't explicitly ban autonomous driving. (Congress has been kicking around legislation for nearly two years.) The sky is far more tightly controlled. The Federal Aviation Administration has to approve any new aircraft and does not easily make changes—let alone usher in a new form of urban flight.
First comes safety certification. For most aircraft, it takes years of testing—and the agency has never certified an electric aircraft for commercial use, let alone one that toggles between vertical and horizontal travel. It's not even clear how the FAA would classify something like Kitty Hawk's Cora, or if it would require an entirely new classification.
The software taking all the work away from the human pilot brings a different problem. Historically, the FAA has allowed only deterministic software, programs that produce the same results from the same inputs. To certify that code, Anderson of Embry-Riddle says, “you have to test all the possible inputs that can go in and show that any output that comes out is not detrimental to the vehicle.” But software that makes decisions about rotor speeds, aircraft angles, and what's an obstacle can't be tested that way. It's too complex.
“Flying cars aren’t like smartphones; you can’t let competing tech and protocols coexist while the market figures it out.”
Before flying cars can take off, the FAA must adopt a methodology that allows for a mathematical proof of safety. “This has been accepted in automotive stuff for years, but the safety aspects of airplanes have made our folks drag their feet,” Anderson says of the FAA. The agency sees significant potential in electric and autonomous aircraft, an FAA spokesperson says, “but the operational rules for a self-flying or autonomous aircraft are not in place, and this type of operation has not been tested.”
Nonetheless, Thrun and others have been encouraged by the FAA's recent efforts to allow for new types of aviation. In 2017 the agency dumped highly prescriptive rules for small planes in favor of performance-based standards. Essentially, it demanded that aircraft be safe instead of specifying how to make them so. And it has steadily relaxed its restrictions on drones. But some worry that this slackening comes with too much risk. “It's not clear the FAA has the staffing or the expertise to independently confirm that the companies who say ‘trust us’ for the unmanned industry have really done due diligence to earn that,” says Ella Atkins, who directs the University of Michigan's Autonomous Aerospace Systems Lab. And the recent crashes of two Boeing 737 MAX 8 jets, which killed 346 people between them, have raised questions about whether the regulator has become lax in how it lets companies certify new aircraft, especially aircraft that are so reliant on complex software.
Making sure flying cars won't crash and kill their occupants—and people on the ground—is not the only serious safety question. More vexing, perhaps, may be what happens when they fill the skies and threaten to crash into each other, or into buildings, birds, cell towers, and power lines. Flying cars, whether piloted by humans or software, will need far more sophisticated systems than anything currently guiding commercial jets. Alphabet's drone delivery company, Wing, has built its own air-traffic management system that assigns craft specific corridors of air space, and it has made parts of it open source in the hope that the industry will adopt the scheme. But flying cars aren't like smartphones; you can't let competing tech and protocols coexist while the market figures it out. Flying cars would require a single operating system—and therefore either a whole lot of cooperation between competing companies or a firm grip by the iron hand of regulators.
Solve all of that (piece of cake!) and the operational challenges of a flying-car service come into focus: What routes to fly, how much to charge, how to schedule maintenance without losing capacity. And what happens when someone gets sick and doesn't grab the barf bag in time. It's the kind of stuff airlines are used to dealing with, while Uber and upstarts like Kitty Hawk have to begin at square one.
IN THE ENTRYWAY of Kitty Hawk's Mountain View headquarters, a sign hangs above a bike rack, laying out the company's principles: “We derive urgency from a new vision for transportation, one that is free of traffic accidents and long rush hour delays.” Everyone working on this preaches the virtue of “going 3D,” obliterating bottlenecks by using the miles of air above our heads. And it's true that it's hard to imagine getting stuck in a jam when you have access to the whole sky. But we've been fooled before.
At the 1939 New York World's Fair, General Motors put on an exhibit called Futurama, designed to extol the virtues of driving. On a 17-minute ride in blue mohair armchairs, visitors took in a massive diorama of the American landscape, striated with ribbons of concrete. It was a vision of a national network of multilane interstates, full of cars but free of traffic. Less than two decades later, President Eisenhower signed the Federal-Aid Highway Act into law, paving the way for a 41,000-mile network of interstates—on which the average modern American driver gets stuck for nearly 100 hours a year.
Sebastian Thrun, though, isn't worried about his company's vision or the difficulty of realizing it. For a man who has taught cars to drive themselves and helped Google chart the world, taking flight is not just the exciting answer, it's the obvious one—what progress looks like. “The first cars were like tricycles,” he says. “When it rained, you got wet.” After he teases me about my umbrella, we walk outside. The skies are clear. Not a drop of rain lands on our heads.