Droneland
A drone for every purpose is on the way, and Stanford is feeling the buzz.
BY MIKE ANTONUCCI
Your life is controlled by mortgage payments because you’ve gone all in on the American dream: a cozy home with a fastidiously gardened yard in a gloriously hospitable neighborhood. Unfortunately, your place turns out to be directly under an off-ramp of the drone superhighway. You can’t tend your roses without smelling the descending pizzas; monogrammed underwear for the guy down the block gets airdropped on your front lawn.
But then you hear the news of the day. That father and daughter who went missing on their backcountry hiking trip were found by a search-and-rescue drone operating with infrared technology in the dead of night. Successfully engineered for lengthy flight times, the robo-hero made a second trip to lower supplies before a wilderness assistance team even reached the scene.
Welcome to the prognostications of the drone revolution, a double-edged sword of progress and turmoil. Stanford professors with expertise in unmanned aerial systems envision a drone-powered future that is overwhelmingly positive. The upside, they believe, will be a cascade of innovation — scientific, commercial and cultural — that rivals some of the biggest shifts in technological history. But along the way there are likely to be so many safety, privacy and nuisance concerns that it may feel like Armageddon for social tranquility.
We’ve been told to expect flocks of automated parcel storks, with baskets landing on every doorstep. Amazon does not yet have regulatory approval to blot out the sun, but citizens have picked up the slack by making drones the latest gizmo craze, with hundreds of thousands bought during the 2015 holiday season alone. Or haven’t you yet seen one of those rotor-propelled “toys” at your local park or even hovering outside your bedroom window? There’s widespread anxiety about rogue drones imperiling passenger air traffic. At Stanford, precise rules have had to be fashioned to protect university grounds from becoming a flyover mecca.
It can be difficult to sort through the hubbub and speculation that obscure a long-term perspective. But Mykel Kochenderfer, an aeronautics and astronautics professor at the national forefront of anticollision engineering, senses what’s coming. “These small consumer drones,” he says, “remind me of the homebrew computer era of the 1970s that led to the personal computer revolution. To a large extent, the technology has been driven by researchers and hobbyists. Start-ups are popping up all around Silicon Valley, and larger corporations are taking notice.”
No one is being cute when they say Stanford is abuzz over drones; there’s a surge of interest among students of almost every discipline. The official student club has become yet another nexus for the university and Silicon Valley: It sometimes hosts speakers from the region’s germinal companies and serves as an incubator for interest beyond the classroom. Professors and graduate students are conducting a variety of drone-related research, some of which is being applied in government and industry projects. Perhaps most tellingly, Stanford’s 10-week skills-and-thrills course in designing, building and flying a drone — popular enough to often go simply by its catalog number, AA241X — is rooted in the kind of interdisciplinary collaboration that has become a touchstone of university activities. The knowledge students acquire can underpin careers that center not only on the many branches of engineering, but also on commerce, management or public administration.
“Unmanned aircraft systems,” notes Juan Alonso, an aero-astro professor who is one of the AA241X instructors, “are first and foremost systems — a combination of multiple physical and functional elements that must work together in order to accomplish a complex task. . . . We place particular emphasis on making sure the students tackle all of the elements in the system and gain experience at all stages of the design, prototyping and operation.”
Stanford’s influence on how drones are made, used and perceived may become one of the university’s signature contributions to social change. Graduates are pouring into drone start-ups, especially in the Valley. And some of these companies, or at least their business models, are bound to burgeon. The impact may transform agriculture, policing, news gathering, filmmaking and photography, construction, emergency services and global exploration. Along the way, controversies are going to heat up about intrusions and misuses, about camera-and-sensor spy drones lurking at schools or peeking at what you do at an automated teller machine. Stanford faculty and alums from all the scientific and mathematical disciplines, not to mention those who will help shape the legislative and regulatory territory, anticipate a world in which drones can help us find parking, examine bridge corrosion and obey virtual barriers — “geofences” that program boundaries — when nearing forbidden locations.
Perhaps the most intriguing factor is one emphasized by aero-astro professor Marco Pavone, who co-authored a recent article with its headline touting “Flying Smartphones.” Pavone, aero-astro professor Mac Schwager, ’00, and Ross Allen, PhD ’16, described aerial drones as on the way to changing consumer electronic technology with as much everyday impact as smartphones had on personal computing.
Seen in that context, there’s much less confusion with drones as war machines. But for many people, merely hearing the word “drone” instantly evokes missile strikes that incinerate their targets. Researchers and entrepreneurs who focus on mainstream applications sometimes recoil from the word, referring only to unmanned aerial systems and vehicles. Nonetheless, drones that gather defense intelligence or serve other military functions may become far more prevalent tools; they’re reportedly being tested for underwater missions. In fact, the definition of a drone is still flexible, with room for nonflying unmanned ships. Predictably there will be periodic ferment about threats from drone smuggling, drone vandalism and drone terrorism.
Pavone, Schwager and Allen envision aerial possibilities that range from the relatively mundane, such as monitoring freeway congestion, to the flamboyantly avant-garde, such as artistic or advertising displays formed by groups of drones radiating colored light. Why not a roadside billboard that you’re driving toward as it coalesces, fluttering into a floating, glowing message? You can hear the differing passenger reactions now: “Awesome” and “Keep your eyes on the road.”
Pavone specializes in research on the analysis, design and control of autonomous systems. For drones that largely means “endowing the robot with flexibility in its decision-making capabilities.” Figuring out how to make drones “smarter” is crucial, says Pavone, because it’s assumed they will encounter far more impediments, safety cues and shifting conditions than engineers can anticipate. In other words, drones must learn to account for the unforeseen or inconstant.
A video filmed in Pavone’s Autonomous Systems Laboratory demonstrates the issue about as entertainingly as possible: Allen, Pavone’s former PhD student, recorded himself “fencing” with a small drone in an indoor space that confined its range of motion. Allen would lunge with his blade, and the swordless quadcopter would maneuver and dodge. The drone was configured to make kinodynamic calculations — figuring out how to instantly react to a problem while also accounting for physical boundaries.
Allen’s primary research objective was to show how planning algorithms (computational procedures) could animate a high-speed robot. And Pavone had thrown down a gauntlet. “I think he was the one,” recalls Allen, “who first brought up the challenge of having the quadrotor avoid a person trying to interfere with it. I think he humorously phrased it as something like, ‘Swing a stick at it, and if it dodges you, I’ll give you your PhD.’”
As amusing as the video comes across, no one is going to suggest it’s the next Zorro. But it garners attention for good reason. Despite a host of developmental, business and governmental complexities, the prospect of experimenting or working with drones seems to have an especially adventurous allure. Even spasms of failure provide inspirational lore.
Consider Peter Blake, MSM ’14, who arrived at the Graduate School of Business with a background that included flying Harrier fighter jets in Iraq for the U.S. Marine Corps. Which didn’t fulfill any of the qualifications for AA241X. When Blake asked in, Alonso interrogated him: “I said, ‘Well, do you know anything about aerodynamics; do you know anything about controls; have you ever done any computation?’” The response, Alonso says, was “No, no, no. But I’m a pilot.”
Alonso decided to see what the teamwork would be like if Blake and another GSB student were planted in a thicket of engineers. Alonso remembers telling Blake, “Initially, your job is to be the pilot for the team, because you have to fly it around remote-controlled and then turn it on so it goes autonomous to see if it works, right?” And then? “He crashed the drone four times in his first five flights.”
Blake’s sense of humor was more durable than the drone, which he says took on a Frankenstein-created appearance from all its duct tape and epoxy repairs. Hard to believe he kept $30 million military planes in the air, he quips. Even so, the group dynamic was strong: In a class competition to perform a simulated search-and-rescue mission, Blake’s team won.
In terms of practical training and cross-disciplinary effort, Blake looks back on AA241X as “very representative” of what he went on to experience in the budding drone industry. After working as director of flight operations and client services for the San Francisco start-up Skycatch, he relocated to Denver and joined Aeryon U.S., a provider of small unmanned aerial systems, as director of worldwide client solutions (flight demonstrations, training and product implementation). He says drone businesses are still in a nascent phase, crafting their strategies. Stanford, he says, “is in a great position to influence how the industry is shaped from a technological standpoint.”
There are a number of academic and commercial hot spots for drones around the United States — Stanford faculty are quick to point out the nationally dispersed momentum in the field — but there may be an extra cachet to the theme of “Silicon Valley meets aviation.” That’s because, as Alonso explains, “It’s not just about some people building the airplanes or building the vehicles. It’s about the sensors; it’s about the vision; it’s about the decision-making; it’s about the data science; it’s about the onboard and off-board computing capabilities.” Stanford’s role as “catalyzer for the energy that’s in the Silicon Valley,” he says, makes it one of the world’s leading spots for the drone revolution.
Kochenderfer, ’03, MS ’03, is director of the Stanford Intelligent Systems Laboratory, whose core focus is on “decision-making under uncertainty.” Before returning to the university as a professor in 2013, he was at the Massachusetts Institute of Technology’s Lincoln Laboratory, where he instigated the ongoing development of a major advancement in an international collision avoidance system for manned aircraft. That work is now being adapted for drones, and his research interests include the computational methodologies for driverless cars.
“How do you make good decisions when you are uncertain about the current state of the world, how the world will evolve and when you have multiple competing objectives?” Kochenderfer’s students undertake that thinking. They go by a lab-based nickname that’s boringly spelled as SISLers and rousingly pronounced as “sizzlers.”
It’s that idea of competing interests that gives drones their image as a double-edged sword. In late spring, for example, Menlo Park made the news for having more commercial drones (176) registered with the Federal Aviation Administration than any other U.S. city. By late summer, the local headline was that drones had been banned at all city parks. Across the country, the media had a field day when a woman blasted a drone out of the air with a shotgun when it flew over the property of her celebrity neighbor, actor Robert Duvall. That happened in Virginia, which shortly thereafter got another kind of attention: Virginia Tech’s involvement with UAS research instigated an experiment in which burritos were delivered to campus for a few weeks by winches lowered from drones.
Kespry, a Menlo Park drone start-up that specializes in functions such as surveying and inspection for the construction and insurance industries, channels an inventive spirit into nitty-gritty business realism. “Drones Will Change Everything” is a prominent slogan on the company’s website, but business and policy vice president Gabriel Dobbs, JD ’14, MBA ’14, supplies the caveat: “Drones can do remarkable things but not everything people imagine they can do.”
Indeed, founder and chief executive Paul Doersch, ’10, notes that he trekked through a slew of vineyards and orchards before determining that agriculture had less immediate potential than other applications. (“Farmers don’t jump on new technology quickly.”) But even the firm’s hard-core geekiness has a poetic aura. In addition to offices, Kespry has a shipping container filled with machines that sit on shelves while they “think” they’re flying. “The drones are dreaming,” say the Kespry employees with sly smiles. It’s more like a guided meditation: The drones are connected to physics simulators that attempt to replicate the challenges of real-world navigation. The drones go everywhere in whirring (snoring) reverie.
Kespry’s day-to-day operations, which as of late August include a 2.0 version of the firm’s drone system, are headlined by proficiencies in flight time, wind resilience and data gathering. It’s also the kind of enterprise that provokes many of the questions about how drones will affect jobs, and in this case highlights the argument that aerial examination of, say, a sprawling quarry is a significant safety enhancement.
Doersch, who majored in computer science, retains a sense of “the big picture” — of what’s happening to increase, with incredible precision, our knowledge of our surroundings. “We scan the physical world into the cloud.”
Sounds socially and statistically gargantuan. Then add mushrooming recreational use and experimental initiatives and you get, yes, a regulatory muddlement. The fog at local, state and federal levels has been pervasive, but the last half year saw some significant clarification. As of late August, new commercial-scientific rules from the FAA eliminate the need to obtain special permissions before using drones for a wide variety of purposes that include education and research.
Details are copious, but the essential provisions are these: The unmanned aircraft must weigh less than 55 pounds, observe an altitude ceiling of 400 feet and a maximum speed of 100 miles per hour, not fly over anyone not participating in the operation, not fly at night, and stay within eyesight of the operator (precluding for now the beginning of grand-scale package bombardment). Another notable requisite is that the operator be at least 16 years old and have a remote pilot certificate with a small-UAS rating (or be under the supervision of a certificate holder).
Certain rules (such as the weight limit) overlap with those already in effect for recreational drone use, but consumers and hobbyists are far less regulated as a result of federal law enacted in 2012. Despite some controversy and puzzlement, restrictions such as an age minimum and training stipulations for operators don’t apply. How much this contributes to drone transgressions is an open question. The practicality of enforcing any rules anywhere is another enigma.
For a private entity such as Stanford, with an especially tempting expanse of flyable landscape, there are considerable dangers for its teeming populace and sprawling property from what amounts to helter-skelter drone trespass. Concerns include possible interference with the Life Flight helicopters bringing patients to Stanford’s two hospitals. Campus security has coped with hobbyists and tourists maneuvering drones over the Main Quad and around Hoover Tower. A crash of a remotely piloted vehicle also started a brush fire in the Lagunita lake bed in July 2015.
As a result, a policy implemented in August prohibits anyone who’s not part of the Stanford community from operating unmanned flying vehicles above or from within university lands. Moreover, faculty, students and staff need to obtain advance permission for vehicle flights at Stanford through a committee designed to streamline the approval process. Members of the student club — Stanford Unmanned Aerial Vehicle Enthusiasts, Engineers and Entrepreneurs (SUAVE) — may operate unmanned aircraft without committee review as a hobby activity in areas preapproved for the club. They must wear identifying vests.
The temptations of the Farm’s towered and mega-lawned terrain are just coincidental to everything else drone-related that’s happening on campus. But if we’re going to be a drone nation, Stanford and Silicon Valley will be among the founding colonies. The proliferation of drones is upon us; the ramifications are still in flux.
Mechanical engineering and computer science grad Hao Yi Ong, ’15, MS ’15, worked under the supervision of Kochenderfer to devise a conflict avoidance system that’s among the concepts being evaluated by NASA, which partners with the FAA on air traffic management research. Collision prevention is crucial to bringing widespread drone use to fruition, and Ong has been building on Kochenderfer’s earlier groundbreaking work on airborne collision avoidance.
Stanford voices will also join the wide-ranging policy discussions that have to keep up with the rapid evolution of drone advancement and activity. A notable recent example: The FAA named Alonso to a 35-person Drone Advisory Committee that draws on an eclectic mix of industry, governmental and academic leaders with an assortment of Valley connections.
Media coverage of drones is mixed, given their privacy and safety implications, not to mention the potential vexation quotient. But Alonso preaches calm and education. Yes, he says, many bedeviling issues are involved. “But they’re going to get solved. I leave you with a message of hope. I think the benefits are going to largely outweigh the drawbacks, and the drawbacks are going to be managed.”
Will drones walk our dogs? Apparently they have already been used to scan for sharks at beachfronts. As with the personal computer revolution, we’re entering an era that portends at least as many surprises as sure things. Indeed, Alonso takes delight in pondering the possibilities. Colleagues in the computer science department, he notes, would like to program drones to serve as campus tour guides. So he’s wondering exactly how that would work. “Does it fly backwards as it talks?” •
