As you read this, an Italian van is driving itself to Shanghai. Thousands of robots are gliding through hospital hallways, standing in for doctors and toting supplies. Seven thousand unmanned planes are patrolling Iraq and Afghanistan, and more than 3,000 robo-sherpas are carrying heavy loads for U.S. soldiers. From Pittsburgh to Tokyo, senior citizens are therapeutically cuddling $6,000 furry “Paro” devices that wriggle and trill in response, 200 humans in Denmark are “certified robot therapists,” and 150,000 robot lawn mowers are clipping grass from St. Andrews, Scotland to the Great Lawn in Central Park.
Robots are now reaching sufficient numbers in some areas to begin changing the way many of us live and work. There’s no doubt that robots can do the heavy lifting for heavy industry — they can build cars, cap deep water oil wells, even explore Mars. But the real growth in robotics will be in the consumer market. Service robots — the artificially intelligent machines that can clean your house, mow your lawn, and teach your children in a classroom setting — are the next challenge and opportunity for the economy.
As we begin to reinvest in the infrastructure, education and health of the nation, where do robots fit into the Great Economic Recovery ? Will they help create jobs? Or eliminate them? Will they help us innovate and automate, increasing productivity on a par with China and Japan?
The White House recently called robotics a national priority: “Tremendous progress in robotics technology development over the past five to 10 years has unleashed new opportunities for automating tasks and enriching the lives of humans. Robotics technology is reaching a ‘tipping point’ and is poised for explosive growth.”
In this two-part series, we’ll look at how robots might serve the needs of business, government, the military and society, and whether or not the United States workforce is ready and able to compete globally in this industry that demands skills in math and science. Welcome to “the knowledge economy.”
Part 1 — The Robots among Us
In a Newville, Pa. warehouse, the inventory is really moving. Early this year, Office Depot opened its new 600,000 square-foot eastern distribution center, built around a fleet of 300 squat, bright orange robots that speed up every aspect of handling goods for the company’s stores and online retail business. Instead of human “pickers” logging miles a day on foot fetching items from warehouse aisles, the ottoman-sized bots made by Kiva Systems slide underneath a metal shelving unit, lift the entire load, and carry the inventory to the people packing orders.
Whirring around the warehouse floor, the Kiva Systems robots avoid collisions, receive commands and communicate with each other wirelessly, forming an orderly line as they wait to deliver their cargo. As the bots return shelves to the warehouse floor, they continually rearrange the shelving layout to increase efficiency. Fulfillment time for an average order is down from two hours in the company’s old warehouses to about 20 minutes in the new one.
The robots are not cheap. The smallest Kiva installation — bots, shelves, and the central software to control the system — runs about $1.5 million, according to the company’s marketing manager, Mitch Rosenberg. But the company has sold or leased 4,000 robots, and based on clients’ experiences, Kivas promise at least two-to-four-fold efficiency increases. Productivity goes up and costs go down as robots displace some workers. Office Depot consolidated several older distribution centers into the new one with fewer employees. Workers who remain, says Rosenberg, have safer and more enjoyable jobs because the robots are doing the most backbreaking and mind-numbing work.
They may have anywhere from 15 to 20 robots actually picking
products [for them]. Our efficiency went up by factors of five or six.”
The Kiva robots installed in 2008 at the online retailer Zappos’ Kentucky distribution center helped to knock order turnaround time from a minimum of 48 minutes using an old conveyor belt system down to an average of 12 minutes. In the past two years, more big-name retailers have also adopted Kiva warehouse bots, including Crate & Barrel, Gap Direct, Saks.com and Staples.com.
A new spinoff this summer from Diapers.com called Soap.com was one of the largest launches in e-commerce history, with $30 million in capitalization. Scott Hilton, executive vice president of parent company Quidisi, has hinted that neither site could be competitive offering overnight delivery of basic corner-drugstore goods without the robots: “The productivity of any individual worker is incredible. They may have anywhere from 15 to 20 robots actually picking products [for them]. Our efficiency went up by factors of five or six,” he says in a company video.
The orange cuboid Kiva bots may not look like most people’s fantasy of a robot, but they are the face of a new generation of so-called service robots that are on the job now. These robots are smaller and smarter than the old industrial behemoths. And the growing field of artificial intelligence — a spinoff of computer science — could give them the brains they need to do even more.
When production slows, as it did in Japan in 2009, the robots can also be laid off without collecting union pensions, unemployment or other benefits.
Here, There, Everywhere
Nonindustrial robots for professional applications number about 80,000 worldwide, accounting for $13.2 billion in sales through the end of 2009; 8.6 million robots for personal entertainment and domestic work are also now in the hands of consumers, representing $2.8 billion in sales. Both categories are projected to double in size by 2013.
Compact, mobile and increasingly agile, service robots include cleaning and lawn-mowing robots, military pack-carrying and bomb-detonating bots, search-and-rescue and surveillance bots, the underwater, remotely operated vehicles used to manipulate BP’s broken oil well head, as well as those used by scientists to detect underwater oil plumes. The category covers any robot that performs a service rather than being directly involved in manufacturing products like industrial robots do.
explicitly designed for interacting with humans — are
the next challenge and opportunity for the economy.
Those giant swinging industrial arms that weld, paint and assemble cars and other heavy equipment have been used for decades — General Motors installed the first one in 1961, and about a million of them are at work around the world right now. The U.S. military has also been making heavy use of remotely operated drone aircraft and rugged ground robots for at least nine years. Many of the military pack bots are made by iRobot, the company better known for its iconic Roomba vacuum that debuted in 2002.
For a long time, Roombas were the only robots an average American was likely to encounter, but new kinds of robots, including “social robots” — the artificially intelligent machines explicitly designed for interacting with humans — and sturdy task-oriented robots, which are steadily penetrating mainstream realms like hospitals and warehouses, are the next challenge and opportunity for the economy.
Aethon TUG bots have become beloved workhorses in more than 100 U.S. hospitals. Summoned wirelessly, or preprogrammed with a route, these diminutive cousins to the Kivas tote carts full of supplies, food trays, medicines or dirty laundry around the buildings, which they navigate using an internal map. “Remote Presence” robots by InTouch Health are another big hit in health care settings. The company’s 5-foot flagship robot, RP-7, serves as a two-way video conferencing system on wheels, and is the only FDA-approved device for performing remote medical examinations. It glides down hospital hallways under the remote user’s control through a dedicated wi-fi network.
By connecting stethoscopes, ultrasound equipment and other devices to the machine and having someone in the patient’s room operate them, a telepresent doctor can look on and talk with the patient while monitoring the data onscreen. The company says that as of this year, more than 250 health care facilities have installed its RP series of telepresence robots.
orderlies, technicians and other workers in the newly built
hospitals never opened up because robots can do them more cheaply.
With chronic shortages of nurses and doctors, hospitals have come to see robots as welcome support, and a handful of newly constructed facilities literally built robots into their daily operations. The Forth Valley Royal Hospital in Scotland just opened in August with dedicated elevators and underground hallways for its fleet of robots that will carry supplies, take out trash, clean operating rooms and even dispense drugs in the pharmacy. Silicon Valley’s high-tech El Camino Hospital opened last year with a fleet of 19 TUGS to make deliveries, and a robotic laboratory system that performs tests automatically.
To date, no one seems to have been fired from an existing hospital to make way for robots, but there’s little doubt jobs that might have gone to orderlies, technicians and other workers in the newly built hospitals never opened up because robots can do them more cheaply.
El Camino’s larger five-story building replaced a 1960s-era high-rise that used dumbwaiters to move food and supplies between floors, says Ken King, head of administrative services for the hospital. He calculated that the new “more horizontal” facility would require at least 12 full-time staff to make deliveries at a minimum cost of $800,000 a year. So he decided instead to lease the TUGs that work 24 hours a day, 7 days a week, for $460,000, including all maintenance and monitoring.
“Conservatively, the cost of a robot is half of a person,” says King. Hospitals like his are embracing robots because, “We’re constantly being pushed to find ways to do things with higher quality and lower cost,” he explains. “Pushing carts around is something that has to be done, but it’s costly. You want to be able to spend your labor dollars with caregivers at the bedside.”
The Human Touch
Although relatively simple service robots are becoming commonplace in hospitals, the robot helpers that could provide home care and companionship for the disabled and elderly are the “killer app” manufacturers are vying to develop. With hefty government funding, Japanese roboticists have been almost exclusively focused on creating caretaker bots for nearly two decades because of a looming demographic crisis in that country that will soon produce more elderly retirees than able-bodied younger people to look after them.
South Korea started throwing similar support to its robotics researchers 10 years ago. So far, the efforts in Asia have yielded some extraordinary prototypes that can perform tasks from loading and unloading a dishwasher, to cooking and serving noodles, and yes, pushing a regular vacuum cleaner. No single machine can do a multitude of household chores, however, and most developers of these prototypes are academic researchers still seeking partners to commercialize the technologies.
is between humanoids — think Star Wars’ C-3PO
with a head and functioning limbs — and everything else.
The majority of professional and personal service robots working today — whether tugs, telepresence bots, or Roombas — highlight what is perhaps a more important distinction in the field. That is, between humanoids – think Star Wars’ C-3PO with a head and functioning limbs — and everything else. Making robots to perform specific tasks is difficult enough, but building a robot that can also walk and handle objects like a human being adds many more layers of technical complexity. Honda has spent nearly 25 years and reportedly upwards of $200 million to develop Asimo , a 4-foot-tall humanoid that may be the Gene Kelly of legged robots. When preprogrammed for demonstrations, it can carry a tray and serve tea, dance the cha-cha, hold a brief conversation and navigate around obstacles in an unfamiliar environment.
Asimo is nevertheless something of a scarecrow still waiting for a brain. The Honda Research Institute is now working with partners on improving the kinds of capabilities any working robot needs in order to be useful: better sensors to detect conditions and objects in its environment and the artificial intelligence (AI) to “understand” and respond autonomously to what it’s seeing.
The $18,000 Nao is one of only a few humanoid robots that are actually for sale today — not to individuals just yet, but to researchers who want to use it for developing their own robotic software applications (though its manufacturer, Aldebaran Robotics, does plan to launch a personal “companion” version of Nao by the end of 2011). But Naos have already been making news this year, with demonstrations of capabilities like emotion-recognition and ethical behavior that will distinguish the next generation of humanoid social robots from the practical service bots on the job now.
Today, robots are still more tools than teachers, nurses, doctors or housekeepers. They are for the most part incapable of completely supplanting a human being. But as they get smarter, more dexterous, and more responsive, the foundations for sophisticated, multitasking robot workers are coming together.
So far, many robots are finding their greatest success in areas where statistics suggest chronic labor shortages prevail — supplementing thinly stretched workers, such as in health care and soon, agriculture. But they are also changing the nature of some jobs, and those in easily automated jobs like equipment monitoring, product assembly, and internal mail delivery could have cause for concern.
On the other hand, an investment in robotics will foster new industries and create new jobs requiring innovative workers skilled in computer science and engineering, sales and marketing. Most important will be workers who can learn to work side by side with robots. The promise of robots has always been that they will do the “dirty, dangerous and dull” work, freeing humans to do the fun stuff.
Tomorrow, we’ll look at what’s driving the robot economy and how robots may influence the larger economy.
Christine Soares is a New York-based journalist and a former editor of Scientific American.