Invasion Of The Robots

by Michael Kanellos – Staff Writer, CNET
March 10, 4:00 AM PT
The robots are coming. And when they get here, they will take out the trash. Mobile, intelligent robots that can perform tasks usually reserved for humans are starting to creep into mainstream society and could become a multibillion-dollar market in a few years.

iRobot says it has sold hundreds of thousands of units of the Roomba, a self-guided, self-propelled vacuum cleaner that sells for around $200, in just one year.

Other inventors are eyeing the health care market. Joe Engelberger, widely known as the father of robotics, is trying to get funding to build robots that will dress, cook for and generally take care of senior citizens. Home health care robots are being tested in Japan, while U.S. hospitals are already using machines to deliver charts, carry medicines or even assist in surgery.

“Nursing homes or live-in help is expensive, and you have personality conflicts,” Engelberger said. “The technology is available. It takes very good engineering, but it does not take invention.”

Another potentially large market exists in creating machines that can operate in hazardous or extreme environments. Workhorse Technologies, founded by Carnegie Mellon University professor William “Red” Whittaker, is working on robots that can map mine shafts. He came up with the idea after Pennsylvania’s Quecreek mine accident in 2002, in which nine miners were trapped for four days as the result of faulty maps.

Whittaker has also created robots that can drive harvesters, clear out sewer lines and clean up nuclear waste.

“When you are in those commodity enterprises and you can introduce efficiency, it does not matter whether it is carrying stuff out of a surface mine for less dollars per ton per mile or whether it is a greater productivity out of the machine in harvest,” Whittaker said. “Any small differences mean a lot, and these technologies can mean larger differences.”

Carnegie Mellon’s technological prowess in this area will be tested this weekend in the DARPA Grand Challenge, when driverless, robotic cars will race from Los Angeles to Las Vegas for a $1 million prize. The university’s Red Team Racing is the favorite. Other contestants include academics from the California Institute of Technology, a team of brothers from upstate New York and a group of students from Palos Verdes High School near Los Angeles.

Evolving automatons
The surge in robot activity is at least partly the result of steady improvements in performance and steadily dropping costs for processors, sensors, navigation software and the other technologies required to put a mobile robot together.

On the performance side, for instance, Seegrid, co-founded by Carnegie Mellon professor Hans Moravec, has developed software that allows a mobile device to create a dense 3D map of a hallway or room after a single pass. Global positioning systems, too, can pinpoint a robot or any other object anywhere on the globe within 10 centimeters, and on-board processors, which crunch sensor data and coordinate a robot’s movements, also continue to increase in performance.

And parts are getting much cheaper. Although robot researchers had to develop their own components in the past, today’s business is attracting chip suppliers and software developers such as Intel, Microsoft and Texas Instruments, as well as a number of start-ups.

Likewise, robot makers are working with other industries to lower manufacturing costs. The basis of the Roomba came out of robotic cleaning machines developed for Johnson Wax, and iRobot worked with Hasbro for three and a half years to learn low-cost manufacturing of toys.

Just as important as performance and costs, from a sales perspective, is customer satisfaction. Robot developers have adjusted their products to meet practical customer needs rather than simply using the machines to showcase a company’s technological abilities or as entertainment devices.

“The industry has disappointed for 40 years,” said Colin Angle, CEO of iRobot. “Ever since Rosie from ‘The Jetsons,’ robots have been the next big thing, but the business case was never there. It is easy to build a robot that is prohibitively expensive.”

iRobot’s PackBot is designed with practicality in mind. It’s a 40-pound battlefield robot designed to perform reconnaissance missions, help wounded soldiers, find chemical leaks and deliver equipment, all over rough, unpredictable terrain.

“One took a 25-foot fall and righted itself,” during an operation recently in Afghanistan, Angle said. “It is an incredibly mobile 800MHz Linux machine.”

PackBots aren’t cheap: They cost between $50,000 and $100,000 apiece. But their main customer–the U.S. government–seems happy. Besides deploying PackBots in Afghanistan and Iraq, the Army has given iRobot $25 million to develop automated battlefield vehicles.

Another recent success is the Predator, an unmanned plane used to attack al-Qaida in Yemen and other locations. The Department of Homeland Security is expected to issue several large research grants to this project in coming years.

“The sky is the limit, but it depends on the robotics entrepreneurs taking a much more customer-oriented approach,” said Kishore Rao, a senior associate at investment firm Trident Capital. The company put some of its capital behind iRobot but acknowledges that the number of VC-backed robot firms “can be counted on one hand.”

Bio of the ‘bot
The idea of automatons that can perform various tasks has been around since ancient Egypt. The word “robot,” however, is of relatively recent vintage, coined by Czech playwright Karel Capek in the 1921 play “R.U.R.”

The first commercial robots appeared in the early 1960s as the world was consumed with the Atomic Age science of the Cold War. Unimation, founded by Engelberger, created robotic manufacturing arms, while Barrett Electronics came out with a driverless electric cart for grocery warehouses that was navigated by signal-emitting wires in the floor.

It wasn’t an easy sell. “I had to go to 46 different companies, including GM and IBM, before I got some money from a railroad company,” Engelberger recalled. Although GM passed the first time, it eventually became the first company to install a Unimation machine.

Although robotic start-ups have spun out of CalTech and the Massachusetts Institute of Technology, one of the largest concentrations is in Pittsburgh. Since Westinghouse funded the development of the robotics program at Carnegie Mellon 25 years ago, roughly 30 robotics companies have set up shop in the region, according to Bill Thomasmeyer, president of the Robotics Foundry, which is trying to establish a robot economy in Western Pennsylvania.

Among the regional residents is Bombardier Transportation, which makes the people movers in airports. “They are looking to expand to downtown transportation,” Thomasmeyer said, an effort that will involve technologies to avoid inevitable obstacles.

Robots still are not a force in the global economy. In all, North American robotics manufacturers ship about $1 billion worth of products a year, according to Robotic Industries Association spokesman Jeff Burnstein. Other statistics show that the international market approaches $5 billion.

The vast majority of the revenue derives from limited-function devices used in packing or manufacturing plants, but versatile robots have entered the market in the last few years.

The da Vinci Surgical System from Intuitive Surgical, for example, is a set of precision robotic hands for doctors. The system, which costs a little more than $1 million, can drill through bone or make incisions.

Surgery is made far less intrusive because the machine can accurately control its movements and “see” in 3D through graphical images. That, in turn, makes recuperation quicker, less painful and cheaper, according to the company. Around 192 da Vinci systems are in use at hospitals to date.

The market for personal and mobile robots could grow to $5.4 billion this year and become larger than the industrial, nonmobile robot market, according to Dan Kara, president of Robotics Trends, which holds conferences and promotes the industry. By 2010, that figure will approach $17 billion, Kara said.

While some say such projections are overly optimistic, Kara can point to anecdotal evidence that robot fever is catching. About 68,000 attended Robodex in Japan last year, he said, and the sales numbers of Roomba and da Vinci are tough to argue with.

Similarities to PC industry
“If we were sitting here 10 years ago, I think no one would have really had a clear sense of what is the content of an intelligent robot,” said Whittaker, who believes that the economics of the robot market could function a lot like those of the PC industry in 20 or so years.

For the business to reach a full-fledged boom, however, some key problems must be solved. Communication and coordination remains a challenge, and such issues are being tackled by several scientists working on systems patterned after the swarming of insects.

Grippers also need work. Mobile robots like the PackBot or like those from Workhorse are good at taking pictures or delivering items, but picking things up is another matter.

Still, robot capabilities are growing rapidly. A 1,500-gram human brain can churn at about 100 trillion instructions a second, according to a paper recently published by Carnegie Mellon’s Moravec–nearly three times the power of the Earth Simulator, the world’s most powerful computer. Under Moore’s Law, processing performance could increase to the point where machines can work almost as well as the human brain.

“Better yet, sufficiently useful robots don’t need full human-scale brainpower,” Moravec wrote in his paper. “Commercial and research experience convinces me that mental power like that of a small guppy, about 1,000 MIPS, will suffice to guide mobile utility robots reliably through unfamiliar surroundings, suiting them for jobs in hundreds of thousands of industrial locations and eventually hundreds of millions of homes.”

The human brain aside, anthropomorphism will grab a piece of the market. Sony, Honda and some other Japanese companies are marketing robot companions, and a $99 “RoboSapien” will be coming soon from tech-toy company Wow Wee.

Mimicking human behavior in most robots remains complex, however, and the demand remains low. “You can get a robot to speak with you, but you can’t get it to talk Spinoza with you,” Engelberger said.

Another absent human characteristic is more obvious: legs. While some experimental robots have legs that mimic how geckos or crabs walk, wheels can work as well in almost every environment and are a lot easier to engineer.

“Legs in my mind are for Hollywood,” Angle said.

The original article can be found at: