Prototype social robotic systems are getting better and better, and the definition of “social robots” likewise continues to evolve. Early definitions described them as “robots that communicate and interact directly with humans, using language or gestures.” Today, some definitions attach the word “autonomous,” raising the bar higher. And at the MIT Media Lab, where a broad range of human-machine research is performed, the abbreviation MDS is used, for “mobile, dexterous, and social,” raising the bar yet again. Still, most wish lists from future robot owners specify even more: mobile, dexterous, social, autonomous, and humanoid.
Next year is the year of the robot, just as was promised five, 10, 15, and even 20 years ago. To date, the only successful consumer robot is iRobot’s Roomba, which may be considered a social robot by those owners who name and talk to their units. Since some surveys find two-thirds of owners name their Roombas, the pent-up demand for an advanced social robot may be considerable, emphasizing the failure of such robots to date.
While 2011 may not be the year of the robot as far as the public is concerned, that same public is served by robots each and every day. More than 800,000 industrial robots make items with better precision and lower cost than human workers, to the point that every car company airs at least one commercial in which welding robots shoot sparks all over the car chassis and television screen. Artificial intelligence (AI) software, a key component in any social robot, helps us search the World Wide Web on Google and recommends books to us on Amazon.com.
This report takes a look at industries that are waiting for social robots, the type of jobs robots will do in those industries, and the likely time frame for commercial products to appear. All bets are officially hedged, because roboticists’ optimism tends to outrun their commercial product releases even more so than in the computer industry.
Robots in Healthcare
The healthcare industry hits the trifecta for robotic advancements: the Baby Boomer population bulge is now reaching Medicare age, the industry is chronically short of personnel to service the population—much less when Baby Boomers start wearing out en masse—and money is available. Adding money, need, and appropriate applications together means social robots will appear in healthcare areas first. Many doctors and hospitals have already purchased robot systems to assist with various surgeries, so a robot precedent has been set.
InTouch Health, based in Santa Barbara, Calif., announced in early 2010 that more than 100,000 clinical sessions have been performed through its Remote Presence (RP) telemedicine network. Indeed, the company’s RP-7, RP-Lite, and RP-Vantage mobile telepresence systems have been installed in hundreds of hospitals. It is the leading robotics company in medical telepresence, and a leading member of the American Telemedicine Association.
The company’s wireless, mobile Remote Presence model RP-7 is self-propelled and controlled by a remote operator over an 802.11wireless network. The unit includes a complete personal telepresence system, including a special video camera, LCD monitor screen that shows the remote doctor to the patient, and a microphone for audio communications.
The InTouch ControlStation includes both a desktop videoconferencing system and controls for the RP-7. InTouch says only one hour of instruction is sufficient for remote operators to control the RP-7 and its videoconferencing equipment. The RP-7’s camera, which the remote operator controls, pans, tilts, and zooms to give the physician multiple viewing options. And, the Federal Drug Administration gave InTouch permission to pull data from electronic diagnostic equipment such as electronic stethoscopes and ultrasound units through the robot to relay to the operator, speeding the flow of information.
RP-7 units cost in the $200,000 range, and a single ControlStation can manage multiple units. The RP-Lite offers the same telepresence features, but must be wheeled around by staff on the patient end of the telepresence call, lowering the cost by removing the self-propelled function. The RP-Vantage is a mounted system often used for training and collaboration.
Other Telepresence Robots
Numerous other companies are developing mobile telepresence robots that work much like the RP-7, but for business purposes rather than medical applications. Each can be controlled remotely, has some autonomous movement such as docking to recharge, and includes all necessary videoconferencing tools. While testing has been ongoing for several years on the following models, none have yet been released commercially:
- Vgo from Vgo Communications Inc.
- TiLR from RoboDynamics
- Texai from Willow Garage Inc.
- QB from Anybots Inc.
- Giraffe from HeadThere Inc.
- Lucas from A*STAR Social Robotics Laboratory (ASORO)
Many are excited about these remote telepresence robots, and some vendors promise to ship before year-end. With announced prices ranging from $5,000 (Vgo) to $15,000 (QB), these will be enterprise appliances, perhaps shared among multiple remote users during the day. With personal videoconferencing systems available free on most modern laptops, this once potentially huge videoconferencing market may actually become too price sensitive to support more than one or two robotic product lines. A key in their success is whether they can move to production before iPhones and iPads become the personal videoconferencing option of choice.
Although in some ways in-home patient monitoring is less critical than doctor-driven medical telepresence sessions, caregiving for the elderly or ill brings another set of challenges. In a person’s living quarters, with narrow doorways and objects that can be moved at any time, navigation requires a level of autonomy not seen in the remotely controlled InTouch RP-7. Therefore, a mobile service robot (MSR) must move autonomously, use AI to gauge if the person cared for needs assistance, and have the ability to initiate communication or a telepresence session in case of trouble.
The French company Robosoft, located in Bidart, makes Kompai, an MSR with a handful of sales in Europe to date. By using off-the-shelf components when possible, Robosoft has kept the price of Kompai under $20,000. The company announced Kompai, and took its first orders, at a long-term care insurance conference this year.
Kompai understands verbal commands, can take orders to track details such as items on a shopping list, and includes a touchscreen for more detailed interaction. Using a wireless Internet connection, Kompai can send streaming video to remote locations to verify help is needed if the client falls or otherwise fails to respond.
An American company, GeckoSystems Intl. Corp. of Conyers, Ga., makes the CareBot, essentially a motion-sensitive video camera on wheels with verbal AI support. Like Kompai, CareBot will query the client and send an alert if needed. CareBot is still in testing, so pricing has not been released, but the company hints that a CareBot will cost less than putting a senior into a nursing home or assisted living situation. That pegs the possible price in the $25,000 to $60,000 range, including onboard monitors for basic tests such as blood pressure and pulse rates.
GeckoSystems touts a variety of other functions for CareBot, such as more intensive healthcare support with telepresence. Being more autonomous than the RP-7, a CareBot can be used as commercial security by patrolling an area and sending notice if its motion detectors are triggered. Facial recognition software can also be loaded for sentry purposes.
Roomba, with more than 3 million units sold, wins the personal assistance race, unless assistance means more than vacuuming. Between Roomba and a basic, limited “Rosie the Robot” of Jetsons fame, lies another decade or more.
These long shots to commercial success include Toyota, whose violinist robot looks much like Honda’s Asimo, which also remains a research project. Anybot has shown its Dexter robot loading dishes into a dishwasher in a video, but nothing is close to release. STAIR (Stanford Artificial Intelligence Robot) is another research project, this one from Stanford University. Intel’s cleverly named HERB (Home Exploring Robotic Butler) is another project at least a decade away from initial commercialization.
Companionship or Therapy?
Elder care, while a huge market, is not the only one robot developers believe is worth exploring. For those wishing for some help losing weight, Hong Kong-based Intuitive Automata Inc. promises to soon release Autom, a nonmobile robot appliance that will gently chide you to exercise more and track what you eat so you lose weight. It will speak, but you must use the torso-mounted touchscreen to communicate in return. Weight program details are tracked on a back-end Web system for use from any computer, useful when the owner travels.
Does Autom straddle the line between caregiver and companion? Are the Kompai and CareBot caregivers or companions that can call for help, like a good babysitter? Academics may argue about classifications, but many companies believe their companion robots are a mixture of companion, therapist, and caregiver.
High on the cute meter is Paro, from Paro Robots USA Inc. in Itasca, Ill. Completely autonomous, Paro looks like a baby harp seal, covered in white fur with a button nose and big black eyes. Under its fur is an intelligent robot able to respond to touch, light, noise, temperature, and posture. It will turn toward the person speaking to it and make seal sounds.
Much like a cat but costing $4,700, Paro calms patients, responds to affection, learns to repeat behavior that generates affection (a concept unknown in most cats)—and convinced the Guinness World Records company to declare it the World’s Most Therapeutic Robot. More than 1,000 have been sold in Japan. Sales in the United States, promised in late 2009, have not yet begun.
Another robot designed to live at home with humans is PaPeRo, from electronics giant NEC, which won recognition as Japan’s most popular robot by Robot Life Magazine. The robot responds to speech and touch and recognizes faces. Somewhat autonomous, PaPeRo can also avoid obstacles and return to its charging station. Touted as therapeutic because children respond well to it, PaPeRo remains another research project.
Therapy and Education
Children with some types of mental disabilities respond more readily to robots than to people. Psychologists believe this is because the children believe the robots are predictable, which is not true of humans. Many robots for the education and therapy markets are being tested, and often financed, at least in part, by government grants and research projects.
AnthroTronix Inc., an engineering research and development company in Silver Spring, Md., uses an animated robot character in its education therapy software, to which children with cognitive issues respond well. They respond even better to the company’s CosmoBot robot prototype. The small, mobile robot made of blue and yellow plastic encourages children to focus and play, thereby developing pre-literacy and pre-numeracy skills. The U.S. Department of Education is working with AnthroTronix and the University of Southern California’s Center for Robotics and Embedded Systems to develop a commercial version of CosmoBot within the next 18 months.
USC research continues to develop robots that can be used to help normal childhood brain development and stroke rehabilitation. Bandit, a new research robot, blows bubbles, mimics the movements of autistic children, and wins them over. Far less complicated looking, but intelligent underneath, is Keepon from BeatBots LLC of Pittsburgh. The small, flexible robot looks like a 5-inch-tall yellow snowman, with cameras hidden in the googly eyes, and a microphone in the nose.
Children with ASD (autism spectrum disorders) have trouble interpreting facial expressions and body language. Many therapy robot systems can engage these children, but Keepon does it by using the four degrees of movement under the yellow snowman body to dance. Children in studies bond with Keepon, going as far as to give it a kiss, something they almost never do with their parents.
Still a research project with units available for purchase at $30,000 each, Keepon has yet to make it to the commercial market. One in 150 children have some level of ASD, according to some reports, so this market will grow considerably as the increasing number of children needing therapy exhaust the supply of trained therapists. Using CosmoBot, Bandit, Keepon, or another system will cost less than adding a therapist for the treating organization. Education funds are more limited than insurance and health funds, but getting preschool ASD children into therapy under their health insurance coverage is not a guarantee. This market has potential for moderate growth over the next decade, lower than that for healthcare but higher than for companionship.
Entertainment and Research
Since Keepon came to the public’s attention while dancing to a pop song on YouTube, is it therapy or education? Are robot toys education or research?
KumoTek LLC, maker of the KT line of humanoid robots as well as giant interactive dinosaurs like a Tyrannosaurus Rex named RoboSue, sells both advanced research robots and consumer hobby robot kits. WowWee’s Robosapiens were introduced as toys, but are so hackable they have become an entry point into robotics for many.
Nao from Aldebaran Robotics made news recently for being the first robot with some ethical programming. Although Nao looks like a cousin to a KT or Robosapien, the cost of $1,200 and up past $14,000 puts it firmly in the realm of research, and about the same price as the top-of-the-line KumoTek advanced research robot.
Eventually, people expect social robots to understand and exhibit humanlike emotions. Research at MIT’s Media Lab uses a humanoid face to display human emotions. Nexi, the lab’s new prototype with movable eyebrows, eyes, upper and lower eyelids, and mouth can display multiple human emotions clearly. Arms and hands add to the emotional resonance.
Follow the Money
Research with Nexi and other humanoid robots will continue to receive funding in the millions of dollars per year. Results of such research will be leveraged by educational therapy units, personal assistants, companions, and caregiving units. But research funds from government and industry will only allow the robot industry to maintain the status quo. Growth for social robot vendors will only come with large-scale commercial sales.
The statistical department of the International Federation of Robotics (IFR), a Frankfurt, Germany-based nonprofit association whose purpose is to promote and strengthen the robotics industry, reported that about 77,000 professional service robots were sold worldwide in 2009. Toys (relatively inexpensive research models) and Roombas have totaled about 5.6 million and 3.1 million units respectively. Total world market value for robot hardware, software, and services is just over $20 billion today. Assuming the economy recovers, IFR predicts annual sales of $17 billion in the global service robot market by 2013. The association further predicts the mobile service robot market will become a major market by that time. These dates and dollars seem optimistic; a more realistic target time frame might be 2015 to 2017. After all, of the 800,000 industrial robots working worldwide today, only a tiny fraction are mobile and social.
Japan, with its rapidly aging population, feels an urgent need to develop service robots quickly. Funding by the government and Trade Ministry is pushing for a full-scale commercial service robot market by 2025. As healthcare and caregiving robots develop, and Baby Boomers demand services for their declining health, social robots will become more important. They won’t be Rosie, but they’ll be helpful.
|The Bottom Line