Scientists around the world are developing robots that can help older people with housework or ease the workload of personal care attendants. Siemens is involved in joint projects to develop a platform for service robots that is suitable for daily use.
By Michael Lang
Seniors at the Parkheim Berg nursing home in Stuttgart, Germany, were flabbergasted when a robot came rolling towards them, offering a cup of water. “CareObot 3 has a job. It is to offer water to older people who forget to drink enough fluids,” says project manager Dr Birgit Graf from the Fraunhofer Institute for Manufacturing Engineering and Automation (IPA) in Stuttgart.
“We were pleasantly surprised by how well residents accepted the robot,” she adds. “On the other hand, they viewed it more as a new form of entertainment and didn’t necessarily feel any obligation to drink the beverages it offered them.” CareObot 3 was developed for “bringremove” household services. For a oneweek practical trial at the nursing home, the robot was programmed to autonomously move towards a water cooler, press the button, use its gripper arm to place a full cup on a tray, and then offer the cup to residents.
This seemingly easy series of tasks actually required years of development work by Fraunhofer researchers. For one thing, the robot must be able to recognise its surroundings and avoid potential obstacles through the use of headmounted colour cameras and laser scanners fitted close to the ground. Each move of the gripper arm results from a series of elaborate calculations.
Robots such as CareObot 3 could become very important in the future as populations age, and as the number of young people who can care for the elderly also declines. Between now and 2050, according to Germany’s Federal Statistical Office, that country’s population segment aged 65 and over will increase by 36 per cent, and the proportion of those over 80 will rise by 136 per cent. At the same time, the number of people of employable age will decline by about 25 per cent.
Scientists are therefore trying to ensure that senior citizens will be able to care for themselves in their own homes for as long as possible. Many projects around the world are now studying ways to alter apartments in senior-friendly ways, such as by equipping them with electronic and sensor systems that would permit remote medical monitoring. In future, those who can afford it may also supplement such AAL (ambient assisted living) technologies with household robots.
“We achieved a high level of acceptance among the seniors in the nursing home because the robot’s utility was clear to them,” Graf explains. Some residents initially feared that the use of the robot would reduce their personal contact with care attendants or place them at the mercy of the machines. However, these concerns soon proved to be groundless. A robot such as Care-O-bot relieves attendants of time-consuming routine activities and actually allows them more time to spend with residents.
For the technology-loving Japanese, there is absolutely no problem with accepting robots. Very popular in Japan at the moment are humanoid robots equipped with a head, two arms, hands, and legs – just like humans. “The Japanese are moving ahead rapidly with robot research,” says Siemens manager Dr Peter Mertens, who until recently headed the company’s Corporate Technology office in Tokyo and has had extensive contact with many Japanese robotics scientists.
“More than 160 teams are now working on service robots in Japan,” he says. Researchers at Waseda University in Tokyo, for example, are developing a robot known as “Twendy-One” that has enough dexterity to operate a toaster. This robot can also help a handicapped person to get out of bed and into a wheelchair.
Already available for purchase today – although only in Japan – is a “robot suit” called HAL (Hybrid Assistive Limb) that was developed by Professor Yoshiyuki Sankai of the University of Tsukuba. HAL, an artificial exoskeleton that can support older people who can no longer walk, or whose strength is waning, is worn like a suit of armour. When humans move their arms or legs, their brains send a nerve signal to the appropriate muscle region. HAL is equipped with electrodes that register the resulting weak signals on the skin surface and convert them into commands for artificial limbs.
Also quite popular are therapy robots such as “Paro” – a robotic seal that produces a friendly look and a purr when caressed. Paro is also capable of learning. Among other things, it can behave in a way people like in order to be petted, for example. The little seal is already being used in Japan and Denmark to calm patients with dementia and reduce their stress levels.
Household robots and robot nurses will be part of the service robot market segment, which experts predict will soon be growing very rapidly. “We expect the number of service robots used in households to increase to 4,8 million units by 2012,” says Gudrun Litzenberger, director of the Statistical Department at the International Federation of Robotics in Frankfurt am Main.
“Most of these robots will mow lawns, vacuum and clean swimming pools. The market for robot nurses is still very small, but it’s expected to expand considerably over the next 10 years.”
A lot like people
Siemens has already developed a number of robotic technologies, although most of these are designed for industrial uses. Examples include cleaning robots for supermarkets, automated cranes for container ports, and robots for medical applications.
Also developed by Siemens Corporate Technology, and now being used by several customers, is a forklift robot that moves around autonomously. “The brain we developed for the forklift navigates in a manner similar to humans,” says Dr Gisbert Lawitzky, who is responsible for robot technology development at CT. “Our forklift relies on its eyes, its brain and a map of its environment.” A laser scanner mounted on the vehicle determines the forklift’s position and compares this data with map information that has already been learned and stored.
Siemens engineers also contributed key research results on robot recordings of everyday scenes and surroundings to the joint DESIRE (German Service Robot Initiative) project, which was managed by the IPA and funded by Germany’s Ministry of Education and Research.
The goal of the project was to develop a reference platform for household-service robots that would be suitable for everyday use. Lawitzky’s team addressed a key prerequisite for a robot’s ability to grip things: recognition of various objects within the robot’s radius of action and identification of the objects’ precise positions.
A stereo camera and intelligent algorithms enable the DESIRE robot to deal with even highly complex situations, such as when an object is partially covered. In such cases, the robot moves in a manner that allows it to gradually register the entire scene as fully as possible. The data thus collected is used by a software program to generate a 3D model of the surroundings, which is then used to plan gripper movements. By the time the DESIRE project was completed, the robot – with the help of Siemens technology – was able to recognise and differentiate between 100 everyday objects.
The technology used for the DESIRE robot’s gripping process was developed by the Karlsruhe Institute of Technology (KIT), and the lightweight arms were produced through a partnership between KUKA, a robot manufacturing company, and the German Aerospace Centre (DLR) under the direction of Prof Gerhard Hirzinger.
Cup of tea, sir?
Whereas the Japanese are the leaders in humanoid robot production – and they have many talking and trumpet-blowing examples to prove it – German researchers are more concerned with enabling robots to do everyday chores. “CT is particularly interested in applications for the Siemens sectors,” says Mertens. “We don’t build household or nurse robots ourselves, but we can support partners in this field by providing them with sensors, intelligent software, control systems, components from other types of service robots, and with information on the results of projects such as DESIRE.”
The other DESIRE partners continued their development of household robots even while the project was running. DLR’s Institute of Robotics and Mechatronics, for example, has a mobile humanoid robot named Justin (previous page) that can make tea, among other things. Justin unscrews the lid of a tea canister, seemingly without effort, turns it on its side, and taps it with its finger so that the instant tea trickles into a cup. The robot then pours water into a cup.
It may look easy – but it’s anything but trivial for Justin, which not only has to identify the objects correctly, but also has to know exactly what amount of force it may exert – otherwise it might end up crushing the tea canister lid. ‘In the future, robots will not only obey commands, but will also learn by watching humans.’
Such actions require delicate robotic hands such as the ones jointly developed by DLR and the Harbin Institute of Technology in China. The new five-finger hand built by researchers has four joints and is only a little larger than a human hand, according to Hirzinger. His latest innovation involves installing the drive for the fingers in the forearm, as is the case with humans, rather than in the hand. This drive unit contains 54 small motors that move the fingers via control wires.
Says Hirzinger: “In a few months, the artificial hand-arm system will be so agile, strong and pliable that it will come very close to the performance capability of a human hand for the first time ever.”
Watching and imitating
A humanoid robot called ARMAR, developed at KIT, can make itself quite useful – for example, by opening, loading and closing a dishwasher. Eighteen academic departments have joined forces at KIT to examine ways in which robots can co-operate with people. Professor Rudiger Dillmann, for example, is teaching ARMAR to learn by watching. As a result, the robot can now watch someone wipe off a table with a sponge and then begin to imitate the wiping movements.
Robots could even learn how to iron by observing humans. In the future, Dillmann's robots will not only obey commands, but also behave proactively. To do so, they will have to be familiar with their surroundings and know where they are allowed to go, and where they shouldn't go.
A team led by Dr Christof Schroter from the University of Ilmenau is now working with 17 partners in a new European Union AAL project called "CompanionAble", which aims to determine the ways in which intelligent homes and robots can complement one another. "Manufacturers of smart homes and robot researchers are now co-ordinating their developments in order to generate synergies," Schroter says.
By the time most members of the baby boomer generation have retired (around 2020), they will be in a position to benefit from this research. Until then, however, robots have a lot to learn.
Siemens Pictures of the Future
Video: From packing a dishwasher to doing the famous dance from the movie Pulp Fiction, you’ll catch these robots in action. [click here]