University Times

By next year, demographers say, more than 25 million Americans will be aged 70 or older. It’s estimated that nearly one in 10 will have difficulty doing simple household tasks – and the U.S. population of frail elderly people is projected to keep on growing for years.

To hasten the day when robotic helpmates will be affordable fixtures in the homes of elderly persons, researchers from Pitt, Carnegie Mellon University and the University of Michigan are testing a mobile “nursebot” named Pearl and a robotic walker named the IMP this summer at area residences for senior adults.

The 4’5″ humanoid Pearl has cameras for eyes, a computer screen for a chest and a tray for carrying items to an elderly or disabled user.

The second robit, the IMP (intelligent mobility platform), is a walker equipped with audio/visual displays as well as sensors to guide spacially disoriented people to programmed destinations. The IMP can be used, un-powered, like a normal walker but also can travel and navigate under its own power. Stored in an out-of-the-way corner, it could be retrieved by remote control.

With funding from the National Institute of Nursing Research, researchers are putting Pearl and the IMP through their paces at the upscale Longwood at Oakmont retirement resort and the urban Lemington Home for the Aged, principal investigator Judith Matthews said in a recent lecture.

Senior adults will be asked to evaluate the prototypes’ performance and usefulness. “These devices are ostensibly being designed for use in people’s homes, but you have to start in a more controlled, institutional setting,” said Matthews, an assistant professor of heath and community systems in Pitt’s School of Nursing. “Pearl has not yet been in anyone’s apartment, much less in a private home. There are too many telephone books and shoes and things like that on the floor in most people’s homes.”

And Pearl, which relies on a laser beam set at 14 inches above the ground to sense solid objects, isn’t sophisticated enough yet to avoid any but the simplest impediments, said Matthews.

Pearl, funded by a $1.4 million grant from the National Science Foundation and now in its third incarnation, is built on a mobile platform with navigational software and infrared and/or solar sensors that enable it to get around (the robot creates and follows a map of its “home” environment) while tracking an elderly person’s movements. Researchers hope that Pearl eventually will be able to detect deviations from its human companion’s normal patterns – if the elderly person has fallen, for example, or is going to the kitchen or bathroom more or less frequently than normal.

Users operate Pearl (named for the color of its plastic and metal casing) through a touch-sensitive screen or by giving the robot vocal commands. According to Matthews, future generations of robot assistants might be able to remind elderly patients to take their medicine, eat meals and drink fluids, or schedule medical appointments. The machines also might help arthritis sufferers operate household appliances.

Nursing Dean Jacqueline Dunbar-Jacob, one of five co-principal investigators working on the nursebot project, observed: “We have succeeded in helping people to live longer, but chronic illnesses and disabilities often prematurely force elderly persons into personal care homes. The nursebot has the potential to help older persons remain independent longer and live a better quality of life.”

One thing that robots never will do is replace human nurses, researchers agree.

“It’s been a bit controversial, calling this a nursebot,” said Matthews. “When the media heard that name they immediately started to say we were developing a robotic nurse. I haven’t gotten much flak about it myself, but my understanding is that some nurses were insulted by the whole idea.”

Actually, Matthews said, nurses stand to benefit from future generations of robot assistants that could help them with physically demanding tasks or serve as “go-fers” to fetch things.

The latest version of Pearl has a round, Mr. Potato Head-type noggin. “You can take off various pieces and change the size and spacing of the eyes,” said Matthews, noting that research has shown that users react best to humanoid robots.

Pearl prototypes have cost about $100,000 to produce, but Matthews estimated that mass-produced robot assistants would cost between $5,000 and $10,000, or about the same as the motorized scooters that some disabled persons use.

While Pearl has received more publicity, the IMP probably has greater commercial potential because it’s a single-function device, according to Matthews. “Any technology that is multi-functional tends to be a harder sell because people don’t necessarily need all of the functions.

“You could make an analogy with cell phones,” Matthews said. “Once people bought into the single function of wireless telephone access and found how convenient it was, adding additional features like a camera was easy and didn’t make the product cost much more. It’s sometimes easier to add features to an established, single-function product than to market a multi-function product to begin with.”

– Bruce Steele