An elderly man grips what look like a set of bicycle handlebars attached to a box on wheels. Looks can be deceptive, however.
The box on wheels is the PAMAID, the personal adaptive mobility aid for frail and elderly blind people. It combines the characteristics of a Zimmer frame with laser rangefinders, sonar and a significant degree of computer intelligence to help provide the elderly blind with a chance to get back on their feet again.
"It is carrying a fair bit of intelligence. It is basically carrying a PC inside," said Dr Gerard Lacey, a lecturer in the Computer Science Department at Trinity College Dublin. He became involved in the research while working on his PhD thesis in robotics at Trinity.
He visited an elderly relation in hospital and happened to see an elderly, frail blind man using a Zimmer frame while being guided by two nurses. "I realised that I could tailor what I was doing in robotics to help him," Dr Lacey said.
He went to discuss the potential with the National Council for the Blind (NCB). He was told that three-quarters of blind people in Ireland are over 65, and many are frail because of age. "There is actually a relationship between visual impairment and frailty," he said. "The lack of exercise is there because you get stuck in bed for your own safety."
The National Rehabilitation Board provided initial funding for a prototype, and Dr Lacey used it to produce an early version. He then assembled a six-member EU research group which included the NCB and research partners in Ireland, Greece, Sweden and Britain. "We set about refining the robot we had and then got out to the nursing homes."
They did what he called "rapid prototyping", which involved using advice and comments from those using the PAM-AID to make improvements. "They were able to participate in the design process. They designed it. You just do what they tell you," Dr Lacey said.
After seven robots and nine versions, the research team has now settled on two types, a powered version that moves along on its own and a passive version which is like a walker with wheels but can turn to avoid obstacles.
"It is less about robotics and more about the user interface," Dr Lacey said. "The interface was a big, big issue."
They used bicycle handlebars which allow a very small amount of movement left or right, movement which can be interpreted by the computer on board the PAMAID. "It is a kind of shared control. It is a bit like the way a guide dog works."
Both versions include a laser rangefinder which is extremely fast and accurate for identifying obstacles ahead. Its beams pass through glass doors, however, so PAM-AID also includes sonar. This is much slower but is good at locating table edges and irregular obstacles, and its signals can also locate a glass door.
The computer uses substantial artificial intelligence which allows it to make decisions about its surroundings and to interpret what the PAM-AID user is trying to do. PAM-AID does not carry maps or information about its surroundings. "It is location-independent," Dr Lacey said. It takes the data from the sensors and decides where it is standing, for example, in a corridor or in a room. It can recognise and identify these general features and tells the user what it "sees" via a computer-driven voice.
It is also making decisions about the commands coming via the handlebars from the user and tries to adapt to the person's behaviour. For example, if PAMAID says it is in a corridor but the user tries to steer it into the wall it can be programmed to stop and ask the user whether he really wants to do that.
PAM-AID is not suited for use by the able-bodied blind, who do much better with a cane. It would suit elderly who even if sighted would need some assistance to move about.
Dr Lacey is now collaborating with the Massachusetts Institute of Technology to improve the design. PAM-AID also has to be shown to be useful as a way to help elderly blind to get more exercise.
If the PAM-AID is proved to enhance independence and quality of life, Dr Lacey would hope to achieve a price comparable to that for an electric wheelchair, somewhere between £3,000 and £5,000. If PAM-AID can be shown to reduce the health detriment caused by lack of exercise in these patients, then it may be cheap at the price.