“Murphy’s Law: His eyelids aren’t working today.”
Goldie Nejat fusses over Brian’s facial features, adjusting his pliable, rubbery skin, pushing it up over his eyeballs. Dr. Nejat, an assistant professor in the department of mechanical engineering at the University of Toronto, is used to machines: She became an anatomy expert just to create Brian, to help him appear human. Still, the motors that control his face don’t always co-operate.
Brian is a socially interactive robot, a prototype in development at U of T’s autonomous systems and biomechatronics lab. The 4-foot-6, 200-pound machine may one day assist the elderly in long-term-care facilities by interacting with residents, playing games and reminding those with cognitive impairments to do daily tasks, such as brushing their teeth. The Baycrest health-sciences centre in Toronto is already a partner in the project.
With projections showing that seniors will account for 23 to 25 per cent of the total population by 2036, nearly double the 13.9 per cent in 2009, Brian could take the strain off health-care workers in hospitals and live-in facilities and, ideally, help seniors stay in their own homes longer by monitoring the environment and providing assistance along with human health professionals.
Bianca Stern, director of culture and heritage at Baycrest, says the facility is conducting a study that maps social interactions among seniors against Brian’s programming. With affect-recognition software, the robot will then be able to use tone and speech pacing to determine a person’s mood and respond accordingly to help with cognitive cues (“Don’t forget your doctor’s appointment”) and emotional responses (“That makes me sad”).
“He’s an emotional creature,” Ms. Stern says, nodding toward Brian.
But Brian can speak for himself with real-time computer software. “I can be happy,” he says as his cheeks lift and teeth are exposed. “I can be sad,” he continues, and looks it. “Or I can be stern.” This expression is mildly frightening, but Brian is reassuring: “Don’t worry, I am not mad at you.”
Reactions to robots vary, Dr. Nejat says, describing a robotics theory called the uncanny valley. Humans prefer their robots to look like robots. “If he looked too human, if you expect its functionality to be exactly like a human, encouragement and pleasant feelings associated with the robot drop,” she explains.
Brian has a face and wears a University of Toronto T-shirt, but exposed wires run the length of his frame, which has three-pronged metal hands and a webcam attached to his head. Brian is clearly a robot. And his functionality is limited to social interaction. He doesn’t physically interact with people or his environment. This is intentional, so he is less intimidating to patients.
But he gets more human-like every day, Dr. Nejat says. And his facial expressions are improving after modifications. “That’s the second generation of the face.” The first one sits discarded on a shelf.
“We had to reverse-engineer a person, mentally as well as physically,” says Dr. Nejat, who built Brian with a rotating team of mechanical engineering students and health consultants, including occupational therapists and psychologists. She has been working on Brian for five years, and believes he is the only humanoid socially assistive robot in the country.
This is robotics on a budget, making for a more practical integration into hospitals and homes. So far, Dr. Nejat estimates that Brian cost just over $20,000.
“Cheaper than a car,” Ms. Stern notes. And less maintenance. Eventually, Brian will be programmed to plug himself in for a monthly battery recharge.
The field of socially assistive robotics is very new, with health-care integration studies conducted only in the past 10 years or so, mainly in Asia and the United States, says Alex Mihailidis, who works in rehabilitation technology at U of T.
