To Pooja Viswanathan, artificial intelligence is about people creating smart tools that maximize human potential.
That’s why, while some researchers are developing robotic wheelchairs that simply transport users from one location to another, she’s adamant about giving humans the final say.
“It’s counterproductive to give people who are already suffering from physical and cognitive impairments wheelchairs that further erode their capacity to make decisions, when the goal is to give them back their independence,” says the University of BC Computer Science PhD student.
Currently, elderly people suffering from both physical disability and degenerative diseases such as Alzheimer’s are not granted access to powered wheelchairs due to safety concerns.
“It’s heartbreaking to see, in many long term care homes, elders slumped over a manual wheelchair because they are too weak or too confused to power their own way to where they want to be. It’s frustrating for them and often causes isolation and depression,” says Viswanathan, whose brother worked at a nursing home.
Viswanathan has developed a prototype smart wheelchair that could give users better quality of life and free up some health care resources. Named Navigation and Obstacle Avoidance Help, or NOAH, the system incorporates a stereo-vision camera that can easily be retrofitted onto any commercially available powered wheelchair, as well as software that learns the behaviour and decision-making patterns of its users.
“The twin cameras work similarly to human eyes,” says Viswanathan. “They memorize landmarks to create maps and calibrate distance to avoid collisions — which is the only time the wheelchair takes over control.”
Designed to operate on a laptop that fits under the wheelchair, and interact with the user through audio suggestions, NOAH is also capable of incorporating the user’s daily schedules.
“For navigation and for people suffering from cognitive impairments, audio prompts have been found to be more effective than visual cues,” says Viswanathan.
Post new comment