Research Profile – Intelligent Motion

Researchers have developed a portable robotic rehabilitation system for stroke survivors.

[ Back to main article ]

Stroke patients who lose the ability to use one of their arms may soon get help from an intelligent robotic rehabilitation system that is small enough to sit on a tabletop. Researchers at the University of Toronto's (UofT) Intelligent Assistive Technology and Systems Lab and at the Toronto Rehabilitation Institute have designed a prototype of this novel system.

According to the World Health Organization (WHO), about 15 million people suffer strokes each year. Of those, 5 million suffer permanent disability, mostly in the upper body. Stroke not only damages part of the brain, it also affects the parts of the body that are controlled by the part of the brain that is damaged.

However, there is evidence that rehabilitation can trigger the brain to rewire itself, and create new pathways to restore function in the affected parts of the body. The new device, a 'haptic robotic system,' is a tactile feedback technology. Haptic (from the Greek word 'touch') technologies respond to the touch of a user and can apply force, vibrations or other movements.

According to Dr. Alex Mihailidis, an associate professor in the Department of Occupational Science and Occupational Therapy at UofT, the new device is important because it is small (about the same size as a fat briefcase) and it uses artificial intelligence (AI).

Other devices on the market are large, can't be easily moved, and are prohibitively expensive. This new portable device could be moved to the bedside of immobile patients, or possibly even used for home-based rehabilitation, Dr. Mihailidis says. And it promises to be a fraction of the cost of other robotic rehabilitation devices – around $10,000 to $15,000 compared to $75,000 and up.

The patient places his or her hand on a dome-shaped grip at the end of a jointed, moveable metal arm that sticks out from the side of the device. The robotic system is connected to a computer that runs several rehabilitation video games. The robotic arm can move on its own, at different speeds, to help guide patient's arm motion.

"The patient holds the end of the mechanical arm and moves it back and forth to practise reaching motions or other arm exercises. On the computer screen in front of them, it looks like they're driving a car, or rearranging objects," says Dr. Mihailidis.

The AI component tracks and learns about the patient's abilities and limitations. The robotic arm's motion and tasks in the games are adjusted by the AI accordingly. "For example, it may set a certain amount of force or resistance that the person is pushing against, or it may adjust how far the person should be reaching," he says. Traditionally, these are parameters a therapist would set manually, but the device can do this automatically and learn over time.

As part of the design process, the researchers conducted a survey of 230 rehabilitation and occupational therapists from Canada, Australia, the US and the UK. Therapist input was taken into account throughout the design and testing stages.

There are two more devices currently under construction, and they will be tested in what may be the first robotic rehabilitation clinic in the world by the end of summer 2012. The clinic will have three or four robotic devices, and it will be the site of clinical studies comparing the effectiveness of the devices to normal therapeutic approaches.

"Once we get the results, and the clinic is set up, and we've proven the model of this sort of clinic approach works, at that point is when we would make the device commercially available," says Dr. Mihailidis.

"The [artificial intelligence] learns about the person's progress through these exercises – how well or how poorly they may be doing – and then changes the parameters of the game."

– Dr. Alex Mihailidis, University of Toronto