Alex Mihailidis

A home that helps us “age in place”

Most people want a home that’s beautiful, but Dr. Alex Mihailidis dreams of one that’s “intelligent” too – and he’s working to make it a reality.

Imagine a place where a voice politely reminds you to turn off the stove or tells you where you’ve put your keys. It’s not as futuristic as it might sound. 

Dr. Mihailidis, a Scientist at Toronto Rehab, believes he’s onto something that could help older people to stay in their homes for a lot longer. He envisions a future where homes actually learn and adapt to the habits of their occupants.

“Our goal is to create an intelligent home that allows people to age in place, giving them independence and reducing the burden on caregivers and the health care system,” he says.

Right now, Dr. Mihailidis and his colleagues are working on a “prompting system” that uses artificial intelligence to monitor older adults with dementia as they wash their hands. Ultimately, it will also be used for other daily activities, such as dressing and cooking.

“The system uses computer vision to track their movements and provides them with verbal prompts, when necessary, to ensure that they don’t miss any steps along the way, such as using soap or turning off the taps,” he explains.

During testing, the device proved remarkably effective. Older adults were able to complete about 25 % more of the steps involved in hand washing on their own, without a caregiver. In one case, the caregiver actually spent less than 30 seconds with the person – instead of the usual seven or eight minutes.  

The researchers are now refining and enhancing the system with visual prompts. They think some users will benefit from seeing video reminders on a screen.  

In his busy office, Dr. Mihailidis and his colleagues are also developing a separate system to detect when someone falls at home. Using a ceiling-mounted camera, the computer vision system tracks a person’s movements in their home.  

“If they do fall, the system automatically detects this and calls for help,” says Dr. Mihailidis. “This means that even if they are unconscious or severely injured, they can still get help.” 

Working with Lifeline Systems Canada, a provider of personal response and support services, the researchers are now adding a feedback feature. “When a fall is detected, the system would ask the person: ‘Are you okay?’ Based on their response or lack thereof, the system makes an intelligent decision as to what level of help is required.”  

Down the road, the device could also be used to monitor the amount of time a person spends sleeping, or the number of visits to the bathroom. It would alert a caregiver if the behaviour did not fit with typical patterns.

Dr. Mihailidis stresses that the cameras do not actually record anything. Hooked up to a computer, all they “see” is moving blobs or silhouettes.

Only a decade ago – when Dr. Mihailidis started his Master’s degree – many people thought artificial intelligence was too complicated to be applied to such real-world problems.

“The computer power didn’t really exist back then. Now, we can write stuff on a standard PC without any problems,” observes Dr. Mihailidis. Something else has changed too, he says. There are now strong, collaborative connections between theoretical computer scientists, biomedical engineers and clinicians, he says.

Dr. Mihailidis believes his research can eventually help us all in our daily lives. “Right now, we’re developing systems for people with moderate-to-severe dementia and for older adults, but in the future this type of technology will likely serve many people.”

Wouldn’t it be great to live in a house that understands your habits and locates your wallet or gloves for you as you rush out of the door in the morning?

Says Dr. Mihailidis: “It would be great if your house could say: “Your wallet’s over there!” I often forget things and a reminder like that would be really, really helpful.”  

Curriculum Vitae and Publications since 2000

Dr. Mihailidis' Sunnybrook Research Institute profile page