The phrase "smart home" usually summons an image of lighting that goes on and off at a voice command, a stereo that automatically plays the music its owners like best, and vacuum cleaners that automatically keep the house shiny and bright.
At Carleton University in Ottawa, however, a multidisciplinary team is working on an entirely different approach to smart houses.
The object of this project, funded by the Ontario Research Network for electronic Commerce (e-Health), is to allow the elderly to remain in their homes rather than having to go to nursing homes, while intercepting incipient health issues before they become emergencies that require hospital stays.
Rafik Goubran, Ph.D. and acting dean of the Faculty of Engineering and Design, and Dr. Frank Knoefel, chief of staff and VP of medical affairs at SCO Health Service, which includes Elisabeth Bruyere Hospital and Research Institute, are leading the effort.
The team of engineering Ph.D. students, doctors and nurses at Elisabeth Bruyere Hospital in Ottawa, and elderly patients, is designing a sensor-rich environment to monitor elderly individuals in their homes unobtrusively.
"We want to monitor small deviations in behavior: are they having trouble sleeping, are they developing a joint problem that makes it hard to get up in the morning, is their weight decreasing," Dean Goubran says. "At the same time, we don't want to make them wear wires and sensors all day. We don't want them really conscious of the sensors, although of course they know the sensors are there and have given permission."
The patients put some additional limitations on the approach. For instance, several objected to cameras as being too obtrusive. The team considered placing microphones throughout the environment, but in the end they rejected this for the initial study, in part because elderly people sometimes speak too softly to be easily understood.
Instead, they use more creative sensors such as a pressure-sensitive bed pad and sensors on the doors of refrigerators and ovens to monitor the patient's eating habits. The Kinotex pad, made by Tactex Controls Inc., goes on the mattress under the sheet, so that you cannot see it. "If you press your hand onto the bedspread, it will create a low resolution image of your hand," Dean Goubran says.
This sensor alone can tell monitoring medical personnel whether the individual is having a good night's sleep or is tossing (a side effect of some medicines), how often he gets up to use the bathroom in the night (a possible indication of diabetes), whether he has trouble getting out of bed, even whether he is losing weight.
A tendency to press harder on the bed with one hand than the other can indicate the beginnings of a hip problem, for instance, that one day might cause the patient to fall and end up in hospital. By recognizing the problem early, the monitors can contact the individual's doctor or a son or daughter to get treatment before the problem becomes severe.
On the technology side, the issues focus on data handling and interpretation. And while for purposes of the experiment an apartment has been set up in the hospital, the plan is eventually to install sensors in seniors' homes throughout the region and use high-speed Internet connections to send the data to a central monitoring location. "These sensors create a loss of data," Dean Goubran says. "We had to decide what data to analyze close to the sensors, and what to send over the network. And we had to figure out how to make sense out of all that data."
Another issue is who interprets the data. As much as of the interpretation as possible needs to be automated, but ultimately human beings, probably nurses, will need to monitor the patients. Then those people need to be able to reach the appropriate doctors, nurses and family members for each individual. The who, what and where questions have yet to be answered.
The team is working with several vendors in the computer and medical areas. "I was particularly happy to find that Nortel was very interested in this project," Dean Goubran says. Nortel is working to build the communication solution that will allow the information from the sensor systems to be communicated to the people that need to be aware. Its approach is a multimedia communications solution allowing people to be reached on their device(s) of choice (computer, PDA, cell phone, etc.), on the access network of choice (cellular, Wi-Fi, wired, etc.) and in the format of their choice (e-mail, IM, voice, XML, etc).
Dean Goubran emphasizes however, that the focus of the study is on using the technology rather than developing it. "Some attempts at similar projects have been done solely by technicians," he says. "That is the wrong approach. The most important thing we have learned is that to do this right you need a team that includes medical personnel, technical people and the patients, themselves. And then you have to promote a great deal of communication among those groups. When we do that we find that what we think is the answer to a problem may not be what the medical people need, and sometimes the answer is much simpler than we expect."