Imagine being monitored when you get up in the middle of the night, when you cook meals, when you take your medicine, when you turn the water on, and when you leave your home—but in a scenario in which it's all done to maintain your freedom, not subtract from it.

Two Washington State University researchers in Pullman are looking into just such a system, where the monitoring—not by video, but by sensors—might help protect older adults who are suffering from memory loss, perhaps allowing them to remain independent longer.

One of the professors, Diane Cook, has a research interest in machine learning and expertise in artificial intelligence, while the other, Maureen Schmitter-Edgecombe, has done research in memory intervention. Together, they are working on a system, known generally as a "smart environment" that could enhance the quality of life of older adults who want to stay in their homes longer despite mental and physical limitations.

"We define a smart environment as one which acquires knowledge about the resident and his or her surroundings and applies the knowledge to improve the resident's experience in the environment," says Cook, a professor in WSU's school of electrical engineering and computer science.

"The need for it has become very timely, because of the aging of the population," she says.

Adds Schmitter-Edgecombe, a professor in WSU's psychology department, "We were both working on the same goal; we were just taking different approaches to it. I'm excited about what we've done so far."

Cook says she's received about $7 million in primarily grant funding during the last 10 years for projects related to smart environments.

"Some of the greatest challenges are gaining acceptance of the technology in our target population and finding the right way to share learned information with residents and caregivers," she says.

There are no cameras in a smart environment, she says. Instead, a couple of sensors, small white boxes about one inch by two inches, would be installed in each room and would monitor motion and collect data about daily activities. Other sensors would track the use of appliances and water.

A computer at the residence would be equipped with software that could recognize and analyze information from the various sensors, and provide feedback, either immediately or in the form of reports, to both the residents and caregivers.

The computer could, for instance, tell caregivers—either while they visited the home or remotely—what the resident is or has been doing. A resident would be tracked as they moved around the home or apartment. The motion sensors would tell the computer how long the older person stayed in one place, and if they were moving around in a single location.

In the bathroom and kitchen, sensors would be attached to faucets and appliances, relaying to the computer what fixture was being used, as well as when, how long, and in what way it was being used. If a stove or oven were left on, or a faucet left running, the resident could be prompted by voice or sound to shut them off.

Sensors also would be attached to doors leading outside, letting the computer, and therefore caregivers, know if the person being monitored were leaving.

In the testing dwellings that have been set up to date, the sensors and wires are visible now throughout the home or apartment, but that likely would change in the future.

"I would envision a home with no (visual) differences at all, because the sensors are located behind walls and ceilings," Cook says.

Cook says she and Schmitter-Edgecombe started working together about two years ago, after Cook arrived at WSU.

"We have many related ideas that will keep us busy for at least the next five years," she says.

Cook says her research interest is in machine learning, and she was looking for a tangible application for her ideas.

"My father-in-law was in the housing industry and recommended I look into smart homes," she says. "Initially I was interested in seeing if smart homes could simply automate resident activities. After seeing how effective the software was, I wanted to focus attention on an application that would be beneficial to society."

Schmitter-Edgecombe, who has been doing work with aging and dementia, says she's interested in finding ways to increase the feeling of independence among older adults, and keeping people functioning in their homes as long as possible.

"My contribution is more the understanding of mild cognitive impairment and dementia and the types of intervention that are likely to prove most helpful," Schmitter-Edgecombe says. She works more directly with the participants in the research and their families and tests the participants. She and those she does research with make decisions about what activities of daily living are going to be tested, she says.

"From my side, we're going to have to make sure people are comfortable with it," she says.

Also, she says, "We want to encourage people to continue to use their cognitive abilities. If difficulties are detected, then that's when we want technology to assist them."

Cook estimates that it will be about five years before a monitoring system like the one she and Schmitter-Edgecombe are working to create will be available for sale to the general public. She estimates that currently, it would cost about $5,000 for total deployment of such a system in a person's living area.

"It's out there being used now for research purposes," set up in homes of older adults who have memory loss, Cook says.

Some aspects of a smart environment system need to be refined, and others, such as monitoring whether someone has taken their medication, need to be developed. That work will take several years, the researchers say.

WSU is doing one test deployment of a smart environment system at a time. The researchers just finished a deployment in a home in Idaho, and they are starting one in Washington this month. Also, there are three test-lab environments in Pullman.

"The tests have gone very well," Cook says.

She says several big-name companies are aware of the research and have contacted the researchers, including IBM, Intel Corp., Samsung Electronics Co., Qualcomm Inc., and others.

"We also have talked with smaller and startup companies, all of whom see that there is a need," Cook says. "They are all interested in the whole package, the sensors, data mining, and machine learning, activity monitoring and intervention. But different companies focus on different aspects of the technology, (such as) a software interface, a way of communicating with caregivers, etcetera."

Cook and Schmitter-Edgecombe are not alone in their field of research, as other schools are working on similar projects, including at the University of Florida, Georgia Institute of Technology, and the University of Missouri, Cook says.