"I've fallen, and I can't get up," Mrs. Fletcher, an elderly woman alone in a blue smock, shouts from her bathroom floor. By some fortune, Fletcher is wearing a Life Call remote control device. Rattled but systematic, she presses a button that activates her telephone. A dispatcher calls her landline, and she is able to yell to a speakerphone in the room adjacent. She cannot go to the phone because, you'll remember, Mrs. Fletcher cannot get up.
Life Call, 1987
"We're sending help immediately, Mrs. Fletcher," says the collected voice on the other end of the line. The owner of said voice, Gerald, is concerned but in control. He is in the sweet spot of his career; old enough to be authoritative and expert, but not so old as to be stereotyped as waning in competence. Despite being a phone dispatcher, Gerald wears a white uniform with a tie and an emblem on the shoulder, like a pilot. He wanted to be a pilot once, and that didn't happen. Sometimes things don't happen the way you hoped they might.
Life Call, 1987
You may remember this scene from a famous 1987 commercial for the Life Call remote control emergency device. It was a subject both of parody and a point of widespread awareness for this type of emergency-monitoring product.
In the next three decades, the number of Americans over 65-years-old will double. That's good news if you are the proprietor of any retirement homes, retirement villages, or retirement centers. But as many Americans opt to "age in place"—a popular term that might conjure the stagnation of a potted plant but is used in rebellion to the nursing-home-exile trope, describing continued independent living in their own home—there will be an increasing need for affordable, large-scale systems to monitor for emergencies.
U.S. Bureau of the Census / U.S. Department of Health and Human Services
The most common injury-related reason that elderly people are admitted to the hospital is that they fell over. It has been reported that almost one in one thousand of those falls results in death. Wearable devices and home-monitoring video cameras have become widely popular, but they have drawbacks in that they can be invasive or annoying. So an expert in radar imaging is on the case.
Dr. Moeness Amin is the director of the Center for Advanced Communications in the college of engineering at Villanova University. He was the lone academic representative at several NATO conferences on through-the-wall radar imaging. Amin's research focuses on various applications for radar motion-detection technology, including search and rescue, military, and law enforcement such as robberies and hostage situations. Now it also includes using radar to identify when people fall in their homes.
For national defense and security, this technology is used when you are trying to find out whether there are people behind walls, inside enclosed structures, and to know how many are there. "We could detect a female versus male walking, a child separate from a dog," Amin told me. "If somebody is shot, for example, the way he or she walks is different than a regular walk. This has been really of interest for the defense industry for many, many years."
Radar monitoring in living spaces is appealing an alternative to cameras, Amin told me. "There's a privacy issue if you put cameras in bedrooms or bathrooms. People don't like that."
Radar motion detection is also an alternative to wearable devices such as the Life Alert bracelet that, Amin says, people may forget or find uncomfortable. The device is meant to work through walls, just like spy equipment. If you put the unit anywhere in the house, it can detect that motion and classify it through different rooms. Though one living space could require multiple units, depending on the type of walls, Amin says. For example, penetrating a bathroom wall of ceramic tiles could be challenging. If a person falls behind a metal file cabinet, they could be lost.
The radar device emits and receives frequencies that vary depending on the motion of a person's body. At the most basic level, the radar will detect movement and classify the frequencies it receives in one of two ways: fall or no fall. That's not always an easy distinction, though. Since a radar signal might look similar when a person is plopping down in a bed and when they are slowly falling forward as their walker slides out of reach, the challenge is in minimizing false alarms. "The last thing you want," Amin says, "is to alert the first responder that somebody fell; they rush in, and the person is just sitting down drinking tea."
Spectrograms of typical motion patterns (Amin et al.)
The last, last thing you want is for the device to think a person is drinking tea when in fact they have sustained a serious fall. So the algorithm must be well trained to know the difference.
The system he is working on would learn an individual person's ways of moving. He calls it a Doppler Frequency Signature. It's what you look like on radar when you're walking, sitting, standing, or falling. An algorithm would learn your habits and know how to differentiate a sit from a fall. For people who use walkers or canes, their patterns will look different from people who don't, and the system can learn to recognize them. "So I want to train the radar to say, this person is, let's say paralyzed, so the way this person walks is different than the general population. So that when this person falls, the radar knows how this person falls, not how [just] anybody falls."
"In the future, I think the radar is going to be like a companion, living with the person, learning about the habits of the person, the way he walks, the way he sits, the way he stands."
Amin's goal with the algorithm is to be able to offer something extremely reliable, "so that when we say there is a fall, then we are very sure it is a fall." The proposed technique to process radar Doppler signatures for fall detection involves three parameters: extreme frequency magnitude, extreme frequency ratio, and length of the event. The algorithm looks something like this, but it goes on for 53 pages.
"There are drop falls—heart-attack falls where a person is standing and then boom. That is very dangerous. There are also tripping falls and the radar can actually distinguish between the two, because a tripping fall, the signature is very different. If it's a person tripping, you will find the person trying to reach to a chair or a table to try to prevent the fall. It's different from a heart-attack fall, which is boom. Drop."
The system will not only detect and classify, but localize the fall. It will say that the person fell in the living room. The radar can make a decision and route and alert to family member's cell phone or LED display in someone else's home. Also, when there are other people in the house, "the radar should detect that and switch off, because you don't want to confuse the radar."
Radar imaging of various movements (Amin et al.)
In developing the initial algorithm, Amin's test subjects are not elderly people. They are mostly Villanova students. "A young man falling is different than an elderly man falling," Amin says. So, we have a nurse practitioner who is coaching our students on how to walk and as though they were much older, so that the radar can be well trained."
Fall experiment setting, with radar detector (Gadde et al)
After the algorithm is complete, which Amin believes will be the end of this year, development will be a matter of finding a company that will put it on a platform. "That should not be difficult," Amin says. "The only challenge I see is basically selling this technology, not in terms of whether it is good or bad or efficient, in terms of whether the person will be comfortable having a radar unit."
"The psychology aspect of it, how the elderly will receive or accept that technology as an alternative to these wearable devices or cameras. That will be interesting."
You are receiving this email because you subscribed to this feed at blogtrottr.com.
No comments:
Post a Comment