Last month, IoT Insider received the news that an AI-powered and IoT-enabled device had been developed to be used for identifying when older or vulnerable individuals living on their own need medical help, an example of how a combination of AI, machine learning (ML) and IoT are being harnessed for good. To learn more, Editor Caitlin Gittins spoke to Nigel Goddard, Associate Professor, Informatics and Lynda Webb, Senior Researcher, Informatics, both from the University of Edinburgh, who collaborated with partners CENSIS, Mydex and Carebuilder on the project.
Against a backdrop of horror stories of individuals taking a fall and laying on the floor for hours after and the strain on hospitals within the UK, the Lifestyle Alert System (LAS) developed by the researchers has the potential to reduce demand on services.
The concept behind the LAS goes back 10 years to 2014, Goddard confirmed. “My lab was running a large project called IDEAL, involving 250 homes in Edinburgh looking at energy and we were working with these methods to identify what was being used so that we could give advice to the people living in these homes as to how they might be able to use less energy and save some money.”
It was Webb who spotted that the project could be applied to a healthcare application, Goddard said. “We co-created the Lifestyle Alert System all the way through,” she said. “So we started off by talking to people who live alone and family members, what is it they worry about? This issue [the scenario of an elderly or vulnerable individual living on their own, needing medical support with nobody to help] was one of the things that came up.”
Taking the learnings of the initial project involving 250 homes resulted in the creation of the device, which connects to a smart or conventional electric meter and uses machine learning to determine when electrical items like a kettle or shower are turned on and off, identifying any anomalies.
An individual living on their own could, for example, typically turn the kettle on at seven am every morning. If they decide not to switch the kettle on then for whatever reason, a check (in the form of an automated text message) is sent to the individual and if there is no response, then an alert is sent to their chosen contacts such as a family member or carer. Originally the researchers checked for use throughout the day of appliance use. As a result of the study and co-creation it was determined that a check on an individual three times a day, was sufficient.
“People were willing to wait for up to 10-15 minutes before a non-response (e.g. they may be unconscious) would trigger the alert to be sent to their chosen responder, Webb was surprised by this as people may already of been on the floor for some hours. But it was the fear of this potentially being a longer period of time that made people want regular checking,” she explained.
Webb explained that one of the surprising lessons to come out of the project was to learn that the “peace of mind” the device and subsequent alerts offered, and Goodard agreed.
“[During the project] we would send an individual the message first if we thought something was wrong so they can say, ‘No, everything’s fine.’ One of the things we thought might happen was that people would say they were receiving too many false alerts,” said Goodard. “As it turned out, even getting those false alerts was reassuring to people. Because it meant something was keeping track of them, something was looking out for them.”
In the UKRI funded Peoplehood project with Blackwood Homes and Care, the Lifestyle Alert System involved 19 households across Scotland.
Goddard said the project enabled us to look at how evolution of in- home technology and appliances may impact the AI from a technical perspective.
“Since we got started on this 10 years ago, people have begun putting in technologies like solar panels, heat pumps, appliances which are more energy efficient. What’s happening in the electricity system and the kind of signal that’s happening at the meter is quite different than when we trained our models.” He noted that these technologies hardly interfere with their system.
On the subject of major findings, Webb said: “Everybody that took part in the project was given the option to extend, chose to extend. “We learned things about people who could have more control of putting in, for example, to say they’re on holiday so the system was paused. One of the major findings was the peace of mind it gave people. If you’ve got peace of mind, as the individual and the carers, that helps to reduce stress and improves health.”
Another finding pointed to the danger of making assumptions about older individuals’ lifestyles, as the researchers found it was very varied. “Older people don’t necessarily have that regular lifestyle – some are regular, but others are varied in what they do.”
The research demonstrated how technologies including IoT and AI can be harnessed for good, providing support to those who need it most.
There’s plenty of other editorial on our sister site, Electronic Specifier! Or you can always join in the conversation by commenting below or visiting our LinkedIn page.
