Radar sensing and AI to cut healthcare costs by monitoring vulnerable people

Falls and fractures amongst people over the age of 65 account for over four million hospital bed days each year in England alone, costing the healthcare system around £2bn.

As well as the strain on accident and emergency departments and hospital wards however, such falls also lead to anxiety and loss of independence in the elderly.

Now a team of UK researchers, funded by EPSRC, is investigating the use of radar sensing and artificial intelligence to monitor vulnerable people, including the elderly and those with cognitive or physical impairments.

The system will be designed to monitor activity levels over long periods of time, to pick up on early signs of cognitive or functional decline, and to detect falls or strokes.

In this way it will provide family members or carers with information on the person’s health and wellbeing, and allow people to remain in their own homes for longer, preserving their quality of life, said Dr Francesco Fioranelli, the project leader at Glasgow University.

The radar system – a small box like that of a Wi-Fi router – will transmit and receive radio waves across rooms within the home, and then use machine learning to analyse the received echoes to understand how and where the person moves.

“The system would build up a map of the individual’s usual routine, and then try to spot any anomalies,” said Fioranelli. “The classic example is if someone crosses the corridor to go to the toilet once per night, and then all of a sudden they do that three or four times, it could be a sign that they are unwell, and they could be referred to a nurse or GP for a visit,” he said.

Unlike wearable sensors and video cameras, radar is contactless and non-intrusive, meaning individuals do not need to carry or interact with the device, which can be particularly difficult if the person has some form of cognitive impairment or dementia, said Fioranelli.

“Another advantage is that unlike cameras you don’t get a plain image of the person, so in terms of privacy it is easier to accept,” he said.

The researchers will take into account the needs and views of older people, carers, healthcare professionals and community members in developing the system.

The project also includes Glasgow-based CENSIS (Centre for Sensors and Imaging Systems), and the Health and Social Care Alliance Scotland.