Pepper, a 'humanoid' robot made in Japan, is being adapted in trials to care for vulnerable groups in the UK. Picture: BLOOMBERG/KIYOSHI OTA
Pepper, a 'humanoid' robot made in Japan, is being adapted in trials to care for vulnerable groups in the UK. Picture: BLOOMBERG/KIYOSHI OTA

Edinburgh — Pepper’s skill set includes making phone calls, identifying missing items in the kitchen and occasional aerobics instruction.

Now, after a surge in loneliness among vulnerable groups during the coronavirus pandemic, this robot’s potential as a companion has earned it a role in a Scottish university’s assisted living experiment with artificial intelligence (AI).

Scientists at Heriot-Watt University in Edinburgh have programmed robots, including Pepper — who was launched as the world’s first humanoid in Japan in 2014 — to perform tasks normally carried out by care workers.

“We are specifically interested in understanding the needs of the most vulnerable at this time and what technology could be used to make their lives better,” Mauro Dragone, the project’s lead scientist, said. “Successful innovation in the field is crucial to alleviate the strain on health and social care services.”

The experiment, named ambient assisted living, will initially focus on finding solutions for priority groups, whose vulnerabilities have been compounded by social isolation measures required during the pandemic.

For the research, Pepper and other robots have been put to work in a university laboratory configured to resemble a standard apartment, with a bedroom, bathroom, kitchen and living room.

Privacy concerns

By using robots to perform basic household tasks for those who have lost their vision or hearing, or suffer from dementia, the project hopes to ease pressure on care workers, who are often encumbered by high workloads.

Researchers, care providers and the end-users of assisted living services are being asked to use cloud and internet of things technologies — in which objects in the house are fitted with sensors linked to the internet — to participate remotely.

“We are transforming this lab into a remote, open-access lab so we can keep doing this work together, even while there is social-distancing in place,” Dragone said.

The project will trial “invisible” signal and sensor technology used to monitor participant’s behaviour, vital signs and constant state of health. Should the sensors detect a health emergency in a patient, an alert can be transmitted, allowing carers or emergency workers to take rapid action.

“In this laboratory we are experts in sensor technology that is invisible,” Dragone said. “Rather than attaching sensors, we use technology such as a Wi-Fi signal to detect the presence and activities of people at home.” He said this meant there would usually be nothing new to install or wear.

Researchers are “mindful” about privacy issues and the ethical issues that could come up in the project, said Dragone. A worldwide panel of ethics experts on AI is overseeing the experiment and will run “constant” risk assessments on the technology as it is developed.

Positive response

The Coalition of Care and Support Providers in Scotland, which represents 80 voluntary care providers that support about 200,000 people, has encouraged its members to collaborate on the project.

Emma Donnelly, the group’s digital programme manager, said Covid-19 has accelerated the need to implement “digital solutions” in care facilities. “There was already an existing drive for digital before the pandemic, but the crisis management answer has been to accelerate the implementation.” 

Donnelly said the reaction to the project had been “really positive” so far. 

“The focus of the project is on the end-user and there is a co-design element to it,” she said. “The care providers know that everything the project produces will support them in making their day-to-day lives a little bit easier.”


Would you like to comment on this article or view other readers' comments?
Register (it’s quick and free) or sign in now.

Speech Bubbles

Please read our Comment Policy before commenting.