A Scottish consortium of partners has developed a new device that can detect when elderly or vulnerable people living alone need urgent medical attention. The device uses artificial intelligence (AI) and the Internet of Things (IoT) to analyse electricity usage in the home to check whether household appliances and other electrical equipment are being used.
The device, developed by CENSIS (the Scottish Innovation Centre for Sensor, Imaging and IoT Technologies) and several others, connects wirelessly to a smart or traditional electricity meter and measures power usage every 10 seconds. By analysing the data, it is possible to see when specific electrical appliances in the home, such as a kettle, microwave or washing machine, are being used normally. Machine learning can be used to detect deviations from this normal usage pattern.
Automatic alerts
For example, if someone gets up at 8am every morning and puts the kettle on to make tea, this will be considered normal behaviour. If the kettle hasn’t been switched on by 9am on a given day, an automated text message is sent to the resident. If there is no response, an alert is sent to the contacts specified, such as a family member, carer, neighbour or response team, asking them to check on the person.
Lynda Webb, a senior researcher at the University of Edinburgh’s School of Informatics, said: “The idea of monitoring electricity usage in the home to see if someone needs help was conceived 10 years ago. The algorithm that underpins the current device was developed during a previous project involving 250 households in Edinburgh. It’s great to see that after all these years, the algorithm can finally be used to help people.”
Commercialisation and further development
Stephen Milne, Director of Strategic Projects at CENSIS, said: “Detecting sudden changes in behaviour is just the start. We want to develop the system even further so that it can eventually be used to detect gradual behavioural changes that could indicate health conditions that are slowly but surely developing, such as dementia.” The consortium now wants to launch the product on the market and deploy it on a large scale. “We are open to discussions with potential partners to build a long-term partnership,” Milne added.
Longer independent living
During the research, the technology was extensively tested in 19 homes in Glasgow, Dundee and Buckie in Morayshire as part of Blackwood Homes and Care’s Peoplehood project. The three-year project aims to develop a future-proof model for enabling older and vulnerable people to live independently for longer.
Lindley Kirkpatrick, Programme Manager for the Peoplehood project at Blackwood Homes and Care, said: “We have developed a range of activities, techniques and technologies throughout this project, and we are delighted with what we have achieved. But this new AI-based device really takes the cake. Thanks to this device, family members and other loved ones can be quickly alerted if something may have happened and, in the future, they may even be alerted to possible signs of illnesses such as dementia.”
Users do not have to worry about privacy. The device uses a special algorithm that processes all data on site, instead of on a central system. Thanks to this algorithm and Mydex’s leading storage of personal data, the user has full control over who can view and use the household data.
AI in Dutch elderly care
There is also growing attention for the potential of AI in Dutch elderly care. This is evident from the whitepaper ‘AI in elderly care’, which was published by ActiZ and Vilans at the end of last year and which extensively discussed the opportunities and challenges of AI in elderly care. The document showed that AI can improve the quality of care and reduce the workload for healthcare workers. It also showed that artificial intelligence can help to realise more efficient processes and increase the self-reliance of clients.
Remote monitoring of vulnerable or elderly people is one of the promising solutions that should make it possible for these people to live in their own environment for longer. Technology such as the bed sensor is used extensively for this purpose.
