RadioMe

Using artificial intelligence to personalise live radio
Community care worker joins an an elderly woman for a cup of tea while sitting at her kitchen table.
RadioMe is a £2.7 million project that uses artificial intelligence to adapt and personalise live radio, with the aim of transforming the lives of people living alone with dementia. It addresses key causes of hospital admission for people with dementia, such as agitation and not taking medication correctly. As a result, it is hoped quality of life will improve and people will be able to remain living independently at home for longer. Funded by the Engineering and Physical Sciences Research Council (EPSRC), it capitalises on the popularity of radio among the age group most likely to be living with dementia, developing a way to seamlessly ‘remix’ live digital broadcast so that listeners receive personalised reminders, information and music.
Running for four years, RadioMe is being trialled among people with dementia in Cambridgeshire and Sussex. The project includes a substantial ethical element, with significant time and money built in to ensure the technology is developed and co-designed with people with dementia and is not open to misuse.
Using a commercial bio-bracelet to measure physical signs like heart rate, as well as wireless speakers and an internet connection, RadioMe output is being produced in users’ homes by artificial intelligence software created at the Â鶹´«Ã½. An electronic diary completed by users and their carers is also a key element.
RadioMe is led by the Â鶹´«Ã½'s Interdisciplinary Centre for Computer Music Research (ICCMR) and Centre for Health Technology , and works with the Centre for Dementia Studies at Brighton and Sussex Medical School, the Glasgow Interactive Systems group at the University of Glasgow, and the Cambridge Institute for Music Therapy Research at Anglia Ruskin University.
Non-academic partners also include Sussex Partnership NHS Trust, the Alzheimer’s Society, MHA, Bauer Media and CereProc. BBC Research and Development and BBC Radio Devon have also been heavily involved in the project.
How RadioMe could work
A user switching on the radio in the morning might find their usual local station. But then, at a point dictated by the electronic diary, and at the start of a song, a DJ-like voice could override the real DJ and remind the listener to have a drink, take medicine, attend a memory café or anything else.
Another time, RadioMe might detect that the listener is becoming agitated via their bio-bracelet readings.
The software will then override the scheduled song choice and select a song from the user's personal library, known to be likely to calm them. Calming material could continue to be played until RadioMe detects the user is no longer agitated.

Research at the Â鶹´«Ã½

RadioMe is being led by Professor Eduardo Miranda from the University's Interdisciplinary Centre for Computer Music Research (ICCMR) and leading dementia researcher and Executive Dean of the University's Faculty of Health, Professor Sube Banerjee MBE . It builds on the substantial research into Brain-Computer Music Interface (BCMI) technology, artificial intelligence, music influencing emotion and the University's long-running involvement in shaping national policy on dementia.