Charlottesville-based University of Virginia researchers are developing an artificial intelligence-powered system that can analyze the voice and tone of caregivers to identify when they are stressed and prompt them to relax, The Daily Progress reports.
The system, which is a collaborative research effort among UVA, Ohio State University in Columbus and Knoxville-based University of Tennessee researchers, will target caregivers of Alzheimer's disease and dementia patients.
The goal of the system is to record caregivers as they talk and interact with the patient, specifically focusing on the pitch and tone of the caregivers' phrases rather than distinct words. The UVA research team's machine learning program then analyzes the recorded sounds and scans for tones of stress or anger, which, if detected, can then send an alert to the caregiver's smartphone to remind him or her to relax. The program uses machine learning to tailor relaxation recommendations to the individual caregiver, such as going for a walk or listening to music.
The research team is currently on the first year of its four-year grant for the project and is working toward improving the accuracy of the technology. The next step will be to recruit an estimated 50 participants to test the system.