Researchers at MIT’s Computer Science Artificial Intelligence Laboratory (CSAIL) published a new paper this week showing that they can identify only a person they have previously monitored. Dina Katabi, a university professor at the university, has been studying the family health device for years, which can be used to monitor signs of human life, including heart rate and breathing. The wireless signal is used for analysis, so no physical contact is required.
The system, called RF-ReID (RADIO Re-identification), can continuously monitor up to 40 people in a collective living environment. Because RF-ReID detects heart rate and breathing, it can be deployed in nursing homes, which allows caregivers to track the vital signs of a person admitted to a hospital.
The system works as a broad trend to monitor the daily health of members. If unusual conditions such as breathing difficulties or decreased heart rate are detected, the caregiver will be alert. They can then separate these people from others in the group and test them for COVID-19. Katabi says:
“The new invention allows us to identify the same person’s measurements, but does not need to collect personal information about their identification, including their appearance. “
While other methods such as installing cameras and wearables have been tried for some time, Katabi says they are impractical and do not guarantee user privacy. RF-ReID also has other uses, such as early identification of Parkinson’s patients with unique walking patterns. So far, the system has been deployed in 19 different homes, and it has been concluded that after training at least 20 people, a new person can be re-identified in less than 10 seconds of physical activity.