Researchers at MIT’s Computer Science and Artificial Intelligence Lab (CSAIL) have developed a way for a fully wireless system to monitor not only movement and vital signs contact-free, but also to track activities – in a more privacy-preserving way without using video. The system has the potential to be used at long-term care and assisted living facilities to provide a higher standard of support, while also ensuring that the privacy of the residents of those facilities is respected.
The system, which is dubbed “computer engineering definition” by the research team that created it, can identify whether a person is sleeping, reading, cooking, watching TV or more, by combining a map of a person’s living space with the types of activities that happen in different activities. The research team behind it trained the system on wireless signals generated by people performing known activities in these spaces, and was subsequently able to categorize activities from new people, in entirely new locations using the knowledge it gained through training.
Not only does the system preserve privacy more effectively than video-based monitoring, but researchers found that RF-Diary was actually more accurate, too. That means that it could accurately identify activity captions for individuals even when they were in dark settings, or blocked by other objects that would’ve thwarted visual checks. Overall, researchers found that their system was able to identify activities accurately over 90 percent of the time, across a range of 30 household activities.
No comments:
Post a Comment