International Journal of Electrical Engineering and Computer Science
E-ISSN: 2769-2507
Volume 3, 2021
A Personalized Fall Detection System for Older People
Authors: , , ,
Abstract: Recent developments and emergent technologies in medicine have improved life quality and increased life span, thus, caused a dramatic increase in the old adult population. This has increased the need for the care of the older people, particularly to deal with problems caused by advancing age, such as physical handicap and activity deficiencies. One of the most serious threats is falling, which can be fatal in the case of the older people who live alone. Several fall detection systems are in use, but there is a high need for a cost effective, compact and individualized system because of drawbacks in current versions, such as cost and privacy issues. We propose a tracking system which monitors people and detects falls using a personal device (a sports watch). The watch is strapped to the chest of the patient, and the 3-axis accelerometer data is used for real time fall detection. The system also allows for personal calibration to adapt to personal movement styles, such as dancing, walking and sleeping, therefore allowing a dramatic decrease in the amount of false fall alarms. In addition, the proposed system enables the monitoring of up to 1024 people in a same location using only one computer without cameras; hence, it avoids privacy problems in institutions, such as rest homes where cameras would be intrusive. An experimental study has been carried out to test the system with 40 participants of various ages and 2 mannequins. In the experiment, the system was calibrated separately for all individuals using their ADLs (Activity of Daily Life). The results showed approximately 85% accuracy in fall detection and 98% in false fall alarm reduction. Proposed system is very cost effective (less than $50 for a patient) and able to detect falls with significant accuracy levels, and differentiate between the falls and other ADLs.
Search Articles
Pages: 32-37
International Journal of Electrical Engineering and Computer Science, E-ISSN: 2769-2507, Volume 3, 2021, Art. #6