A Wearable Fall Detector for Elderly People Based on AHRS and Barometric Sensor

Abstract

Falls and their consequences are among the major health care problems affecting functional mobility and quality of life of elderly people. Even for people living independently, falls are common occurrences. In this paper, we present a waist-mounted device useful to detect possible falls in elderly people. Through data coming from a three-axis accelerometer, a three-axis gyroscope, a three-axis magnetometer, and a barometer sensor integrated into our device, we are able to obtain a highly accurate estimation about posture and altitude of the subject. By means of such information, we have developed an extremely efficient system for fall detection, reaching 100% of sensitivity in commonly adopted testing protocols. In particular, the algorithm was tested according to three different experimental protocols, where volunteers performed several scenarios, including various types of falls, falls with recovery, and daily living activities frequent in the elderly. Results show that the proper combined use of the four sensors and efficient data fusion algorithms allow to achieve noticeable better performances to those obtained with similar systems proposed in the literature.