Location-independent fall detection with smartphone

TLDR: This paper presents a location-independent fall detection method implemented as a smartphone application for an inconspicuous use in nearly every situation of the daily life and applies a modular threshold-based algorithm which uses the acceleration sensor with moderate energy consumption.

Abstract: Due to demographic changes in developed industrial countries and a better medical care system, the number of elderly people who still live in their home environment is rapidly growing because there they feel more comfortable and independent as in a clinical environment or in a residential care home. The elderly often live alone and receive only irregular visits. Due to impaired physical skills the probability of falls significantly increases. The detection of falls is a crucial aspect in the care of elderly. Falls are often detected very late with severe consequential damages. There are existing approaches for automatic fall detection. They usually deploy special external devices. Elderly people often do not accept these devices because they expose their frailty. In this paper, we present a location-independent fall detection method implemented as a smartphone application for an inconspicuous use in nearly every situation of the daily life. The difficulty of our approach is in the low resolution range of integrated acceleration sensors and the limited energy supply of the smartphone. As solution, we apply a modular threshold-based algorithm which uses the acceleration sensor with moderate energy consumption. Its fall detection rate is in the average of current relevant research.