Evaluating the specifications of built-in accelerometers in smartphones on fall detection performance
Journal
Instrumentation Science and Technology
Journal Volume
46
Journal Issue
2
Pages
194-206
Date Issued
2018
Author(s)
Abstract
In the daily life of the elderly, falling is a common accident and results in severe consequences. Hence, considerable research has been conducted on fall detection for the elderly. Smartphones, which have motion sensors and transmitting modules, have been applied as fall detectors. However, the performance of human activity recognition using various smartphones is usually lower than those using a dedicated smartphone due to different specification values of built-in smartphone accelerometers. This presents a challenge in the development of technology for using smartphones as fall detectors. In this study, we compared the specifications of built-in accelerometers which are essential for fall detection in four smartphones from different manufacturers. To test fall detection performance on various smartphones, 10 young healthy volunteers were included to wear the four smartphones and perform daily activities and fall event trials. Then the accuracies of a fall detection algorithm performed on four different smartphones were calculated and compared. The results show that the same algorithm has different accuracies when performed using smartphones from four different manufacturers. Also, a smartphone with insufficient accelerometer specifications exhibits low accuracy in fall detection. Based on our results, we suggest the effect of specifications of built-in accelerometers in smartphones should be considered when developing a fall detection system for multiple models of smartphones. To maintain high accuracy, a smartphone-based fall detection algorithm should only be used in cases in which built-in accelerometers meet certain specifications. ? 2017 Taylor & Francis.
Subjects
Accelerometer; elderly; fall detection; smartphones
SDGs
Other Subjects
Accelerometers; Manufacture; Signal detection; Specifications; Daily activity; elderly; Fall detection; Fall detectors; Healthy volunteers; Human activity recognition; Motion sensors; Transmitting module; Smartphones
Type
journal article