Fall Detection Using Accelerometer, Gyroscope & Impact Force Calculation on Android Smartphones

Thailand has become an aging society. An accidental fall is a common problem found for the elderly. This research developed an Android mobile application for fall detection. The application compared threshold values of accelerometer and gyroscope together with impact force with the present values obtained from sensors to detect a fall. In this study, threshold values of impact force were computed based on an elder's weight to improve an accuracy of fall detection. Four types of elders' behaviors were studied. The first type was non-movement behavior (sitting and standing). The second group was constantly moving behavior (walking and running). Next group involved a change of movement (standing then sitting and sitting then standing). The last type were behaviors with falling (walking then falling on the buttocks, walking then falling on the back, running then falling on the back, and hopping then falling on the back). Results showed that the application could correctly identify the fall 97.33 percent.