Fall detection of the elderly using a smartphone

Fall detection of the elderly is a major public health problem. The probability of falls makes them dependent on others and restricts their freedom of movement. Although many fall detection methods have been developed to recognize falls in a real-time, most are inaccurate and inconvenient to use. In this paper we describe two methods for detecting the fall of a human body that can be implemented for the smartphones with built-in accelerometer. The first one used the raw data obtained from the sensor, and the second one - filtered data. In addition to the measuring a load factor, an important role in the algorithms has also a mobile device orientation to the ground. The assumption for the study was the localization of the smartphone in a right pocket of trousers - common in right-handed people. The experiment consisted in simulation the falls from different initial postures (standing, sitting, kneeling) in four directions (front, back, left, right). The results are satisfactory for detection of falls from a standing position. In conclusion, correct detection of falls based on the accelerometer built into the smartphone is possible after the filtration of the raw data, although the location of this device, the initial body position and direction of the fall have significant impact.

---

Opracowanie ze środków MNiSW w ramach umowy 812/P-DUN/2016 na działalność upowszechniającą naukę (zadania 2017).