Fall detection using adaptive neuro-fuzzy inference system

A factor seriously endangers the people health is falling, particularly for patients and the elderly. Fall detection systems contribute in preventing the consequences of the late medical aid and injuries endangering the people health. The main problem within fall detection systems is how to correctly distinguish between a fall and the other daily activities. There are various types of fall detection systems each of which has different advantages and disadvantages. Wireless motion-sensor based systems such as accelerometer and gyroscope provide higher efficiency with lower limits. This study introduces a new fall detection method employing motion sensors in smart phones to collect data due to the ease of access and application. To provide high efficiency for people with various ages and conditions, this method also takes advantages of adaptive-fuzzy neural networks for learning and inference. These methods correctly detect all 4 types of fall from 9 main daily activity groups. © 2016 SERSC.