A Novel Machine Learning Based Fall Detection System using Acclerometer and Gyroscope

Falls are regarded as a very serious problem for the majority of elderly individuals. In this study, a prototype of a machine learning-based fall detection system is developed. In the developed system, an Artificial Neural Network (ANN) is used to detect different types of falls, namely, ‘falling forward’, ‘falling right’, ‘falling left’, ‘falling backward’, and ‘no fall’ (standing, walking, sitting) using the data received from 3-axis accelerometer and gyroscopic sensor.