Real Time Human Fall Detection Using Accelerometer and IoT

Abstract

Fall detection is an important research area where the focus is on detecting the human fall events and thus reducing the rate of mortality and resulting in better health care monitoring. In this paper, the main objective is to design a fall detection system which can classify the real-time sensor data into fall or non-fall events. In the proposed methodology, we stored the real time sensor data from a three-axis accelerometer sensor in an IoT data aggregator. This data is fetched from the cloud to the MATLAB desktop environment. And finally, the classification of data is done using supervised machine learning algorithm like logistic regression, and a real time alert will be sent on the same IoT data aggregator once the fall event is detected. This alert message will be used to provide immediate assistance and treatment. Such type of system can boost confidence among the elderly people who prefer to live alone.