Automatic Fall Detection Using Mobile Devices (Abstract Only)

Mobile devices are becoming an integrated part of every person's life regardless of age. Several studies have shown that falls account for over sixty percent of injuries in adults older than 65. These falls can leave permanent scars and limit the everyday lives of those people. Since people that fall are instructed not to move, it is crucial to provide them with a means to call for help. Systems like LifeAlert© have provided means for their clients to call for assistance by pressing a button they wear around their neck. However, if a person who has fallen is rendered unconscious, they are unable to do anything to obtain the assistance they need. This project focuses on developing a mobile application and related algorithms to allow mobile devices to determine if a person has fallen. The algorithms developed in this project takes in real-time data from a smartphone's accelerator and determines whether the sequential readings satisfy the three stages of falling (the free fall stage, collision with the floor stage, and remaining at rest stage). The results of the experiments demonstrate promising hands-free responses to the smartphone user falling. By using a smartphone or similar device, it is possible to autonomously call for assistance without the user having to manually press a button.