A Smartphone-Based Elderly Fall Detection Approach

Placed on the waist, an Android smartphone collects acceleration sensor data to power a proposed human fall detection system, designed to mitigate elderly fall injuries and detect falls in real time. By employing an algorithm integrating gesture recognition and fall detection, the system can distinguish fall behavior from normal daily activities. Upon detecting an abnormal fall, an alarm message, accompanied by GPS-acquired location, is dispatched. Simulations and experiments indicate the system’s high efficacy in distinguishing falls from daily behaviors, boasting high real-time performance, sensitivity, and specificity in its algorithm.