Mobile Sensing Based Classification System For Human Fall Detection
Rohit Singh Rana, Neelam Bora, Shubham Chaudhary, Tarun Kumar · 2023
The elderly are at a considerable risk of falls, which is a primary cause of hospitalization and injury. Statistics from the CDC pointed out that one in four Americans 65 years of age or older are susceptible to falling incidents. Each year, resulting in over 800,000 hospitalizations and 27,000 deaths annually. Fall detection and rescue systems based on technology have been developed to address this issue, provide prompt emergency assistance, and reduce the incidence of injuries and associated healthcare costs. This paper systematically reviews sensor-based systems for detecting fallen individuals. This study aims to stimulate additional research in this area of study. One low-cost approach to fall detection illustrated in this paper involves using the built-in sensors of iOS phones to detect falls. A dataset and a classification system were developed to identify fall events utilizing the phone's accelerometer and gyroscope. This approach could benefit elderly individuals living alone without access to more expensive Systems designed to detect falls. Other sensor-based fall detection systems use various sensors, such as accelerometers, gyroscopes, pressure sensors, or cameras. By utilizing the Artificial Neural Network methodology, the gathered data has been evaluated, resulting in a proposed model with an impressive accuracy rate of 99%. This showcases the efficacy of the algorithm in identifying incidents of human falls.