A study on real-time fall detection systems using acceleration sensor and tilt sensor

With society entering an aging society, a growing number of elderly people-who have lower bone mineral density levels and lower sense of equilibrium than young people-conduct social activities and sustain frequent fall injuries. They thus frequently sustain fatal fractures. Fracture is one of four major causes of elderly people's death, and causes fatal complications if not leading to death, so a system is very much required to sense falls and prevent bone fractures. Thus, herein was developed a system to sense falls in order to accurately assess and monitor falls. The proposed fall sensing system used three-axis acceleration sensors and two-axis tilt sensors, measured body movements generated during falls, and discriminated falls using fall detection algorithms. Fall experiments were conducted using a fall induction system developed on the basis of an air-pressure actuator which induced subjects' natural falls. Accelerations and inclinations of various major parts of the body, generated during falls, were measured, analyzed and combined to assess falls and their directions. To distinguish falls from ordinary life activities, accelerations and inclinations generated during ordinary life activities were measured, compared with those during falls, so as to develop fall detection algorithms. As a result of experiments, the fall detection system using such algorithms was capable of distinguishing falls from ordinary life movements such as walking, running, sitting down, standing up and lying. The system was capable of detecting falls and their direction. The system is believed to be helpful in developing our planned bone fracture prevention system.