A fall detection study based on neural network algorithm using AHRS

Human fall detection devices with high recognition rate have an important significance for the elderly and patient to detect their falls which may lead to dangerous or even death. In this paper, attitude angle and tri-axial acceleration of the Attitude and Heading Reference System (AHRS) module on the waist was used for the fall detection system. A fall detection method based on neural network was presented which could accurately distinguish falls from activities of daily living (ADL) including walking, jumping, sitting, bending, squatting, lying down, etc. The experiment was carried out with different groups of objects. The experimental results demonstrated that the proposed method was efficient, reliable as well as practical.