Fall detection analysis with wearable MEMS-based sensors

Accidental falls are frequent and dangerous events for the elderly population, which can result into serious injury or fracture of bones especially hip bone injury or other joint fractures. There are several methods for detecting falls of elderly, such as camera-based, personal emergency response System (PERS), and wearable sensor-based. However, the camera-based method is limited by instrumented spaces and the PERS is suffer from inability to give an alarm after a fall. The wearable sensor-based fall detection is not limited to instrumented spaces, moreover, it is easily to detect the falls through tracking the kinematic information about the monitored person. In this paper, a wearable Micro-electromechanical Systems (MEMS)-based sensors module is designed for fall detection including one three-axis accelerometer, one three-axis gyroscope and one three-axis magnetometer. However, falls from activities of daily living (ADL) make it difficult to distinguish real falls from certain fall-like activities such as sitting down quickly and jumping. An approach is proposed using attitude angles to reduce false falls through tests of static postures and dynamic transitions. Meanwhile, the proposed method has real-time response and high computation efficiency.