Fall Detection Using Multi-sensory Accelerometer Sensor Network
Hans Hallez, Paul Alcala, Jeroen Boydens · Lirias · 2014
In this study we developed a measurement setup consisting of a sensor network to detect falls. A fall is characterized by a large peak in the magnitude of the acceleration vector. In contrast to the commonly used single sensor, we use a network of two sensors in order to increase accuracy and decrease false positives. The measurement setup consists of an acceleration sensor at the hip and one at the wrist. After gathering the acceleration data, the norm is calculated and we used Kaiser-Teager Energy Operator to detect sharp peaks. By setting the threshold on both the Kaiser-Teager Energy Operator peaks of the acceleration data of hip and the wrist, we can more accurately detect a fall.