Human Movement Detection Based on Acceleration Measurements and k-NN Classification

This paper addresses the problem of human movement detection and recognition using acceleration measurements and classification of acquired data with k-NN classification algorithm. For achieving the functionality of movement detection, two Crossbow's Mica2 motes are positioned on a person's body in order to measure the acceleration in the X, Y and Z axes. Several characteristic movements, such as falling, walking, running, sitting and standing can be successfully classified. We have developed a data acquisition, analysis and simulation environment based on the Tiny-OS, nesC and . NET technology. High level specialized movement detection tool was created. This tool can acquire, save, replay (simulate saved data), step-by-step present and classify all events during the measuring process. The paper presents the obtained results along with the system configuration and the initially required conditions.