Methods for Person Recognition and Abnormal Gait Detection Using Tri-axial Accelerometer and Gyroscope

Abstract

This work reports methods to identify a person from gaits as well as identifying anomaly of gaits using "Koala". Koala, that was developed in National Chiao Tung University, Hsinchu, Taiwan, consists of an accelerometer and a gyroscope, and support Bluetooth communication. A gait is segmented from the signals acquired by Koala. Features extracted from the signals were used to identify a person. If a person can be defined by gait, the normal gait of a person should be able to be identified. In this work, we designed a method to detect abnormal gaits. If a practical wearable device available, the presented method can be applied to early detection of diseases, for example, an early stroke.