Classification of movements based on wearable device data in biometric authentication

Abstract

The paper proposes a neural network algorithm for classifying human movements according to the accelerometer data, which is located in a mobile device. Intelligent algorithms for classifying movement types (single step, walking, walking on stairs, running) are considered on 9 types of different movements that a person performs in everyday life. The developed algorithm is proposed to be used in biometric authentication systems based on mobile phone data.