A Study about Identification of Pedestrian by Using 3-Axis Accelerometer

Abstract

Recently, various authentication systems are used in our society, for example, WYK system, WYH system and WYA system. However, most of them require user's some actions at domain of authentication system. Therefore, we suggest walking authentication system without user's special action by using 3-axis accelerometer embedded in mobile phone. We extract two feature amounts from basic experimentation and decide efficient feature amounts for walking authentication system. Furthermore, we suggest a method that detects mobile phone's position to authenticate more correctly and conduct main experimentation by using the method to decide effective feature amount. Moreover, we use data mining tool to classify mobile phone's position and five subjects and one unknown person in main experimentation. Then, we discussed the suitable feature amount of identification of pedestrian and examined the effectively of the system.