Author(s)

Heng Yin

Corresponding Author

Heng Yin

Abstract

A real-time identification scheme based on cycle variation law of android phone's accelerometer is proposed in this article to overcome the drawbacks of the current identification system such as simplicity, lag-behind in real-time responsiveness and low intelligence level. The implementation of this scheme includes three stages, namely owner feature data collection, feature analysis and identity recognition, where a feature extraction algorithm based on a kind of clustering idea is used for the identification. Various tests have demonstrated a high practicality of this scheme in real life usage due to a high accuracy result as much as 90% achieved in the tests.

Keywords

Smart phones, Gait, Identification, Accelerometer.