Person Identification Based on Walking Motion

• Main Authors: CHUNG, YUAN, 鍾原
• Other Authors: CHEN, JEN-YANG
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/3zjk24

碩士 === 銘傳大學 === 電子工程學系碩士班 === 107 === Intelligent recognition has gradually become a trend of technological development. Many identification techniques have been applied in our daily life, especially biometric identification is widely used nowadays. Therefore, this research proposes a Device-Free Identification(DFI) method, in which an Access Point(AP) is set up to send wireless signals in the environment, and uses Wireless Network Interface Controller to receive signals at the other end. We analyze Channel State Information(CSI) and Received Signal Strength(RSS) from the wireless network. In the environment with wireless network, human activity will cause reflection or scattering of wireless signals. Therefore, this research collects wireless signal data based on walking motion, establishes multiple training datasets. Then, we use Deep Learning and also Machine Learning for training and prediction. After experiments and comparison, it proves the feasibility of this person identification method and also has high accuracy.