Indoor Human Detection Using Wi-Fi

• Main Authors: Xiu-Qi Huang, 黃脩棋
• Other Authors: Chien-Cheng Lee
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/h48t3h

碩士 === 元智大學 === 通訊工程學系 === 105 === This study proposes a system based on the Wi-Fi signal to detect the status of indoor human. This method captures useful amplitude information from the channel state information and converts the amplitude information to the 2D images. Next, this study uses the "self-learning" nature of the deep neural network. The 2D images are used as input to learn with deep neural network, and to distinguish the differences in amplitude information due to different human states. This method can effectively reduce the recognition error that caused by the signal measured different environment, and it improves the accuracy of the classification of indoor human status. In this study, the human recognition rate was 99.96%.