Person Identification using Wi-Fi CSI Focusing on Door Opening and Closing Motions

The identification of persons entering a room or building contributes to improved security and smart home services. Conventionally, person identification has used camera images and devices in the possession of the person. However, there are privacy protection concerns with unintended persons appearing in the camera image, as well as the burden of carrying and wearing the device. In this study, we propose a method to identify the identity of a person among several pre-registered persons based on door opening and closing motions, using Wi-Fi's Channel State Information (CSI), which is device-free and better than conventional methods in terms of privacy. In the evaluation experiments, 10 subjects were selected as the subjects. In the evaluation experiment, identification was performed on 10 subjects, and a high average F value of 0.83 was obtained overall.