One hour term authentication for Wi-Fi information captured by smartphone sensors

• Published in: ISITA 2016 : proceedings of 2016 International Symposium on Information Theory and Its Applications : Hyatt Regency Monterey Hotel, Monterey, California, USA, October 30 - November 2, 2016 pp. 330 - 334
• Main Authors: Kobayashi, Ryosuke, Yamaguchi, Rie Shigetomi
• Format: Conference Proceeding
• Language: English
• Published: IEICE 01.10.2016
• Subjects: Authentication; Biometrics (access control); Electronic mail; IEEE 802.11 Standard; Performance evaluation; Sensors; Usability

In recent years, an authentication method using humans' behavior is paid attention to such as risk-based authentication. A behavior authentication method may offer high usability for users because they do not require registration and authentication information to be entered manually. On the other hand, such an authentication method using one behavior factor is generally seen as having low authentication accuracy. For acquiring higher authentication accuracy, several behavior information is needed as authentication factors. The current subject is to find behavior information which can be an authentication factor. In a previous study to find a behavior authentication factor, a personal authentication method has been proposed utilizing Wi-Fi information which is captured by smartphone sensors. Generally, it is necessary that behavior information of a certain term is used as authentication information for behavior authentication, while it is sufficient that information at a moment of authentication is used for biometrics such as fingerprint authentication. The previous study has proposed using 24 hours data as authentication information. However, the long authentication term may increase the possibility of spoofing when a smartphone device is stolen, because the user's information remains in the device for the term. Therefore, in this paper, we propose a new authentication method using one hour data as authentication information, and apply our proposal algorithm to 47 subjects' data.