Lifestyle Authentication Using a Correlation Between Activity and GPS/Wi-Fi Data

In recent years, lifestyle authentication, which combines multiple personal behavioral data for authentication, has been proposed as a new authentication method in addition to traditional knowledge-based authentication, possession-based authentication, and biometrics-based authentication. In previous research on lifestyle authentication, authentication scores of each authentication element were often calculated independently and used for the final authentication, ignoring the correlation between each element. It was also often difficult to apply lifestyle authentication methods in the real world because they required a large amount of preliminary data. In this paper, we propose a new method that solves these problems by using the correlation between GPS/Wi-Fi data from smartphones and activity data (activity types that are inferred from the metabolic equivalent of task (MET)) from activity trackers. We applied our method to the data collected in the MITHRA project, which is a proof-of-concept experiment of lifestyle authentication. As a result, we achieved an equal error rate (EER) of 0.087 and 0.130 when ideal data were obtained and not obtained, respectively.