Studying WiFi and Accelerometer Data Based Authentication Method on Mobile Phones

Abstract

Accessing applications on mobile phones has become a habit of our daily activities. These applications either require a PIN (Personal Identification Number) / password for authentication or donot have any authentication measurement at all because of the priority from user friendliness. In order to enhance the security of these mobile applications while not much compromising the user friendliness, we develop an approach to authenticate the user by using WiFi and accelerometer data collected within 3 seconds when the user opens a mobile application. The proposed approach has been evaluated on a dataset collected from a real-life scenario in which the data was collected from the participants' own mobile phones, and the participants have the full right to decide where/when to open the application. According to the experimental results, the proposed score-level fusion approach achieved EER (Equal Error Rate) at 9.19%, which indicates the feasibility of deploying such authentication mechanism on the mobile application.