Hybrid Smartwatch Multi-factor Authentication

We propose Hybrid Smartwatch Multi-Factor Authenticator (HS-MFA), a system for leveraging a smartwatch as an additional form of user authentication. HS-MFA examines both inherent and subtle intentional gestures as a means of appropriately identifying a user. The system leverages a custom-developed Android Wear OS smartwatch app that records accelerometer sensor data for both user keystrokes and touchscreen interactions using a smartwatch paired with a smartphone. Observed authentication methods include username and password entry on a keyboard, pattern unlocking for smartphones, and PIN entry with a smartphone. Five different pattern matching methods were examined for a total of 96,880 genuine and impostor comparison tests by processing data from 246 unique user samples. The two best performing analysis methods achieved Equal Error Rate (EER) values between 0 and 67%, with an average of 28%, across the observed three axes of accelerometer sensor data captured through smartwatch. With notable accuracy and ease of use, this method would be a novel and intuitive multifactor authentication system for regular users as well as severely vision-impaired users to provide security for their digital assets in cyberspace.